Re: D2D on AIX
Hi Ian! Because you can't reclaim DISK volumes. TSM stores files in aggregates which will only be freed when all files within the aggregate become expired. Kindest regards, Eric van Loon KLM Royal Dutch Airlines -Original Message- From: Ian Hobbs [mailto:[EMAIL PROTECTED] Sent: Tuesday, September 21, 2004 00:26 To: [EMAIL PROTECTED] Subject: Re: D2D on AIX Question, Why not use the DISK device class with RAW volumes? Personally, I find FILE classes a pain for user storage because you DO have to perform reclamation on them. Ian Hobbs On Mon, 20 Sep 2004 14:48:07 -0400, Eliza Lau wrote: >Okay. I got it. It is a pain to manually define the volumes, but it can be >done. I also received your pdf file. > >Thanks to everyone who answered, >Eliza > >> >> It depends upon how you configure things. For dynamic allocation of >> volumes, then yes you are limited to the size of the file system that >> you mount on that mount point. However if you define the stgpool >> volumes explicitly using the DEFINE VOLUME command, you can place the >> volumes across as many file systems as you want. I will email you a PDF >> presentation IBM has on Disk Only backups. >> >> >> H. Milton Johnson >> >> -Original Message- >> From: ADSM: Dist Stor Manager [mailto:[EMAIL PROTECTED] On Behalf Of >> Eliza Lau >> Sent: Monday, September 20, 2004 12:11 PM >> To: [EMAIL PROTECTED] >> Subject: D2D on AIX >> >> Our 3494 with 3590K tapes in 3 frames is getting full. Instead of >> adding another frame or upgrading to 3590H or 3592 tapes we are looking >> into setting up a bunch of cheap ATA disks as primary storage. >> >> The FILE devclass defines a directory as its destination and JFS2 has a >> max file system size of 1TB. Does it mean the largest stgpool I can >> define is 1TB? >> >> My Exchange stgpool alone has 8TB of data. Do I have to split it up >> into 8 pieces? >> >> server: TSM 5.2.2.5 on AIX 5.2 >> database 90GB at 70% >> Total backup data - 22TB >> >> Eliza Lau >> Virginia Tech Computing Center >> [EMAIL PROTECTED] >> >> Ian Hobbs [EMAIL PROTECTED] === "Never argue with an idiot. They drag you down to their level then beat you with experience." -Dilbert ** For information, services and offers, please visit our web site: http://www.klm.com. This e-mail and any attachment may contain confidential and privileged material intended for the addressee only. If you are not the addressee, you are notified that no part of the e-mail or any attachment may be disclosed, copied or distributed, and that any other action related to this e-mail or attachment is strictly prohibited, and may be unlawful. If you have received this e-mail by error, please notify the sender immediately by return e-mail, and delete this message. Koninklijke Luchtvaart Maatschappij NV (KLM), its subsidiaries and/or its employees shall not be liable for the incorrect or incomplete transmission of this e-mail or any attachments, nor responsible for any delay in receipt. **
Re: D2D on AIX
Hi Eliza! You do want several smaller files, rather than a few very large files because each client session will allocate a volume. File volumes cannot be used concurrently by more than one session. Kindest regards, Eric van Loon KLM Royal Dutch Airlines -Original Message- From: Eliza Lau [mailto:[EMAIL PROTECTED] Sent: Monday, September 20, 2004 19:11 To: [EMAIL PROTECTED] Subject: D2D on AIX Our 3494 with 3590K tapes in 3 frames is getting full. Instead of adding another frame or upgrading to 3590H or 3592 tapes we are looking into setting up a bunch of cheap ATA disks as primary storage. The FILE devclass defines a directory as its destination and JFS2 has a max file system size of 1TB. Does it mean the largest stgpool I can define is 1TB? My Exchange stgpool alone has 8TB of data. Do I have to split it up into 8 pieces? server: TSM 5.2.2.5 on AIX 5.2 database 90GB at 70% Total backup data - 22TB Eliza Lau Virginia Tech Computing Center [EMAIL PROTECTED] ** For information, services and offers, please visit our web site: http://www.klm.com. This e-mail and any attachment may contain confidential and privileged material intended for the addressee only. If you are not the addressee, you are notified that no part of the e-mail or any attachment may be disclosed, copied or distributed, and that any other action related to this e-mail or attachment is strictly prohibited, and may be unlawful. If you have received this e-mail by error, please notify the sender immediately by return e-mail, and delete this message. Koninklijke Luchtvaart Maatschappij NV (KLM), its subsidiaries and/or its employees shall not be liable for the incorrect or incomplete transmission of this e-mail or any attachments, nor responsible for any delay in receipt. **
Which TSM client code to use for Linux
We are ordering 2 IBM Blade servers for a Oracle project. They will be running Red Hat Linux Enterprise Advanced Server 3.0. What is the most stable TSM client code? The information transmitted is intended only for the person or entity to which it is addressed and may contain CONFIDENTIAL material. If you receive this material/information in error, please contact the sender and delete or destroy the material/information.
Urgent .. Empty files restored ..!!!
HI TSMers under ZOS , using TSM4.2 one client W2000 formated the hard disk and he needs to restore his files After restore operation completed . objects failed =0 objects restored = 13766 but number of bytes transferred = 0! some file names looks rubbish and all the files are empty , i mean the size=0 bytes for all but the filespace size capacity (37283.6 MB) and Pct util=38.3 note :- the format of the filespace is NTFS & the format of the hard disk partion also is NTFS -any one can guide me where is the data itself ...is it lost ?? why this problem happened please help .. it is urgent Thanks in advance = h2mso4 __ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com
Re: Which TSM client code to use for Linux
We've been using 5.1.6.0 with no problems. I expect we'll move to the latest 5.2 sometime in the not so distant future just to stay current. David Ehresman University of Louisville >>> [EMAIL PROTECTED] 9/21/2004 7:21:29 AM >>> We are ordering 2 IBM Blade servers for a Oracle project. They will be running Red Hat Linux Enterprise Advanced Server 3.0. What is the most stable TSM client code? The information transmitted is intended only for the person or entity to which it is addressed and may contain CONFIDENTIAL material. If you receive this material/information in error, please contact the sender and delete or destroy the material/information.
Re: Urgent .. Empty files restored ..!!!
On Sep 21, 2004, at 8:54 AM, hussein hasan wrote: ... some file names looks rubbish and all the files are empty , i mean the size=0 bytes for all ... Hussein - My first thought would be to do 'dsmc q backup -ina ' for at least one of the problem filenames and see what's in the server for files: there may be older versions that are valid. This is to say that the most recent backup was performed when the system was trashy. Before restoring, I would run some good disk-checking tools to assure that the disk is a viable container for the data you seek to restore. And, as always, make sure that the system is up to date on patches which keep malicious intruders out of the system...who might cause such damage. Richard Sims
Re: Urgent .. Empty files restored ..!!!
If he used dsmc to run the restore, is there a chance that all he got was directories? I've done that in the past with poorly typed dsmc commands, albeit under Unix, not Windows. [EMAIL PROTECTED] Richard Sims <[EMAIL PROTECTED]> Sent by: "ADSM: Dist Stor Manager" <[EMAIL PROTECTED]> 09/21/2004 08:33 AM Please respond to "ADSM: Dist Stor Manager" <[EMAIL PROTECTED]> To [EMAIL PROTECTED] cc Subject Re: Urgent .. Empty files restored ..!!! On Sep 21, 2004, at 8:54 AM, hussein hasan wrote: > ... some file names looks rubbish and all the files > are empty , i mean the size=0 bytes for all ... Hussein - My first thought would be to do 'dsmc q backup -ina ' for at least one of the problem filenames and see what's in the server for files: there may be older versions that are valid. This is to say that the most recent backup was performed when the system was trashy. Before restoring, I would run some good disk-checking tools to assure that the disk is a viable container for the data you seek to restore. And, as always, make sure that the system is up to date on patches which keep malicious intruders out of the system...who might cause such damage. Richard Sims
Re: D2D on AIX
IBM gave a webinar on DISK ONLY backups including the advantages of DISK vs. FILE device classes. While your mileage may vary, in general it seems that a FILE devclass will give better performance for large pools (read TB not GB). Two quick examples: 1) With DISK TSM keeps track of each 4K block in the DISK volumes. This means that TSM must maintain a map of all those blocks and search/update that map every time a file is saved/expired. Also your files will be fragmented within those DISK volumes leading to further performance problems. 2) When it's time to backup what's on disk to offsite tapes, TSM has a speedy shortcut with SEQUENTIAL device classes. TSM keeps a flag for each sequential volume, when the sequential volume is backed the flag is set, and when the sequential volume is written to the flag is cleared. This means that when it's time to back-up those primary stgpool sequential volumes to a copypool TSM only needs to examine those files in the sequential volumes with the flag cleared. With the way TSM works, this greatly reduces the amount of time required by TSM to determine what data needs to be backed up. With a DISK device class, TSM has no choice but to examine each file in the STGPOOL being backed up to determine if it has been previously backed-up to the copypool. Incidentally, IBM hinted that a future enhancement would be to allow a list of mount points (directories) to be assigned as the destination to FILE device classes. This would allow utilization of dynamic allocation across multiple file systems. Of course one drawback with dynamic allocation is that fragmentation can occur overtime. Your particular OS will greatly influence the severity of this problem, however defining the stgpool volume explicitly will prevent that problem. H. Milton Johnson -Original Message- From: ADSM: Dist Stor Manager [mailto:[EMAIL PROTECTED] On Behalf Of Ian Hobbs Sent: Monday, September 20, 2004 5:26 PM To: [EMAIL PROTECTED] Subject: Re: D2D on AIX Question, Why not use the DISK device class with RAW volumes? Personally, I find FILE classes a pain for user storage because you DO have to perform reclamation on them. Ian Hobbs On Mon, 20 Sep 2004 14:48:07 -0400, Eliza Lau wrote: >Okay. I got it. It is a pain to manually define the volumes, but it >can be done. I also received your pdf file. > >Thanks to everyone who answered, >Eliza > >> >> It depends upon how you configure things. For dynamic allocation of >> volumes, then yes you are limited to the size of the file system that >> you mount on that mount point. However if you define the stgpool >> volumes explicitly using the DEFINE VOLUME command, you can place the >> volumes across as many file systems as you want. I will email you a >> PDF presentation IBM has on Disk Only backups. >> >> >> H. Milton Johnson >> >> -Original Message- >> From: ADSM: Dist Stor Manager [mailto:[EMAIL PROTECTED] On Behalf >> Of Eliza Lau >> Sent: Monday, September 20, 2004 12:11 PM >> To: [EMAIL PROTECTED] >> Subject: D2D on AIX >> >> Our 3494 with 3590K tapes in 3 frames is getting full. Instead of >> adding another frame or upgrading to 3590H or 3592 tapes we are >> looking into setting up a bunch of cheap ATA disks as primary storage. >> >> The FILE devclass defines a directory as its destination and JFS2 has >> a max file system size of 1TB. Does it mean the largest stgpool I >> can define is 1TB? >> >> My Exchange stgpool alone has 8TB of data. Do I have to split it up >> into 8 pieces? >> >> server: TSM 5.2.2.5 on AIX 5.2 >> database 90GB at 70% >> Total backup data - 22TB >> >> Eliza Lau >> Virginia Tech Computing Center >> [EMAIL PROTECTED] >> >> Ian Hobbs [EMAIL PROTECTED] === "Never argue with an idiot. They drag you down to their level then beat you with experience." -Dilbert
AIX/TSM Paging Space and Memory Settings
Well, interesting results from the changes we made to our vmtune settings yesterday. We have one TSM instance running with JFS2 for database, logs and storage pools, and another on all raw. We run two database backups daily, a full and a snapshot. Last evening's snapshot was 5,000,000 pages an hour faster on the JFS2 TSM instance than any previous DB backup we've had. It was about the same speed as a backup on the server using raw volumes. Of course, one night doesn't mean a whole lot in the big scheme of things. But if this continues, I'll be impressed. Curtis [EMAIL PROTECTED]
Re: D2D on AIX
And with DISK device class there is no multi-session restore. -Original Message- From: Ian Hobbs [mailto:[EMAIL PROTECTED] Sent: Monday, September 20, 2004 5:26 PM To: [EMAIL PROTECTED] Subject: Re: D2D on AIX Question, Why not use the DISK device class with RAW volumes? Personally, I find FILE classes a pain for user storage because you DO have to perform reclamation on them. Ian Hobbs On Mon, 20 Sep 2004 14:48:07 -0400, Eliza Lau wrote: >Okay. I got it. It is a pain to manually define the volumes, but it can be >done. I also received your pdf file. > >Thanks to everyone who answered, >Eliza > >> >> It depends upon how you configure things. For dynamic allocation of >> volumes, then yes you are limited to the size of the file system that >> you mount on that mount point. However if you define the stgpool >> volumes explicitly using the DEFINE VOLUME command, you can place the >> volumes across as many file systems as you want. I will email you a PDF >> presentation IBM has on Disk Only backups. >> >> >> H. Milton Johnson >> >> -Original Message- >> From: ADSM: Dist Stor Manager [mailto:[EMAIL PROTECTED] On Behalf Of >> Eliza Lau >> Sent: Monday, September 20, 2004 12:11 PM >> To: [EMAIL PROTECTED] >> Subject: D2D on AIX >> >> Our 3494 with 3590K tapes in 3 frames is getting full. Instead of >> adding another frame or upgrading to 3590H or 3592 tapes we are looking >> into setting up a bunch of cheap ATA disks as primary storage. >> >> The FILE devclass defines a directory as its destination and JFS2 has a >> max file system size of 1TB. Does it mean the largest stgpool I can >> define is 1TB? >> >> My Exchange stgpool alone has 8TB of data. Do I have to split it up >> into 8 pieces? >> >> server: TSM 5.2.2.5 on AIX 5.2 >> database 90GB at 70% >> Total backup data - 22TB >> >> Eliza Lau >> Virginia Tech Computing Center >> [EMAIL PROTECTED] >> >> Ian Hobbs [EMAIL PROTECTED] === "Never argue with an idiot. They drag you down to their level then beat you with experience." -Dilbert
FW: unsubscribe
bye
Re: D2D on AIX
From: ADSM: Dist Stor Manager [mailto:[EMAIL PROTECTED] On Behalf Of Rushforth, Tim >And with DISK device class there is no multi-session restore. Are you sure? I seem to recall using RESOURCEUTILIZATION to run a multi-threaded restore or two from DISKPOOL. -- Mark Stapleton ([EMAIL PROTECTED]) Berbee Information Networks Office 262.521.5627
Re: Duel tape write to LTO's
Hi Milton, When TSM writes simultaneously to the copypool would this be on the 2nd Library for duel tape backup? "Johnson, Milton" wrote: > You should be able to create a PRIMARY STGPOOL named TAPEPOOL and a > COPY STGPOOL named COPYPOOL with both of them having a sequential access > (tape) DEVICE CLASS such as DLT or LTO. Both stgpools can be in the > same library. On the stgpool TAPEPOOL definition you set the > COPYSTGPOOLS parameter to COPYPOOL. Then when your client backs up to > TAPEPOOL, TSM will simultaneously write to COPYPOOL. Of course having > an adequate number of tape drives is required. > > H. Milton Johnson > > > -Original Message- > From: ADSM: Dist Stor Manager [mailto:[EMAIL PROTECTED] On Behalf Of > Timothy Hughes > Sent: Friday, September 17, 2004 8:01 AM > To: [EMAIL PROTECTED] > Subject: Re: Duel tape write to LTO's > > Hello, > > I was told that this could work If I have 2 backup disk pools. > > Like I have backup diskpool, then I can have like say a DB2 backup > diskpool then I can have the next storage pool setting for the db backup > pool so I can migrate to the one Library then for the other backup disk > pool I can have it migrate to the other Library. > > I think I can have simultaneous write to two different libraries this > way. Still not sure if this would work. > >TSM Library setup > > TSM SERVER LTO_LIB LIBRARY > > TSM SERVER RMT1 DRIVE LTO_LIB > TSM SERVER RMT2 DRIVE LTO_LIB > TSM SERVER RMT3 DRIVE LTO_LIB > TSM SERVER RMT4 DRIVE LTO_LIB > > TSM SERVER RMT_LTO LIBRARY > > TSM SERVER RMT5 DRIVE RMT_LTO > TSM SERVER RMT6 DRIVE RMT_LTO > TSM SERVER RMT7 DRIVE RMT_LTO > TSM SERVER RMT8 DRIVE RMT_LTO > > Any other ideas comments are welcome! > > Thanks > > "Johnson, Milton" wrote: > > > It is the basic philosophy of TSM to have only one copy of a file in a > > > PRIMARY STORAGE POOL. With TSM 5.x you can simultaneously write to a > > PRIMARY STORAGE POOL and COPY STORAGE POOL (see HELP DEFINE STGPOOL). > > > > H. Milton Johnson > > > > > > -Original Message- > > From: ADSM: Dist Stor Manager [mailto:[EMAIL PROTECTED] On Behalf > > Of Timothy Hughes > > Sent: Thursday, September 16, 2004 10:28 AM > > To: [EMAIL PROTECTED] > > Subject: Duel tape write to LTO's > > > > Hello all, > > > > We are backing our server from disk to tape on a LTO 3584 Library > > drives > > 4 drives (1-4) > > > > We just added a 2nd library LTO (same 3584) Is it possible to backup > > to the 2nd library LTO simultaneously. > > > > In other words were backing up to tape 1 on Lib_LTO and would like to > > backup to Tape 2 on the new library at the same time. We want Duel > > tape backup and don't want to use copy groups if we don't have to. > > > > 3584 Lib_LTO A has 4 drives (1-4) > > 3584 Rmt_LTO B has 4 drives (5-8) > > > > TSM version 5.2.3.1 > > AIX 5.2 > > > > Thanks in advance for any help!
Re: D2D on AIX
Yes, try again. If it works it is a bug (don't tell IBM)! If data is on a DISK device class and Tape (or file device class) you can have 1 session from disk and other sessions from tape. -Original Message- From: Stapleton, Mark [mailto:[EMAIL PROTECTED] Sent: Tuesday, September 21, 2004 9:51 AM To: [EMAIL PROTECTED] Subject: Re: D2D on AIX From: ADSM: Dist Stor Manager [mailto:[EMAIL PROTECTED] On Behalf Of Rushforth, Tim >And with DISK device class there is no multi-session restore. Are you sure? I seem to recall using RESOURCEUTILIZATION to run a multi-threaded restore or two from DISKPOOL. -- Mark Stapleton ([EMAIL PROTECTED]) Berbee Information Networks Office 262.521.5627
OT: Flushing 3494 ATL Queue
This is slightly OT but came up through TSM on z/OS and hoping some 3494 gurus are listening. Started noticing lots of tape failures/abends (S613-1C) on TSM 5.1.8.x server on z/OS. After doing some digging and noticing the problem was almost always occuring on 1-drive, came upon this weird anomoly on the 3494ATL. When you check the queues, this drive is flagged as having a CLEAN "IN PROGRESS". Eventhough the tape drive (3590B) is empty, we can't seem to convince the ATL that the clean has finished. Any tape mount that gets assigned to this drive (via z/OS) eventually fails since the ATL never completes the mount process. We have reset and power-recycled the drive numerous times. Rebooted the A60 controller. Ejected and re-entered the cleaning tape it thinks it is using. Forced it to rescan/inventory the frame with the drive in it. Still no luck. Even the CE has no suggestions/ideas other than to shut down the whole library, which we are planning to do tonight, unless someone comes up with another idea ! Any thoughts ? Suggestions ? Seen this before ?
Re: Duel tape write to LTO's
It depends upon where you define the copypool to reside. If it is contained in the "2nd library" then yes. Has anyone out there in TSM land actually used this feature? What happens to the back-up when one of the tape volumes fills up? Does it go into a media wait state until the next volume is mounted? What happens if there isn't a volume available in the copypool? Any other gotcha's? H. Milton Johnson -Original Message- From: ADSM: Dist Stor Manager [mailto:[EMAIL PROTECTED] On Behalf Of Timothy Hughes Sent: Tuesday, September 21, 2004 10:05 AM To: [EMAIL PROTECTED] Subject: Re: Duel tape write to LTO's Hi Milton, When TSM writes simultaneously to the copypool would this be on the 2nd Library for duel tape backup? "Johnson, Milton" wrote: > You should be able to create a PRIMARY STGPOOL named TAPEPOOL and a > COPY STGPOOL named COPYPOOL with both of them having a sequential > access > (tape) DEVICE CLASS such as DLT or LTO. Both stgpools can be in the > same library. On the stgpool TAPEPOOL definition you set the > COPYSTGPOOLS parameter to COPYPOOL. Then when your client backs up to > TAPEPOOL, TSM will simultaneously write to COPYPOOL. Of course having > an adequate number of tape drives is required. > > H. Milton Johnson > > > -Original Message- > From: ADSM: Dist Stor Manager [mailto:[EMAIL PROTECTED] On Behalf > Of Timothy Hughes > Sent: Friday, September 17, 2004 8:01 AM > To: [EMAIL PROTECTED] > Subject: Re: Duel tape write to LTO's > > Hello, > > I was told that this could work If I have 2 backup disk pools. > > Like I have backup diskpool, then I can have like say a DB2 backup > diskpool then I can have the next storage pool setting for the db > backup pool so I can migrate to the one Library then for the other > backup disk pool I can have it migrate to the other Library. > > I think I can have simultaneous write to two different libraries this > way. Still not sure if this would work. > >TSM Library setup > > TSM SERVER LTO_LIB LIBRARY > > TSM SERVER RMT1 DRIVE LTO_LIB > TSM SERVER RMT2 DRIVE LTO_LIB > TSM SERVER RMT3 DRIVE LTO_LIB > TSM SERVER RMT4 DRIVE LTO_LIB > > TSM SERVER RMT_LTO LIBRARY > > TSM SERVER RMT5 DRIVE RMT_LTO > TSM SERVER RMT6 DRIVE RMT_LTO > TSM SERVER RMT7 DRIVE RMT_LTO > TSM SERVER RMT8 DRIVE RMT_LTO > > Any other ideas comments are welcome! > > Thanks > > "Johnson, Milton" wrote: > > > It is the basic philosophy of TSM to have only one copy of a file in > > a > > > PRIMARY STORAGE POOL. With TSM 5.x you can simultaneously write to a > > PRIMARY STORAGE POOL and COPY STORAGE POOL (see HELP DEFINE STGPOOL). > > > > H. Milton Johnson > > > > > > -Original Message- > > From: ADSM: Dist Stor Manager [mailto:[EMAIL PROTECTED] On > > Behalf Of Timothy Hughes > > Sent: Thursday, September 16, 2004 10:28 AM > > To: [EMAIL PROTECTED] > > Subject: Duel tape write to LTO's > > > > Hello all, > > > > We are backing our server from disk to tape on a LTO 3584 Library > > drives > > 4 drives (1-4) > > > > We just added a 2nd library LTO (same 3584) Is it possible to backup > > to the 2nd library LTO simultaneously. > > > > In other words were backing up to tape 1 on Lib_LTO and would like > > to backup to Tape 2 on the new library at the same time. We want > > Duel tape backup and don't want to use copy groups if we don't have to. > > > > 3584 Lib_LTO A has 4 drives (1-4) > > 3584 Rmt_LTO B has 4 drives (5-8) > > > > TSM version 5.2.3.1 > > AIX 5.2 > > > > Thanks in advance for any help!
Re: AIX/TSM Paging Space and Memory Settings
Hi, Since nobody else went in for the big explanation I thought I might. AIX memory tuning can help TSM performance significantly. We run on RAW LVs and set the AIX box to not page at all. >From our experience, the key seems to be the paging of the bufferpool. TSM keeps this cache of DB pages in memory to speed things up, if they end up put back on disk (usually slower system disks) then the advantage is lost, even if the cache hit% is high, the performance will be low as the system disk and the paging space on it becomes the bottleneck. My understanding is that AIX memory gets divded up into two types: computational (application code/storage) and non-computational/file (basically jfs file caching). The minperm/maxperm settings determine which of these gets paged out when AIX gets short on memory. When computational pages get paged out they go to the paging space, when non-computational pages get paged out they go back to their backing disk storage, where they came from. To stop the system paging out the TSM bufferpool, we found the right balance for our systems. This involved working out how much various applications used and setting the minperm/maxperm to match. For example, a theoretical system with 2GB may use memory in the following way: 512MB - AIX + local apps (about right on our hosts) 512MB - TSM bufferpool 512MB - TSM code etc ???MB - AIX file caching The aim would then be to tune the system so that it always had at least 1.5GB of computational memory available. That way TSM should never get paged out. To do this on this system would require setting the maxperm/minperm combination to something like 20/10, meaning 80% (5% "breathing" space) of memory should remain computational (AIX, TSM, the TSM bufferpool etc) and a max of 20% for AIX's file caching. With these settings when AIX gets a little short on memory it should eat into the file cache and not the TSM bufferpool. If you don't use RAW LVs for your LOG/DB/DISK STGPOOLS then you have an added factor to think about. The JFS file caching (non-computational) may actually be helping you, it is another layer of caching being utilised before the data hits the physical media. Careful consideration needs to be given to how much room to give to this (especially if it helps with STGPOOL which doesn't get its own bufferpool). For better explanations and a better overview on AIX performance: AIX 5L Performance Tools Handbook http://www.redbooks.ibm.com/abstracts/sg246039.html Sorry for the long post, Pete -- Peter Jones Unix Systems Programmer (HFS) Oxford University Computing Services
Re: AIX/TSM Paging Space and Memory Settings
Did you also try with -c Flag too along with -P and -p Flag!!!topas will show its effect too. -Original Message- From: ADSM: Dist Stor Manager [mailto:[EMAIL PROTECTED] On Behalf Of Peter Jones Sent: Tuesday, September 21, 2004 10:52 AM To: [EMAIL PROTECTED] Subject: Re: AIX/TSM Paging Space and Memory Settings Hi, Since nobody else went in for the big explanation I thought I might. AIX memory tuning can help TSM performance significantly. We run on RAW LVs and set the AIX box to not page at all. >From our experience, the key seems to be the paging of the bufferpool. TSM keeps this cache of DB pages in memory to speed things up, if they end up put back on disk (usually slower system disks) then the advantage is lost, even if the cache hit% is high, the performance will be low as the system disk and the paging space on it becomes the bottleneck. My understanding is that AIX memory gets divded up into two types: computational (application code/storage) and non-computational/file (basically jfs file caching). The minperm/maxperm settings determine which of these gets paged out when AIX gets short on memory. When computational pages get paged out they go to the paging space, when non-computational pages get paged out they go back to their backing disk storage, where they came from. To stop the system paging out the TSM bufferpool, we found the right balance for our systems. This involved working out how much various applications used and setting the minperm/maxperm to match. For example, a theoretical system with 2GB may use memory in the following way: 512MB - AIX + local apps (about right on our hosts) 512MB - TSM bufferpool 512MB - TSM code etc ???MB - AIX file caching The aim would then be to tune the system so that it always had at least 1.5GB of computational memory available. That way TSM should never get paged out. To do this on this system would require setting the maxperm/minperm combination to something like 20/10, meaning 80% (5% "breathing" space) of memory should remain computational (AIX, TSM, the TSM bufferpool etc) and a max of 20% for AIX's file caching. With these settings when AIX gets a little short on memory it should eat into the file cache and not the TSM bufferpool. If you don't use RAW LVs for your LOG/DB/DISK STGPOOLS then you have an added factor to think about. The JFS file caching (non-computational) may actually be helping you, it is another layer of caching being utilised before the data hits the physical media. Careful consideration needs to be given to how much room to give to this (especially if it helps with STGPOOL which doesn't get its own bufferpool). For better explanations and a better overview on AIX performance: AIX 5L Performance Tools Handbook http://www.redbooks.ibm.com/abstracts/sg246039.html Sorry for the long post, Pete -- Peter Jones Unix Systems Programmer (HFS) Oxford University Computing Services
Re: D2D on AIX
What about combining both worlds...have the DISK storage pool for your daily backups to get the multi-session backups and faster backups, then migrate to a FILE storage pool for retention. Now you'll get the multi-session restore, less overhead than the large DISK pool, but still have to do reclamation. Just substitute FILE for the onsite TAPE. Bill Boyer DSS, Inc. -Original Message- From: ADSM: Dist Stor Manager [mailto:[EMAIL PROTECTED] Behalf Of Rushforth, Tim Sent: Tuesday, September 21, 2004 11:16 AM To: [EMAIL PROTECTED] Subject: Re: D2D on AIX Yes, try again. If it works it is a bug (don't tell IBM)! If data is on a DISK device class and Tape (or file device class) you can have 1 session from disk and other sessions from tape. -Original Message- From: Stapleton, Mark [mailto:[EMAIL PROTECTED] Sent: Tuesday, September 21, 2004 9:51 AM To: [EMAIL PROTECTED] Subject: Re: D2D on AIX From: ADSM: Dist Stor Manager [mailto:[EMAIL PROTECTED] On Behalf Of Rushforth, Tim >And with DISK device class there is no multi-session restore. Are you sure? I seem to recall using RESOURCEUTILIZATION to run a multi-threaded restore or two from DISKPOOL. -- Mark Stapleton ([EMAIL PROTECTED]) Berbee Information Networks Office 262.521.5627
Re: D2D on AIX
Eric, What is the recommended volume size. I have seen someone mentioned 5G, but then the number of volumes will explode from about 800 (current # of 3590 primary tapes) to thousands. How about keeping the staging space so clients backup to staging then migrate to FILE volumes. Then every volume will be filled up. Eliza > > Hi Eliza! > You do want several smaller files, rather than a few very large files > because each client session will allocate a volume. File volumes cannot be > used concurrently by more than one session. > Kindest regards, > Eric van Loon > KLM Royal Dutch Airlines > > > -Original Message- > From: Eliza Lau [mailto:[EMAIL PROTECTED] > Sent: Monday, September 20, 2004 19:11 > To: [EMAIL PROTECTED] > Subject: D2D on AIX > > > Our 3494 with 3590K tapes in 3 frames is getting full. Instead of adding > another frame or upgrading to 3590H or 3592 tapes we are looking into > setting > up a bunch of cheap ATA disks as primary storage. > > The FILE devclass defines a directory as its destination and JFS2 > has a max file system size of 1TB. Does it mean the largest stgpool > I can define is 1TB? > > My Exchange stgpool alone has 8TB of data. Do I have to split it up > into 8 pieces? > > server: TSM 5.2.2.5 on AIX 5.2 > database 90GB at 70% > Total backup data - 22TB > > Eliza Lau > Virginia Tech Computing Center > [EMAIL PROTECTED] > > > ** > For information, services and offers, please visit our web site: http://www.klm.com. > This e-mail and any attachment may contain confidential and privileged material > intended for the addressee only. If you are not the addressee, you are notified that > no part of the e-mail or any attachment may be disclosed, copied or distributed, and > that any other action related to this e-mail or attachment is strictly prohibited, > and may be unlawful. If you have received this e-mail by error, please notify the > sender immediately by return e-mail, and delete this message. Koninklijke Luchtvaart > Maatschappij NV (KLM), its subsidiaries and/or its employees shall not be liable for > the incorrect or incomplete transmission of this e-mail or any attachments, nor > responsible for any delay in receipt. > ** >
Re: OT: Flushing 3494 ATL Queue
On Sep 21, 2004, at 11:16 AM, Zoltan Forray/AC/VCU wrote: ...Eventhough the tape drive (3590B) is empty, we can't seem to convince the ATL that the clean has finished. Any tape mount that gets assigned to this drive (via z/OS) eventually fails since the ATL never completes the mount process. ... I'm impressed with your problem pursuit measures, Zoltan - you have tried quite a lot. If your CE is stumped... I presume that the drive's panel indicates a neutral drive state, rather than any indication of cleaning. (I would not want to see a situation where the OS was for some reason persistently forcing the drive to cleaning mode.) Whereas the drive has been reset and power-cycled, it ostensibly doesn't think that it is in Cleaning mode; but the library still does. This might indicate a problem in communication between library and drive, which occurs over the ARTIC manifold of RS-422 cabling...which might be as simple as a loose connection, or may be a fault in some of the electronics. The CE should be able to watch the 3494's internal logs as you perform actions like reset and power cycling on the drive, to see if communication is happening - particularly comparing its behavior with that of the other drives. If the communication is all happening correctly, then something may be awry in the library database. I would try to jog state by making the drive Unavailable then Available at the 3494 operator panel. An experienced CE should be able to track this down. what immediately comes to mind, Richard Sims
Re: OT: Flushing 3494 ATL Queue
Thanks for the suggestions. The library won't let us take the drive offline. I forgot to mention that we tried that, as well. It simply says that something is in progress and won't let us disable the drive/controller/frame ! It is very persistant about letting this process complete ! The CE has run complete diagnostics on the drive. As usually, the diagnostics showed nothing wrong (I always thought that running diagnostics simply proves that the drive can run the diagnostics...;--))) We even ran a cleaning cartridge through it, numerous times, including the one it thinks it is using ! We tried just pressing the RESET button on the A60. No difference. Then we completely powered off the A60. Still nothing. Even tried unplugging the connectors on the A60 that go to the RTIC (which the controller, live) to see if an interrupt of that type would cause the LM to rethink its position/status of the drive ! Still nothing. This is why we are going to just shut down the LM Richard Sims <[EMAIL PROTECTED]> Sent by: "ADSM: Dist Stor Manager" <[EMAIL PROTECTED]> 09/21/2004 12:04 PM Please respond to "ADSM: Dist Stor Manager" <[EMAIL PROTECTED]> To [EMAIL PROTECTED] cc Subject Re: OT: Flushing 3494 ATL Queue On Sep 21, 2004, at 11:16 AM, Zoltan Forray/AC/VCU wrote: > ...Eventhough the tape drive (3590B) is empty, we can't seem to > convince the > ATL that the clean has finished. Any tape mount that gets assigned to > this drive (via z/OS) eventually fails since the ATL never completes > the > mount process. ... I'm impressed with your problem pursuit measures, Zoltan - you have tried quite a lot. If your CE is stumped... I presume that the drive's panel indicates a neutral drive state, rather than any indication of cleaning. (I would not want to see a situation where the OS was for some reason persistently forcing the drive to cleaning mode.) Whereas the drive has been reset and power-cycled, it ostensibly doesn't think that it is in Cleaning mode; but the library still does. This might indicate a problem in communication between library and drive, which occurs over the ARTIC manifold of RS-422 cabling...which might be as simple as a loose connection, or may be a fault in some of the electronics. The CE should be able to watch the 3494's internal logs as you perform actions like reset and power cycling on the drive, to see if communication is happening - particularly comparing its behavior with that of the other drives. If the communication is all happening correctly, then something may be awry in the library database. I would try to jog state by making the drive Unavailable then Available at the 3494 operator panel. An experienced CE should be able to track this down. what immediately comes to mind, Richard Sims
Re: D2D on AIX
Yes that is basically what we are doing. We are doing it more for fault tolerance. Our file storage pool is on a different disk subsystem. If there were major problems with that we could still do the nightly backups. Note that if you went directly to a File storage pool you would get multi-session backup also - you don't need a DISK pool for that. -Original Message- From: Bill Boyer [mailto:[EMAIL PROTECTED] Sent: Tuesday, September 21, 2004 11:09 AM To: [EMAIL PROTECTED] Subject: Re: D2D on AIX What about combining both worlds...have the DISK storage pool for your daily backups to get the multi-session backups and faster backups, then migrate to a FILE storage pool for retention. Now you'll get the multi-session restore, less overhead than the large DISK pool, but still have to do reclamation. Just substitute FILE for the onsite TAPE. Bill Boyer DSS, Inc. -Original Message- From: ADSM: Dist Stor Manager [mailto:[EMAIL PROTECTED] Behalf Of Rushforth, Tim Sent: Tuesday, September 21, 2004 11:16 AM To: [EMAIL PROTECTED] Subject: Re: D2D on AIX Yes, try again. If it works it is a bug (don't tell IBM)! If data is on a DISK device class and Tape (or file device class) you can have 1 session from disk and other sessions from tape. -Original Message- From: Stapleton, Mark [mailto:[EMAIL PROTECTED] Sent: Tuesday, September 21, 2004 9:51 AM To: [EMAIL PROTECTED] Subject: Re: D2D on AIX From: ADSM: Dist Stor Manager [mailto:[EMAIL PROTECTED] On Behalf Of Rushforth, Tim >And with DISK device class there is no multi-session restore. Are you sure? I seem to recall using RESOURCEUTILIZATION to run a multi-threaded restore or two from DISKPOOL. -- Mark Stapleton ([EMAIL PROTECTED]) Berbee Information Networks Office 262.521.5627
disk storage pools and windows compressed file systems
Have any of you used disk primary storage pools which use windows compressed file systems? Comments on performance, etc? We are investigating use of a multi TB raid5 array to use as a buffer between our local primary disk pool and the tapepool. Have seen the posts regarding file vs disk device classes but what about compression? Good, bad, etc. Win 2000 sp4 with TSM server 5.2.3.2 -- Steve Bennett, (907) 465-5783 State of Alaska, Enterprise Technology Services, Technical Services Section
Re: TDP Log file backup running for a very long time
Data Protection for SQL creates a separate TSM Server session for each stripe and then waits for the SQL Server to send data to each stripe. The SQL Server determines which data goes to which stripe, and writes the data to it. >From what you have put into this append, it appears as if sending the data to the TSM Server is the bottleneck in your scenario. How many tapes are mounted for the log backup? How many TSM Server sessions are being started for the log backup? Can you examine the session statistics on the TSM Server? What is happening on each session? How much data is being sent for each session? What is the speed like when you use a lower number of stripes for the log backup? (For example, using 2 stripes.) Finding the bottleneck on the TSM Server/Storage Agent is the key to this one. Thanks, Del "ADSM: Dist Stor Manager" <[EMAIL PROTECTED]> wrote on 09/20/2004 05:39:02 PM: > TSM Client V5.3.2.1, TDP for SQL V5.2.1, TSM Storage Agent V5.2.3.2 > TSM Server V5.2.3.2. > Windows 2003 Itanium Cluster with 2 servers. > > We are doing some SQL backup testing today. We have a 2.5 TB > database that we backed up at 477 GB/hr to 10 tape drives using > LANFree (2.5 TB's in 5.1 hours). > > We started a logs backup of a 13 GB Log and it has been running for > nearly 90 minutes so far. The backup mounted a tape via LanFree > right away and has had the same tape mounted in the same tape drive > this whole time. There are no errors being reported anywhere. > > On the command window where we started the TDP backup we see > "Waiting for TSM server" with pages of periods ".". Then we see a > message showing that is wrote some data. It keeps repeating that > process. At the rate being displayed of 620.12 Kb/Sec it looks like > the Log file backup is running at around 2 GB/hr. Any ideas what > might be going wrong? > > We still have all the settings maxed: > Buffers 8 > Buffersize 8191 > SQLBuffers 0 > SQLBuffersize 4096 > Stripes 10 > > These same settings didn't seem to effect the full backup adversly? >
Re: D2D on AIX
Yes, you would still get the multi-session backup, but only to the limit of your MAXNUMMP,right? What if you're running the FILE stgpool as collocated? -Original Message- From: ADSM: Dist Stor Manager [mailto:[EMAIL PROTECTED] Behalf Of Rushforth, Tim Sent: Tuesday, September 21, 2004 12:36 PM To: [EMAIL PROTECTED] Subject: Re: D2D on AIX Yes that is basically what we are doing. We are doing it more for fault tolerance. Our file storage pool is on a different disk subsystem. If there were major problems with that we could still do the nightly backups. Note that if you went directly to a File storage pool you would get multi-session backup also - you don't need a DISK pool for that. -Original Message- From: Bill Boyer [mailto:[EMAIL PROTECTED] Sent: Tuesday, September 21, 2004 11:09 AM To: [EMAIL PROTECTED] Subject: Re: D2D on AIX What about combining both worlds...have the DISK storage pool for your daily backups to get the multi-session backups and faster backups, then migrate to a FILE storage pool for retention. Now you'll get the multi-session restore, less overhead than the large DISK pool, but still have to do reclamation. Just substitute FILE for the onsite TAPE. Bill Boyer DSS, Inc. -Original Message- From: ADSM: Dist Stor Manager [mailto:[EMAIL PROTECTED] Behalf Of Rushforth, Tim Sent: Tuesday, September 21, 2004 11:16 AM To: [EMAIL PROTECTED] Subject: Re: D2D on AIX Yes, try again. If it works it is a bug (don't tell IBM)! If data is on a DISK device class and Tape (or file device class) you can have 1 session from disk and other sessions from tape. -Original Message- From: Stapleton, Mark [mailto:[EMAIL PROTECTED] Sent: Tuesday, September 21, 2004 9:51 AM To: [EMAIL PROTECTED] Subject: Re: D2D on AIX From: ADSM: Dist Stor Manager [mailto:[EMAIL PROTECTED] On Behalf Of Rushforth, Tim >And with DISK device class there is no multi-session restore. Are you sure? I seem to recall using RESOURCEUTILIZATION to run a multi-threaded restore or two from DISKPOOL. -- Mark Stapleton ([EMAIL PROTECTED]) Berbee Information Networks Office 262.521.5627
Re: D2D on AIX
I haven't tried this setup but that makes sense. So you would ensure MAXNUMP would be equal to the number of backup sessions you wanted. (You want to update MAXNUMP for restores anyway when restoring from file device class.) I asked the presenter of the Disk Only Backup Technical Exchange about collocating file stgpools and he said it made no sense. I tend to agree. Tim Rushforth City of Winnipeg -Original Message- From: Bill Boyer [mailto:[EMAIL PROTECTED] Sent: Tuesday, September 21, 2004 12:34 PM To: [EMAIL PROTECTED] Subject: Re: D2D on AIX Yes, you would still get the multi-session backup, but only to the limit of your MAXNUMMP,right? What if you're running the FILE stgpool as collocated? -Original Message- From: ADSM: Dist Stor Manager [mailto:[EMAIL PROTECTED] Behalf Of Rushforth, Tim Sent: Tuesday, September 21, 2004 12:36 PM To: [EMAIL PROTECTED] Subject: Re: D2D on AIX Yes that is basically what we are doing. We are doing it more for fault tolerance. Our file storage pool is on a different disk subsystem. If there were major problems with that we could still do the nightly backups. Note that if you went directly to a File storage pool you would get multi-session backup also - you don't need a DISK pool for that. -Original Message- From: Bill Boyer [mailto:[EMAIL PROTECTED] Sent: Tuesday, September 21, 2004 11:09 AM To: [EMAIL PROTECTED] Subject: Re: D2D on AIX What about combining both worlds...have the DISK storage pool for your daily backups to get the multi-session backups and faster backups, then migrate to a FILE storage pool for retention. Now you'll get the multi-session restore, less overhead than the large DISK pool, but still have to do reclamation. Just substitute FILE for the onsite TAPE. Bill Boyer DSS, Inc.
Re: D2D on AIX
Eliza: At the Disk only Backups Technical Exchange, IBM recommended 2-4 GB volume size. (This was stated by the presenter, it was not written on the PDF presentation.) We started with 25 GB volumes and have now switched to 4 GB volumes. Using smaller volume sizes allows a better utilization of space and increases restore performance with multi-session restore. (Also helps eliminate contention if multiple clients are restoring from the same volume) Tim Rushforth City of Winnipeg -Original Message- From: Eliza Lau [mailto:[EMAIL PROTECTED] Sent: Tuesday, September 21, 2004 11:20 AM To: [EMAIL PROTECTED] Subject: Re: D2D on AIX Eric, What is the recommended volume size. I have seen someone mentioned 5G, but then the number of volumes will explode from about 800 (current # of 3590 primary tapes) to thousands. How about keeping the staging space so clients backup to staging then migrate to FILE volumes. Then every volume will be filled up. Eliza > > Hi Eliza! > You do want several smaller files, rather than a few very large files > because each client session will allocate a volume. File volumes cannot be > used concurrently by more than one session. > Kindest regards, > Eric van Loon > KLM Royal Dutch Airlines > > > -Original Message- > From: Eliza Lau [mailto:[EMAIL PROTECTED] > Sent: Monday, September 20, 2004 19:11 > To: [EMAIL PROTECTED] > Subject: D2D on AIX > > > Our 3494 with 3590K tapes in 3 frames is getting full. Instead of adding > another frame or upgrading to 3590H or 3592 tapes we are looking into > setting > up a bunch of cheap ATA disks as primary storage. > > The FILE devclass defines a directory as its destination and JFS2 > has a max file system size of 1TB. Does it mean the largest stgpool > I can define is 1TB? > > My Exchange stgpool alone has 8TB of data. Do I have to split it up > into 8 pieces? > > server: TSM 5.2.2.5 on AIX 5.2 > database 90GB at 70% > Total backup data - 22TB > > Eliza Lau > Virginia Tech Computing Center > [EMAIL PROTECTED] > > > ** > For information, services and offers, please visit our web site: http://www.klm.com. This e-mail and any attachment may contain confidential and privileged material intended for the addressee only. If you are not the addressee, you are notified that no part of the e-mail or any attachment may be disclosed, copied or distributed, and that any other action related to this e-mail or attachment is strictly prohibited, and may be unlawful. If you have received this e-mail by error, please notify the sender immediately by return e-mail, and delete this message. Koninklijke Luchtvaart Maatschappij NV (KLM), its subsidiaries and/or its employees shall not be liable for the incorrect or incomplete transmission of this e-mail or any attachments, nor responsible for any delay in receipt. > ** >
Re: D2D on AIX
Now we get into religion. IBM did offer a figure of ~5GB during the webinar, but there are a lot of factors that would affect this such as: REUSE DELAY: you want to be able to use those TSM DB backups RECLAMATION THRESHOLD: A lower threshold should lead to more efficient usage of volumes except that it causes more frequent tape reclamation leading to more pending volumes causing wasted space. Of course the exact opposite is true regarding higher reclamation thresholds. What yin yang is right for you? Experiment and find out. AVG SIZE OF STORED OBJECTS? EXPIRATION RATE OF STORED OBJECTS? I'm sure others will bring up other factors. How many volumes are too many? If TSM is keeping track of the volumes and you are not handling the physical volumes (i.e. loading/unloading tapes), is 4,000 too many? If so, why? H. Milton Johnson -Original Message- From: ADSM: Dist Stor Manager [mailto:[EMAIL PROTECTED] On Behalf Of Eliza Lau Sent: Tuesday, September 21, 2004 11:20 AM To: [EMAIL PROTECTED] Subject: Re: D2D on AIX Eric, What is the recommended volume size. I have seen someone mentioned 5G, but then the number of volumes will explode from about 800 (current # of 3590 primary tapes) to thousands. How about keeping the staging space so clients backup to staging then migrate to FILE volumes. Then every volume will be filled up. Eliza > > Hi Eliza! > You do want several smaller files, rather than a few very large files > because each client session will allocate a volume. File volumes > cannot be used concurrently by more than one session. > Kindest regards, > Eric van Loon > KLM Royal Dutch Airlines > > > -Original Message- > From: Eliza Lau [mailto:[EMAIL PROTECTED] > Sent: Monday, September 20, 2004 19:11 > To: [EMAIL PROTECTED] > Subject: D2D on AIX > > > Our 3494 with 3590K tapes in 3 frames is getting full. Instead of > adding another frame or upgrading to 3590H or 3592 tapes we are > looking into setting up a bunch of cheap ATA disks as primary storage. > > The FILE devclass defines a directory as its destination and JFS2 has > a max file system size of 1TB. Does it mean the largest stgpool I can > define is 1TB? > > My Exchange stgpool alone has 8TB of data. Do I have to split it up > into 8 pieces? > > server: TSM 5.2.2.5 on AIX 5.2 > database 90GB at 70% > Total backup data - 22TB > > Eliza Lau > Virginia Tech Computing Center > [EMAIL PROTECTED] > > > ** > For information, services and offers, please visit our web site: http://www.klm.com. This e-mail and any attachment may contain confidential and privileged material intended for the addressee only. If you are not the addressee, you are notified that no part of the e-mail or any attachment may be disclosed, copied or distributed, and that any other action related to this e-mail or attachment is strictly prohibited, and may be unlawful. If you have received this e-mail by error, please notify the sender immediately by return e-mail, and delete this message. Koninklijke Luchtvaart Maatschappij NV (KLM), its subsidiaries and/or its employees shall not be liable for the incorrect or incomplete transmission of this e-mail or any attachments, nor responsible for any delay in receipt. > ** >
Re: D2D on AIX
What do use for a reuse delay? How many pending volumes do you average? H. Milton Johnson -Original Message- From: ADSM: Dist Stor Manager [mailto:[EMAIL PROTECTED] On Behalf Of Rushforth, Tim Sent: Tuesday, September 21, 2004 1:00 PM To: [EMAIL PROTECTED] Subject: Re: D2D on AIX Eliza: At the Disk only Backups Technical Exchange, IBM recommended 2-4 GB volume size. (This was stated by the presenter, it was not written on the PDF presentation.) We started with 25 GB volumes and have now switched to 4 GB volumes. Using smaller volume sizes allows a better utilization of space and increases restore performance with multi-session restore. (Also helps eliminate contention if multiple clients are restoring from the same volume) Tim Rushforth City of Winnipeg -Original Message- From: Eliza Lau [mailto:[EMAIL PROTECTED] Sent: Tuesday, September 21, 2004 11:20 AM To: [EMAIL PROTECTED] Subject: Re: D2D on AIX Eric, What is the recommended volume size. I have seen someone mentioned 5G, but then the number of volumes will explode from about 800 (current # of 3590 primary tapes) to thousands. How about keeping the staging space so clients backup to staging then migrate to FILE volumes. Then every volume will be filled up. Eliza > > Hi Eliza! > You do want several smaller files, rather than a few very large files > because each client session will allocate a volume. File volumes > cannot be used concurrently by more than one session. > Kindest regards, > Eric van Loon > KLM Royal Dutch Airlines > > > -Original Message- > From: Eliza Lau [mailto:[EMAIL PROTECTED] > Sent: Monday, September 20, 2004 19:11 > To: [EMAIL PROTECTED] > Subject: D2D on AIX > > > Our 3494 with 3590K tapes in 3 frames is getting full. Instead of > adding another frame or upgrading to 3590H or 3592 tapes we are > looking into setting up a bunch of cheap ATA disks as primary storage. > > The FILE devclass defines a directory as its destination and JFS2 has > a max file system size of 1TB. Does it mean the largest stgpool I can > define is 1TB? > > My Exchange stgpool alone has 8TB of data. Do I have to split it up > into 8 pieces? > > server: TSM 5.2.2.5 on AIX 5.2 > database 90GB at 70% > Total backup data - 22TB > > Eliza Lau > Virginia Tech Computing Center > [EMAIL PROTECTED] > > > ** > For information, services and offers, please visit our web site: http://www.klm.com. This e-mail and any attachment may contain confidential and privileged material intended for the addressee only. If you are not the addressee, you are notified that no part of the e-mail or any attachment may be disclosed, copied or distributed, and that any other action related to this e-mail or attachment is strictly prohibited, and may be unlawful. If you have received this e-mail by error, please notify the sender immediately by return e-mail, and delete this message. Koninklijke Luchtvaart Maatschappij NV (KLM), its subsidiaries and/or its employees shall not be liable for the incorrect or incomplete transmission of this e-mail or any attachments, nor responsible for any delay in receipt. > ** >
Re: D2D on AIX
We use 5 days for reuse delay. I did a quick comparison using 25GB and 4GB volumes on our pilot with the following results: Disk Volumes - 25 GB Volumes Stored Data - 236 GB # of Disk Vols - 14 (including 2 pending volumes) Total allocation - 14 * 25GB = 350 GB 67% Utilization Disk Volumes - 4 GB Volumes (TSM2) Stored Data - 333 GB # of Disk Vols - 100 (including 12 pending volumes) Total allocation - 100 * 4GB = 400 GB 83% Utilization Reclamation was set at 25% for both of these. It would be interesting to see others peoples results. Thanks, Tim Rushforth City of Winnipeg -Original Message- From: Johnson, Milton [mailto:[EMAIL PROTECTED] Sent: Tuesday, September 21, 2004 1:04 PM To: [EMAIL PROTECTED] Subject: Re: D2D on AIX What do use for a reuse delay? How many pending volumes do you average? H. Milton Johnson -Original Message- From: ADSM: Dist Stor Manager [mailto:[EMAIL PROTECTED] On Behalf Of Rushforth, Tim Sent: Tuesday, September 21, 2004 1:00 PM To: [EMAIL PROTECTED] Subject: Re: D2D on AIX Eliza: At the Disk only Backups Technical Exchange, IBM recommended 2-4 GB volume size. (This was stated by the presenter, it was not written on the PDF presentation.) We started with 25 GB volumes and have now switched to 4 GB volumes. Using smaller volume sizes allows a better utilization of space and increases restore performance with multi-session restore. (Also helps eliminate contention if multiple clients are restoring from the same volume) Tim Rushforth City of Winnipeg -Original Message- From: Eliza Lau [mailto:[EMAIL PROTECTED] Sent: Tuesday, September 21, 2004 11:20 AM To: [EMAIL PROTECTED] Subject: Re: D2D on AIX Eric, What is the recommended volume size. I have seen someone mentioned 5G, but then the number of volumes will explode from about 800 (current # of 3590 primary tapes) to thousands. How about keeping the staging space so clients backup to staging then migrate to FILE volumes. Then every volume will be filled up. Eliza > > Hi Eliza! > You do want several smaller files, rather than a few very large files > because each client session will allocate a volume. File volumes > cannot be used concurrently by more than one session. > Kindest regards, > Eric van Loon > KLM Royal Dutch Airlines > > > -Original Message- > From: Eliza Lau [mailto:[EMAIL PROTECTED] > Sent: Monday, September 20, 2004 19:11 > To: [EMAIL PROTECTED] > Subject: D2D on AIX > > > Our 3494 with 3590K tapes in 3 frames is getting full. Instead of > adding another frame or upgrading to 3590H or 3592 tapes we are > looking into setting up a bunch of cheap ATA disks as primary storage. > > The FILE devclass defines a directory as its destination and JFS2 has > a max file system size of 1TB. Does it mean the largest stgpool I can > define is 1TB? > > My Exchange stgpool alone has 8TB of data. Do I have to split it up > into 8 pieces? > > server: TSM 5.2.2.5 on AIX 5.2 > database 90GB at 70% > Total backup data - 22TB > > Eliza Lau > Virginia Tech Computing Center > [EMAIL PROTECTED] > > > ** > For information, services and offers, please visit our web site: http://www.klm.com. This e-mail and any attachment may contain confidential and privileged material intended for the addressee only. If you are not the addressee, you are notified that no part of the e-mail or any attachment may be disclosed, copied or distributed, and that any other action related to this e-mail or attachment is strictly prohibited, and may be unlawful. If you have received this e-mail by error, please notify the sender immediately by return e-mail, and delete this message. Koninklijke Luchtvaart Maatschappij NV (KLM), its subsidiaries and/or its employees shall not be liable for the incorrect or incomplete transmission of this e-mail or any attachments, nor responsible for any delay in receipt. > ** >
Re: D2D on AIX
==> In article <[EMAIL PROTECTED]>, Eliza Lau <[EMAIL PROTECTED]> writes: > What is the recommended volume size. I have seen someone mentioned 5G, but > then the number of volumes will explode from about 800 (current # of 3590 > primary tapes) to thousands. Consider, this doesn't really cost you much. Seek time in a directory of thousands of files is still tiny compared to tape behavior. I probably wouldn't go as low as 5G, but 10G (much less than the average size of my 3590 vols) seems pretty reasonable to me. 20G is getting big, from my perspective. > How about keeping the staging space so clients backup to staging then > migrate to FILE volumes. Then every volume will be filled up. I like this, too. - Allen S. Rout
Re: Relabel of 3592
On Thu, 2004-09-16 at 19:08, Jim Sporer wrote: > Have you tried the dsmlabel command with the -overwrite option? Worked. Thanks. -- Michael Prix
Re: disk storage pools and windows compressed file systems
The 5.2.2 Performance Tuning Guide says: "NTFS file compression should not be used on disk volumes that are used by the TSM server, because of the potential for performance degradation." We use client compression so I don't think it would buy us anything. Report back if you try this out! -Original Message- From: Steve Bennett [mailto:[EMAIL PROTECTED] Sent: Tuesday, September 21, 2004 12:11 PM To: [EMAIL PROTECTED] Subject: disk storage pools and windows compressed file systems Have any of you used disk primary storage pools which use windows compressed file systems? Comments on performance, etc? We are investigating use of a multi TB raid5 array to use as a buffer between our local primary disk pool and the tapepool. Have seen the posts regarding file vs disk device classes but what about compression? Good, bad, etc. Win 2000 sp4 with TSM server 5.2.3.2 -- Steve Bennett, (907) 465-5783 State of Alaska, Enterprise Technology Services, Technical Services Section
Re: Huge system object
> For now, at the 5.1.5 client level, it might be easier to put the > following in your include/exclude list: > >exclude.systemobject frs > > This should cause systemobject backups to skip the FRS object, and avoids > the batch file method you mentioned. Thanks. That worked perfectly, and was a lot easier to implement than my proposal would have been.
Re: D2D on AIX
True. Seek time is tiny compared to tape mounts. I am just concerned that the TSM db has to keep track of thousands of volume. How much will it increase the size of the db. Ours is already 90G at 70% utilized. Eliza > > ==> In article <[EMAIL PROTECTED]>, Eliza Lau <[EMAIL PROTECTED]> writes: > > > What is the recommended volume size. I have seen someone mentioned 5G, but > > then the number of volumes will explode from about 800 (current # of 3590 > > primary tapes) to thousands. > > Consider, this doesn't really cost you much. Seek time in a directory of > thousands of files is still tiny compared to tape behavior. > > I probably wouldn't go as low as 5G, but 10G (much less than the average size > of my 3590 vols) seems pretty reasonable to me. 20G is getting big, from my > perspective. > > > > > How about keeping the staging space so clients backup to staging then > > migrate to FILE volumes. Then every volume will be filled up. > > > I like this, too. > > > - Allen S. Rout >
Re: TDP EXCHANGE
How do you perform a schedule archive for Exchange using TDP for Mail? Thank you James
Re: TDP EXCHANGE
From: ADSM: Dist Stor Manager [mailto:[EMAIL PROTECTED] On Behalf Of James Lepre >How do you perform a schedule archive for Exchange using TDP for Mail? IIRC, archives are not supported with most of the TDP products. The best workaround that comes to mind quickly is to restore a backup to a test Exchange server, shut down Exchange on that server, and use the TSM b/a client to archive the cold directory and information store. -- Mark Stapleton ([EMAIL PROTECTED]) Berbee Information Networks Office 262.521.5627
Re: TDP EXCHANGE
James, Mark is correct that there is no TSM "archive" function for Data Protection for Exchange. However, there are a few things that many customers have done: - Set up a special NODENAME (like EXCHSRV1_ARCHIVE) that will bind the backup objects to special management classes that meet your desired retention policies. Then, perform COPY (or DBCOPY) type backups using that NODENAME on a schedule that you want. - Use IBM DB2 CommonStore for Exchange Server: http://www-306.ibm.com/software/data/commonstore/exchange/index.html which supports archive to TSM Servers. Thanks, Del "ADSM: Dist Stor Manager" <[EMAIL PROTECTED]> wrote on 09/21/2004 02:57:08 PM: > How do you perform a schedule archive for Exchange using TDP for Mail? > > Thank you > > James
Re: UNIX
Hello *smers I am having a problem with backing up a certain file structure in AIX 5.2. The file structure is about seven layers deep and in the last directory layer are a set of 10 files that the schedule skips every night. I checked there are no excludes, I tried to put in includes without anything working. If I try to backup them up manually, the results are that TSM inspects the files but does not back them up. Any ideas would greatly be appreciated, and Thank you in advance Thank you James
Re: UNIX
On Tue, 21 Sep 2004, James Lepre wrote: > Hello *smers > > I am having a problem with backing up a certain file structure in AIX > 5.2. The file structure is about seven layers deep and in the last > directory layer are a set of 10 files that the schedule skips every > night. I checked there are no excludes, I tried to put in includes > without anything working. If I try to backup them up manually, the > results are that TSM inspects the files but does not back them up. Any > ideas would greatly be appreciated, and Thank you in advance How about posting your dsm control files and what files are being skipped? Maybe the timestampe is out of whack? Mike
schedule of SQL LOG backup
Hi all Is there a way to make a schedule to take the backup of the SQL LOG every 20 minutes .. I'm not able to find the way to put it in minutes ... thanks TSM SERVER - win2000 5.2.0.2 SQL 2000 - 5.2.1.0 Luc Beaudoin Administrateur Réseau / Network Administrator Hopital General Juif S.M.B.D. Tel: (514) 340-8222 ext:8254
Re: schedule of SQL LOG backup
From: ADSM: Dist Stor Manager [mailto:[EMAIL PROTECTED] On Behalf Of Luc Beaudoin >Is there a way to make a schedule to take the backup of the >SQL LOG every 20 minutes .. Now think about that for a minute; that's three log backups per hour. If you have to restore the database 20 hours after the full backup is done, that means there are 61 files you have to restore, most likely from tape. That will make for a unnecessarily long restore. Do you really need *that* fine a granularity? -- Mark Stapleton ([EMAIL PROTECTED]) Berbee Information Networks Office 262.521.5627
Re: schedule of SQL LOG backup
thanks Mark .. I thought of doing Full backup every 8 hours and LOG backup every 20 minutes ... Is there a best pratice for SQL backup ?? thanks again Luc "Stapleton, Mark" <[EMAIL PROTECTED]> Sent by: "ADSM: Dist Stor Manager" <[EMAIL PROTECTED]> 2004-09-21 04:20 PM Please respond to "ADSM: Dist Stor Manager" To: [EMAIL PROTECTED] cc: Subject:Re: schedule of SQL LOG backup From: ADSM: Dist Stor Manager [mailto:[EMAIL PROTECTED] On Behalf Of Luc Beaudoin >Is there a way to make a schedule to take the backup of the >SQL LOG every 20 minutes .. Now think about that for a minute; that's three log backups per hour. If you have to restore the database 20 hours after the full backup is done, that means there are 61 files you have to restore, most likely from tape. That will make for a unnecessarily long restore. Do you really need *that* fine a granularity? -- Mark Stapleton ([EMAIL PROTECTED]) Berbee Information Networks Office 262.521.5627
Re: schedule of SQL LOG backup
From: ADSM: Dist Stor Manager [mailto:[EMAIL PROTECTED] On Behalf Of Luc Beaudoin >I thought of doing Full backup every 8 hours and LOG backup >every 20 minutes ... >Is there a best pratice for SQL backup ?? What works best is whatever meets your business needs. Most of my customers do a full backup of databases once a day, and periodic log backups (every 4 hours, for example) throughout the day. YMMV. -- Mark Stapleton ([EMAIL PROTECTED]) Berbee Information Networks Office 262.521.5627
Re: schedule of SQL LOG backup
Hi Mark ... So with that setup ... worst case ... they will loose 4 hours of work ??? I'm working in a hospital ... so even 1 hours lost of lab result or patient appointment can be kind of hell anyway .. if there is no way of putting minutes ... I will put the minimum ... 1 hours ... thanks a lot Mark Luc "Stapleton, Mark" <[EMAIL PROTECTED]> Sent by: "ADSM: Dist Stor Manager" <[EMAIL PROTECTED]> 2004-09-21 04:28 PM Please respond to "ADSM: Dist Stor Manager" To: [EMAIL PROTECTED] cc: Subject:Re: schedule of SQL LOG backup From: ADSM: Dist Stor Manager [mailto:[EMAIL PROTECTED] On Behalf Of Luc Beaudoin >I thought of doing Full backup every 8 hours and LOG backup >every 20 minutes ... >Is there a best pratice for SQL backup ?? What works best is whatever meets your business needs. Most of my customers do a full backup of databases once a day, and periodic log backups (every 4 hours, for example) throughout the day. YMMV. -- Mark Stapleton ([EMAIL PROTECTED]) Berbee Information Networks Office 262.521.5627
Re: disk storage pools and windows compressed file systems
Only a few of our clients do compression before sending backup/archive data and since clients backup directly to the local primary disk storage pool with no NTFS compression it should not affect them. Then later as part of the daily maint cycle the local disk pool will get migrated to the compressed array. Then as the array fills the data gets migrated to tape. If we use dev disk there will be no reclamation process but if we used dev file then performance for reclamations could suffer. The only time clients would be impacted by compression is when they do restores that come from the array which very infrequent. With lots of cpu power I would think the NTFS compression wouldn't be too much of a issue. I'll post what I find should we ever get there. Rushforth, Tim wrote: The 5.2.2 Performance Tuning Guide says: "NTFS file compression should not be used on disk volumes that are used by the TSM server, because of the potential for performance degradation." We use client compression so I don't think it would buy us anything. Report back if you try this out! -Original Message- From: Steve Bennett [mailto:[EMAIL PROTECTED] Sent: Tuesday, September 21, 2004 12:11 PM To: [EMAIL PROTECTED] Subject: disk storage pools and windows compressed file systems Have any of you used disk primary storage pools which use windows compressed file systems? Comments on performance, etc? We are investigating use of a multi TB raid5 array to use as a buffer between our local primary disk pool and the tapepool. Have seen the posts regarding file vs disk device classes but what about compression? Good, bad, etc. Win 2000 sp4 with TSM server 5.2.3.2 -- Steve Bennett, (907) 465-5783 State of Alaska, Enterprise Technology Services, Technical Services Section -- Steve Bennett, (907) 465-5783 State of Alaska, Enterprise Technology Services, Technical Services Section
Upgrade time...
I could use a little help. I am at TSM Win 4.2.3.1 on my server and need to upgrade to 5.2 or so. Probably to 5.2.2.3. I have a 3583 with LTO1's, all freshly upgraded to current microcode a week or two ago. Any special gotcha's I should look for? >From what I have skimmed (I would say read, but I know I miss stuff), it looks like: Do an un-install of the TSM device driver. Do an un-install of the TSM server. Install the new versions of both. This seems to simple. Are there any conversion utilities I should use? Special commands to clean up the database? Etc? etc? Just not wanting to miss anything.
Re: Upgrade time...
Have you done the infamous CLEANUP BACKUPGROUPS ? "Coats, Jack" <[EMAIL PROTECTED]> Sent by: "ADSM: Dist Stor Manager" <[EMAIL PROTECTED]> 09/21/2004 05:54 PM Please respond to "ADSM: Dist Stor Manager" <[EMAIL PROTECTED]> To [EMAIL PROTECTED] cc Subject Upgrade time... I could use a little help. I am at TSM Win 4.2.3.1 on my server and need to upgrade to 5.2 or so. Probably to 5.2.2.3. I have a 3583 with LTO1's, all freshly upgraded to current microcode a week or two ago. Any special gotcha's I should look for? >From what I have skimmed (I would say read, but I know I miss stuff), it looks like: Do an un-install of the TSM device driver. Do an un-install of the TSM server. Install the new versions of both. This seems to simple. Are there any conversion utilities I should use? Special commands to clean up the database? Etc? etc? Just not wanting to miss anything.
Re: disk storage pools and windows compressed file systems
I have two instances of storage pool volume corruption due to file compression on Windows 2003. Required a db audit of the storage to get rid of the volumes. It seemed like a good idea to the customer . . . Bill Smoldt STORServer, Inc. -Original Message- From: ADSM: Dist Stor Manager [mailto:[EMAIL PROTECTED] On Behalf Of Rushforth, Tim Sent: Tuesday, September 21, 2004 12:39 PM To: [EMAIL PROTECTED] Subject: Re: disk storage pools and windows compressed file systems The 5.2.2 Performance Tuning Guide says: "NTFS file compression should not be used on disk volumes that are used by the TSM server, because of the potential for performance degradation." We use client compression so I don't think it would buy us anything. Report back if you try this out! -Original Message- From: Steve Bennett [mailto:[EMAIL PROTECTED] Sent: Tuesday, September 21, 2004 12:11 PM To: [EMAIL PROTECTED] Subject: disk storage pools and windows compressed file systems Have any of you used disk primary storage pools which use windows compressed file systems? Comments on performance, etc? We are investigating use of a multi TB raid5 array to use as a buffer between our local primary disk pool and the tapepool. Have seen the posts regarding file vs disk device classes but what about compression? Good, bad, etc. Win 2000 sp4 with TSM server 5.2.3.2 -- Steve Bennett, (907) 465-5783 State of Alaska, Enterprise Technology Services, Technical Services Section
Re: D2D on AIX
How does TSM access the data on file volumes? Does it keep an offset of the start of every file or aggregate? If it does, then yes we could skip to the start of each file or aggregate. If it does not, then we need to read through the volume to find the file we are going to restore. Where we have a large number of concurrent restores happening, this could cause performance issues on the array. Now TSM has some smarts on later technology tape drives that have block addressability and on-cartridge memory and can find a spot on the tape quickly, but does this translate to file volumes? Regards Steve. >>> [EMAIL PROTECTED] 22/09/2004 4:49:55 >>> True. Seek time is tiny compared to tape mounts. I am just concerned that the TSM db has to keep track of thousands of volume. How much will it increase the size of the db. Ours is already 90G at 70% utilized. Eliza > > ==> In article <[EMAIL PROTECTED]>, Eliza Lau <[EMAIL PROTECTED]> writes: > > > What is the recommended volume size. I have seen someone mentioned 5G, but > > then the number of volumes will explode from about 800 (current # of 3590 > > primary tapes) to thousands. > > Consider, this doesn't really cost you much. Seek time in a directory of > thousands of files is still tiny compared to tape behavior. > > I probably wouldn't go as low as 5G, but 10G (much less than the average size > of my 3590 vols) seems pretty reasonable to me. 20G is getting big, from my > perspective. > > > > > How about keeping the staging space so clients backup to staging then > > migrate to FILE volumes. Then every volume will be filled up. > > > I like this, too. > > > - Allen S. Rout > *** This email, including any attachments sent with it, is confidential and for the sole use of the intended recipient(s). This confidentiality is not waived or lost, if you receive it and you are not the intended recipient(s), or if it is transmitted/received in error. Any unauthorised use, alteration, disclosure, distribution or review of this email is prohibited. It may be subject to a statutory duty of confidentiality if it relates to health service matters. If you are not the intended recipient(s), or if you have received this email in error, you are asked to immediately notify the sender by telephone or by return email. You should also delete this email and destroy any hard copies produced. ***
Re: AIX/TSM Paging Space and Memory Settings
I am also curious to see what your settings are for AIXDIRECTIO and AIXSYNCIO. For JFS2 only, we recommend that the AIXDIRECTIO setting be YES and AIXSYNCIO be set to no. In general, we see that if the direct i/o is set to yes, then JFS2 is roughly equal to RLV (raw logical volumes) in terms of performance. These settings apply to JFS2 ONLY. At 09:36 AM 9/21/2004 -0500, you wrote: Well, interesting results from the changes we made to our vmtune settings yesterday. We have one TSM instance running with JFS2 for database, logs and storage pools, and another on all raw. We run two database backups daily, a full and a snapshot. Last evening's snapshot was 5,000,000 pages an hour faster on the JFS2 TSM instance than any previous DB backup we've had. It was about the same speed as a backup on the server using raw volumes. Of course, one night doesn't mean a whole lot in the big scheme of things. But if this continues, I'll be impressed. Curtis [EMAIL PROTECTED] Dave Canan TSM Performance IBM Advanced Technical Support [EMAIL PROTECTED]
Re: UNIX
>I am having a problem with backing up a certain file structure in AIX >5.2. The file structure is about seven layers deep and in the last >directory layer are a set of 10 files that the schedule skips every >night. I checked there are no excludes, I tried to put in includes >without anything working. If I try to backup them up manually, the >results are that TSM inspects the files but does not back them up. Any >ideas would greatly be appreciated, and Thank you in advance When posting client problems such as this, always tell us whether there were any allied messages or indications in the dsmerror.log . And it would help us to know the details of the backup attempt: Incremental or Selective, whether any special options in effect, how the file system objects were specified - and whether the backup was attempted as root. The type of file system and type of file objects (ordinary, Special, symlink, directory, etc.) would be useful to know. Richard Sims
Re: AIX Paging Space Utilization
Curtis, I'm guessing that the restriction on Maxclient might do the trick. Shortly after updating our TSM and Oracle servers to AIX 5.2, I started using JFS2 for some of the file systems. In a very short time, I saw page file usage in the 60-70% range after YEARS of running less than 10%. I eventually tracked it down to the MaxClient% which sets the percentage of memory used for file pages and caching of JFS2. I had been setting MaxPerm% but that only applies to JFS(1). After restricting that parameter to about 30%, pagefile usage dropped to its normal range of about 5%. It is currently 4.6% as I write this. Tab Trepagnier TSM Administrator Laitram, L.L.C. Curtis Stewart <[EMAIL PROTECTED]> Sent by: "ADSM: Dist Stor Manager" <[EMAIL PROTECTED]> 09/20/2004 03:50 PM Please respond to "ADSM: Dist Stor Manager" To: [EMAIL PROTECTED] cc: Subject:Re: AIX Paging Space Utilization We changed our settings to the following. I'll let you all know if this fixes the problem in our case. Thanks for all the advice. So far so good. After we made the changes, I stopped both instances of TSM, and the paging space was reclaimed in about 5 minutes. It's sitting at 74.3% free now, after a couple of hours. I'll give it a couple of days and report for anyone who's interested. Oh, and Richard still didn't get the size of the paging space increased. The wet noodle flogging didn't work completely, but it got the VM Tuning done! 1) Maxperm = 50% 2) Minperm = 10% 3) Maxclient = 20% 4) Strict_Maxperm = 1 Curtis [EMAIL PROTECTED]
Re: Upgrade time...
Yep, Thanks for asking :) -Original Message- From: Zoltan Forray/AC/VCU [mailto:[EMAIL PROTECTED] Sent: Tuesday, September 21, 2004 4:57 PM To: [EMAIL PROTECTED] Subject: Re: Upgrade time... Have you done the infamous CLEANUP BACKUPGROUPS ? "Coats, Jack" <[EMAIL PROTECTED]> Sent by: "ADSM: Dist Stor Manager" <[EMAIL PROTECTED]> 09/21/2004 05:54 PM Please respond to "ADSM: Dist Stor Manager" <[EMAIL PROTECTED]> To [EMAIL PROTECTED] cc Subject Upgrade time... I could use a little help. I am at TSM Win 4.2.3.1 on my server and need to upgrade to 5.2 or so. Probably to 5.2.2.3. I have a 3583 with LTO1's, all freshly upgraded to current microcode a week or two ago. Any special gotcha's I should look for? >From what I have skimmed (I would say read, but I know I miss stuff), it looks like: Do an un-install of the TSM device driver. Do an un-install of the TSM server. Install the new versions of both. This seems to simple. Are there any conversion utilities I should use? Special commands to clean up the database? Etc? etc? Just not wanting to miss anything.
Re: D2D on AIX
Eliza, We are using 25 GB volumes right now without any issues but we are still collocateing by node. We are evaluting the savings of using smaller volumes when we move to noncollocated storage pools. I agree that it doesn't make sense to collocate file device class pools but managment wouldn't let us change that at first. If you are talking about a Windows server then you need to think about file handles and their effect on the 256 MB of nonpagged memory. Also, the size of your MFT should be considered. Using more smaller files will have an effect on these things. I think as more and more of us implement these solutions we will have more collective knowledge from which to guide these decisions. Again, 25 GB volumes have been working well for us for nearly 8 months. We set our storage pools to relaim=30. We run expiration at 6:00 AM which starts a reclamation process soon after. We never have reclamation processes running past 8:30 AM. We have run many restores and have been extremely pleased with our speeds, much faster than tape for servers with many small files. "Rushforth, Tim" <[EMAIL PROTECTED]> wrote: Eliza: At the Disk only Backups Technical Exchange, IBM recommended 2-4 GB volume size. (This was stated by the presenter, it was not written on the PDF presentation.) We started with 25 GB volumes and have now switched to 4 GB volumes. Using smaller volume sizes allows a better utilization of space and increases restore performance with multi-session restore. (Also helps eliminate contention if multiple clients are restoring from the same volume) Tim Rushforth City of Winnipeg -Original Message- From: Eliza Lau [mailto:[EMAIL PROTECTED] Sent: Tuesday, September 21, 2004 11:20 AM To: [EMAIL PROTECTED] Subject: Re: D2D on AIX Eric, What is the recommended volume size. I have seen someone mentioned 5G, but then the number of volumes will explode from about 800 (current # of 3590 primary tapes) to thousands. How about keeping the staging space so clients backup to staging then migrate to FILE volumes. Then every volume will be filled up. Eliza > > Hi Eliza! > You do want several smaller files, rather than a few very large files > because each client session will allocate a volume. File volumes cannot be > used concurrently by more than one session. > Kindest regards, > Eric van Loon > KLM Royal Dutch Airlines > > > -Original Message- > From: Eliza Lau [mailto:[EMAIL PROTECTED] > Sent: Monday, September 20, 2004 19:11 > To: [EMAIL PROTECTED] > Subject: D2D on AIX > > > Our 3494 with 3590K tapes in 3 frames is getting full. Instead of adding > another frame or upgrading to 3590H or 3592 tapes we are looking into > setting > up a bunch of cheap ATA disks as primary storage. > > The FILE devclass defines a directory as its destination and JFS2 > has a max file system size of 1TB. Does it mean the largest stgpool > I can define is 1TB? > > My Exchange stgpool alone has 8TB of data. Do I have to split it up > into 8 pieces? > > server: TSM 5.2.2.5 on AIX 5.2 > database 90GB at 70% > Total backup data - 22TB > > Eliza Lau > Virginia Tech Computing Center > [EMAIL PROTECTED] > > > ** > For information, services and offers, please visit our web site: http://www.klm.com. This e-mail and any attachment may contain confidential and privileged material intended for the addressee only. If you are not the addressee, you are notified that no part of the e-mail or any attachment may be disclosed, copied or distributed, and that any other action related to this e-mail or attachment is strictly prohibited, and may be unlawful. If you have received this e-mail by error, please notify the sender immediately by return e-mail, and delete this message. Koninklijke Luchtvaart Maatschappij NV (KLM), its subsidiaries and/or its employees shall not be liable for the incorrect or incomplete transmission of this e-mail or any attachments, nor responsible for any delay in receipt. > ** > - Do you Yahoo!? vote.yahoo.com - Register online to vote today!
Re: TDP Log file backup running for a very long time
Thanks Del. Last night I changed everything to the defaults and used one stripe. The backup ran in 2 or 3 minutes. I plan to run more tests next week experiementing with different settings for stripe, buffers and such to see what yeilds the best results. Also, the restore of the 2.5 TB DB I hit the "commtimeout" issue. A quick search of the list archives and I found many notes from you on the subject so changing commtimeout to 1 solved that problem as well. Thanks for that as well. Del Hoobler <[EMAIL PROTECTED]> wrote: Data Protection for SQL creates a separate TSM Server session for each stripe and then waits for the SQL Server to send data to each stripe. The SQL Server determines which data goes to which stripe, and writes the data to it. >From what you have put into this append, it appears as if sending the data to the TSM Server is the bottleneck in your scenario. How many tapes are mounted for the log backup? How many TSM Server sessions are being started for the log backup? Can you examine the session statistics on the TSM Server? What is happening on each session? How much data is being sent for each session? What is the speed like when you use a lower number of stripes for the log backup? (For example, using 2 stripes.) Finding the bottleneck on the TSM Server/Storage Agent is the key to this one. Thanks, Del "ADSM: Dist Stor Manager" wrote on 09/20/2004 05:39:02 PM: > TSM Client V5.3.2.1, TDP for SQL V5.2.1, TSM Storage Agent V5.2.3.2 > TSM Server V5.2.3.2. > Windows 2003 Itanium Cluster with 2 servers. > > We are doing some SQL backup testing today. We have a 2.5 TB > database that we backed up at 477 GB/hr to 10 tape drives using > LANFree (2.5 TB's in 5.1 hours). > > We started a logs backup of a 13 GB Log and it has been running for > nearly 90 minutes so far. The backup mounted a tape via LanFree > right away and has had the same tape mounted in the same tape drive > this whole time. There are no errors being reported anywhere. > > On the command window where we started the TDP backup we see > "Waiting for TSM server" with pages of periods ".". Then we see a > message showing that is wrote some data. It keeps repeating that > process. At the rate being displayed of 620.12 Kb/Sec it looks like > the Log file backup is running at around 2 GB/hr. Any ideas what > might be going wrong? > > We still have all the settings maxed: > Buffers 8 > Buffersize 8191 > SQLBuffers 0 > SQLBuffersize 4096 > Stripes 10 > > These same settings didn't seem to effect the full backup adversly? > - Do you Yahoo!? Read only the mail you want - Yahoo! Mail SpamGuard.
Re: schedule of SQL LOG backup
I have a customer that runs log backups every hour. For some systems they can't loose more than 5 minutes of data. For those systems they implement clustering. Going with Marks idea can they afford to wait 10 hours for you to replay logs. In that time they can't backup the data because you are running a restore. If the real need is full time access to data without 0 potential for data loss then they shouldn't be relying solely on a backup application. Luc Beaudoin <[EMAIL PROTECTED]> wrote: Hi Mark ... So with that setup ... worst case ... they will loose 4 hours of work ??? I'm working in a hospital ... so even 1 hours lost of lab result or patient appointment can be kind of hell anyway .. if there is no way of putting minutes ... I will put the minimum ... 1 hours ... thanks a lot Mark Luc "Stapleton, Mark" Sent by: "ADSM: Dist Stor Manager" 2004-09-21 04:28 PM Please respond to "ADSM: Dist Stor Manager" To: [EMAIL PROTECTED] cc: Subject: Re: schedule of SQL LOG backup From: ADSM: Dist Stor Manager [mailto:[EMAIL PROTECTED] On Behalf Of Luc Beaudoin >I thought of doing Full backup every 8 hours and LOG backup >every 20 minutes ... >Is there a best pratice for SQL backup ?? What works best is whatever meets your business needs. Most of my customers do a full backup of databases once a day, and periodic log backups (every 4 hours, for example) throughout the day. YMMV. -- Mark Stapleton ([EMAIL PROTECTED]) Berbee Information Networks Office 262.521.5627 - Do you Yahoo!? vote.yahoo.com - Register online to vote today!
Re: D2D on AIX
Tim, we recently ran a bunch of tests on client side compression. In every test the backup ran for 2 to 3 times longer. In some cases this wouldn't be a big deal when you look at the backup alone being incremental and all. However, we also believed that it would also cause the restore to run 2 to 3 times as long to uncompress the data. As a result of these tests and thoughts we decided not to implement client side compression. Have you run any tests to see how compression effects your backups and restores? "Rushforth, Tim" <[EMAIL PROTECTED]> wrote: We use 5 days for reuse delay. I did a quick comparison using 25GB and 4GB volumes on our pilot with the following results: Disk Volumes - 25 GB Volumes Stored Data - 236 GB # of Disk Vols - 14 (including 2 pending volumes) Total allocation - 14 * 25GB = 350 GB 67% Utilization Disk Volumes - 4 GB Volumes (TSM2) Stored Data - 333 GB # of Disk Vols - 100 (including 12 pending volumes) Total allocation - 100 * 4GB = 400 GB 83% Utilization Reclamation was set at 25% for both of these. It would be interesting to see others peoples results. Thanks, Tim Rushforth City of Winnipeg -Original Message- From: Johnson, Milton [mailto:[EMAIL PROTECTED] Sent: Tuesday, September 21, 2004 1:04 PM To: [EMAIL PROTECTED] Subject: Re: D2D on AIX What do use for a reuse delay? How many pending volumes do you average? H. Milton Johnson -Original Message- From: ADSM: Dist Stor Manager [mailto:[EMAIL PROTECTED] On Behalf Of Rushforth, Tim Sent: Tuesday, September 21, 2004 1:00 PM To: [EMAIL PROTECTED] Subject: Re: D2D on AIX Eliza: At the Disk only Backups Technical Exchange, IBM recommended 2-4 GB volume size. (This was stated by the presenter, it was not written on the PDF presentation.) We started with 25 GB volumes and have now switched to 4 GB volumes. Using smaller volume sizes allows a better utilization of space and increases restore performance with multi-session restore. (Also helps eliminate contention if multiple clients are restoring from the same volume) Tim Rushforth City of Winnipeg -Original Message- From: Eliza Lau [mailto:[EMAIL PROTECTED] Sent: Tuesday, September 21, 2004 11:20 AM To: [EMAIL PROTECTED] Subject: Re: D2D on AIX Eric, What is the recommended volume size. I have seen someone mentioned 5G, but then the number of volumes will explode from about 800 (current # of 3590 primary tapes) to thousands. How about keeping the staging space so clients backup to staging then migrate to FILE volumes. Then every volume will be filled up. Eliza > > Hi Eliza! > You do want several smaller files, rather than a few very large files > because each client session will allocate a volume. File volumes > cannot be used concurrently by more than one session. > Kindest regards, > Eric van Loon > KLM Royal Dutch Airlines > > > -Original Message- > From: Eliza Lau [mailto:[EMAIL PROTECTED] > Sent: Monday, September 20, 2004 19:11 > To: [EMAIL PROTECTED] > Subject: D2D on AIX > > > Our 3494 with 3590K tapes in 3 frames is getting full. Instead of > adding another frame or upgrading to 3590H or 3592 tapes we are > looking into setting up a bunch of cheap ATA disks as primary storage. > > The FILE devclass defines a directory as its destination and JFS2 has > a max file system size of 1TB. Does it mean the largest stgpool I can > define is 1TB? > > My Exchange stgpool alone has 8TB of data. Do I have to split it up > into 8 pieces? > > server: TSM 5.2.2.5 on AIX 5.2 > database 90GB at 70% > Total backup data - 22TB > > Eliza Lau > Virginia Tech Computing Center > [EMAIL PROTECTED] > > > ** > For information, services and offers, please visit our web site: http://www.klm.com. This e-mail and any attachment may contain confidential and privileged material intended for the addressee only. If you are not the addressee, you are notified that no part of the e-mail or any attachment may be disclosed, copied or distributed, and that any other action related to this e-mail or attachment is strictly prohibited, and may be unlawful. If you have received this e-mail by error, please notify the sender immediately by return e-mail, and delete this message. Koninklijke Luchtvaart Maatschappij NV (KLM), its subsidiaries and/or its employees shall not be liable for the incorrect or incomplete transmission of this e-mail or any attachments, nor responsible for any delay in receipt. > ** > - Do you Yahoo!? vote.yahoo.com - Register online to vote today!
Re: D2D on AIX
Good questions. Our real world example:We went from around 8 - 12 GB/hr restore off of tape to over 40 GB/hr from the file device classes. Our test was a file server with a little over 300 GB of data. The File server and the TSM server both had 1 GB NIC's. Resource utilization was set to 10 in both cases. The data Steve Harris <[EMAIL PROTECTED]> wrote:How does TSM access the data on file volumes? Does it keep an offset of the start of every file or aggregate? If it does, then yes we could skip to the start of each file or aggregate. If it does not, then we need to read through the volume to find the file we are going to restore. Where we have a large number of concurrent restores happening, this could cause performance issues on the array. Now TSM has some smarts on later technology tape drives that have block addressability and on-cartridge memory and can find a spot on the tape quickly, but does this translate to file volumes? Regards Steve. >>> [EMAIL PROTECTED] 22/09/2004 4:49:55 >>> True. Seek time is tiny compared to tape mounts. I am just concerned that the TSM db has to keep track of thousands of volume. How much will it increase the size of the db. Ours is already 90G at 70% utilized. Eliza > > ==> In article <[EMAIL PROTECTED]>, Eliza Lau writes: > > > What is the recommended volume size. I have seen someone mentioned 5G, but > > then the number of volumes will explode from about 800 (current # of 3590 > > primary tapes) to thousands. > > Consider, this doesn't really cost you much. Seek time in a directory of > thousands of files is still tiny compared to tape behavior. > > I probably wouldn't go as low as 5G, but 10G (much less than the average size > of my 3590 vols) seems pretty reasonable to me. 20G is getting big, from my > perspective. > > > > > How about keeping the staging space so clients backup to staging then > > migrate to FILE volumes. Then every volume will be filled up. > > > I like this, too. > > > - Allen S. Rout > *** This email, including any attachments sent with it, is confidential and for the sole use of the intended recipient(s). This confidentiality is not waived or lost, if you receive it and you are not the intended recipient(s), or if it is transmitted/received in error. Any unauthorised use, alteration, disclosure, distribution or review of this email is prohibited. It may be subject to a statutory duty of confidentiality if it relates to health service matters. If you are not the intended recipient(s), or if you have received this email in error, you are asked to immediately notify the sender by telephone or by return email. You should also delete this email and destroy any hard copies produced. *** __ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com
Re: D2D on AIX
Good questions. Our real world example:We went from around 8 - 12 GB/hr restore off of tape to over 40 GB/hr from the file device classes. Our test was a file server with a little over 300 GB of data. The File server and the TSM server both had 1 GB NIC's. Resource utilization was set to 10 in both cases. The d Steve Harris <[EMAIL PROTECTED]> wrote:How does TSM access the data on file volumes? Does it keep an offset of the start of every file or aggregate? If it does, then yes we could skip to the start of each file or aggregate. If it does not, then we need to read through the volume to find the file we are going to restore. Where we have a large number of concurrent restores happening, this could cause performance issues on the array. Now TSM has some smarts on later technology tape drives that have block addressability and on-cartridge memory and can find a spot on the tape quickly, but does this translate to file volumes? Regards Steve. >>> [EMAIL PROTECTED] 22/09/2004 4:49:55 >>> True. Seek time is tiny compared to tape mounts. I am just concerned that the TSM db has to keep track of thousands of volume. How much will it increase the size of the db. Ours is already 90G at 70% utilized. Eliza > > ==> In article <[EMAIL PROTECTED]>, Eliza Lau writes: > > > What is the recommended volume size. I have seen someone mentioned 5G, but > > then the number of volumes will explode from about 800 (current # of 3590 > > primary tapes) to thousands. > > Consider, this doesn't really cost you much. Seek time in a directory of > thousands of files is still tiny compared to tape behavior. > > I probably wouldn't go as low as 5G, but 10G (much less than the average size > of my 3590 vols) seems pretty reasonable to me. 20G is getting big, from my > perspective. > > > > > How about keeping the staging space so clients backup to staging then > > migrate to FILE volumes. Then every volume will be filled up. > > > I like this, too. > > > - Allen S. Rout > *** This email, including any attachments sent with it, is confidential and for the sole use of the intended recipient(s). This confidentiality is not waived or lost, if you receive it and you are not the intended recipient(s), or if it is transmitted/received in error. Any unauthorised use, alteration, disclosure, distribution or review of this email is prohibited. It may be subject to a statutory duty of confidentiality if it relates to health service matters. If you are not the intended recipient(s), or if you have received this email in error, you are asked to immediately notify the sender by telephone or by return email. You should also delete this email and destroy any hard copies produced. *** - Do you Yahoo!? vote.yahoo.com - Register online to vote today!
Re: D2D on AIX
Good questions. Our real world example:We went from around 8 - 12 GB/hr restore off of tape to over 40 GB/hr from the file device classes. Our test was a file server with a little over 300 GB of data. The File server and the TSM server both had 1 GB NIC's. Resource utilization was set to 10 in both cases. The data was Steve Harris <[EMAIL PROTECTED]> wrote:How does TSM access the data on file volumes? Does it keep an offset of the start of every file or aggregate? If it does, then yes we could skip to the start of each file or aggregate. If it does not, then we need to read through the volume to find the file we are going to restore. Where we have a large number of concurrent restores happening, this could cause performance issues on the array. Now TSM has some smarts on later technology tape drives that have block addressability and on-cartridge memory and can find a spot on the tape quickly, but does this translate to file volumes? Regards Steve. >>> [EMAIL PROTECTED] 22/09/2004 4:49:55 >>> True. Seek time is tiny compared to tape mounts. I am just concerned that the TSM db has to keep track of thousands of volume. How much will it increase the size of the db. Ours is already 90G at 70% utilized. Eliza > > ==> In article <[EMAIL PROTECTED]>, Eliza Lau writes: > > > What is the recommended volume size. I have seen someone mentioned 5G, but > > then the number of volumes will explode from about 800 (current # of 3590 > > primary tapes) to thousands. > > Consider, this doesn't really cost you much. Seek time in a directory of > thousands of files is still tiny compared to tape behavior. > > I probably wouldn't go as low as 5G, but 10G (much less than the average size > of my 3590 vols) seems pretty reasonable to me. 20G is getting big, from my > perspective. > > > > > How about keeping the staging space so clients backup to staging then > > migrate to FILE volumes. Then every volume will be filled up. > > > I like this, too. > > > - Allen S. Rout > *** This email, including any attachments sent with it, is confidential and for the sole use of the intended recipient(s). This confidentiality is not waived or lost, if you receive it and you are not the intended recipient(s), or if it is transmitted/received in error. Any unauthorised use, alteration, disclosure, distribution or review of this email is prohibited. It may be subject to a statutory duty of confidentiality if it relates to health service matters. If you are not the intended recipient(s), or if you have received this email in error, you are asked to immediately notify the sender by telephone or by return email. You should also delete this email and destroy any hard copies produced. *** - Do you Yahoo!? vote.yahoo.com - Register online to vote today!
Re: schedule of SQL LOG backup
Hello Luc, I think the only way to do this from the TSM scheduler is to define three schedules, one with a start time of 00:00, one at 00:20 and one at 00:40. Then set the periodicity to one hour for each of the three schedules. Cheers, Paul. -Original Message- From: ADSM: Dist Stor Manager [mailto:[EMAIL PROTECTED] On Behalf Of Luc Beaudoin Sent: Tuesday 21 September 2004 22:33 To: [EMAIL PROTECTED] Subject: Re: schedule of SQL LOG backup Hi Mark ... So with that setup ... worst case ... they will loose 4 hours of work ??? I'm working in a hospital ... so even 1 hours lost of lab result or patient appointment can be kind of hell anyway .. if there is no way of putting minutes ... I will put the minimum ... 1 hours ... thanks a lot Mark Luc "Stapleton, Mark" <[EMAIL PROTECTED]> Sent by: "ADSM: Dist Stor Manager" <[EMAIL PROTECTED]> 2004-09-21 04:28 PM Please respond to "ADSM: Dist Stor Manager" To: [EMAIL PROTECTED] cc: Subject:Re: schedule of SQL LOG backup From: ADSM: Dist Stor Manager [mailto:[EMAIL PROTECTED] On Behalf Of Luc Beaudoin >I thought of doing Full backup every 8 hours and LOG backup >every 20 minutes ... >Is there a best pratice for SQL backup ?? What works best is whatever meets your business needs. Most of my customers do a full backup of databases once a day, and periodic log backups (every 4 hours, for example) throughout the day. YMMV. -- Mark Stapleton ([EMAIL PROTECTED]) Berbee Information Networks Office 262.521.5627 Any e-mail message from the European Central Bank (ECB) is sent in good faith but shall neither be binding nor construed as constituting a commitment by the ECB except where provided for in a written agreement. This e-mail is intended only for the use of the recipient(s) named above. Any unauthorised disclosure, use or dissemination, either in whole or in part, is prohibited. If you have received this e-mail in error, please notify the sender immediately via e-mail and delete this e-mail from your system.
