Re: [zfs-discuss] ZFS snapshot GUI
Very good points Rang, I'm going to add to them with a few of my own. It should be possible to restore individual files rather than rolling back the snapshot and I guess that's what was meant here. I think the terminology in the original post may not be too clear. However, my impression reading this is that this is an application that runs directly on the machine. If so, we're missing an opportunity here. Solaris isn't really an end user OS, it's more of a server OS. If you are going to implement a nice GUI for restoring files from a snapshot, you really want that to work over a network as well as on the local machine. Ironically, if you're a windows user you already have that ability over the network with Solaris. Run ZFS and Samba and windows users can use Microsoft's Shadow Copy Client to right-click any file and easily restore it from a snapshot: http://helpdesk.its.uiowa.edu/windows/instructions/shadowcopy.htm What's really needed is a way to do that on Solaris and Linux machines over the network. Integration with Apple's time machine would be great too (especially as it sounds like they may be making it compatible with ZFS), but unless somebody high up in Sun speaks to Apple I don't see that happening. So you need two UI's: - On the server side a simple UI is needed for creating and scheduling snapshots of the filesystem. Tim Foster's service would be a good starting point for that: http://blogs.sun.com/timf/entry/zfs_automatic_for_the_people - On the client side a simple UI is needed that allows users to easily see previous versions of files and folders, and either restore them in place or copy old versions to a new location. And the client side of this would want to be capable of running either locally or over the network. I think you could probably bodge this by virtue of the fact that you can browse the files in a snapshot. Performance would probably be slow however and I've no doubt that far better performance could be achieved with hooks into ZFS (which incidentally would benefit apple if they want to move time machine to ZFS). That kind of thing is way outside my experience however, but it would be good if somebody at Sun could think about it. This message posted from opensolaris.org ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] ZFS snapshot GUI
Hi, In respect of snapshots :- a) should the snapshot process it self be modified to allow restoring of individual files via zfs rollback b) should there be a zfs rollfile to selectively restore files from a snapshot c) should there be a zfs purge which would allow file(s) to be removed from zfs including all snapshots Ross wrote: > Very good points Rang, I'm going to add to them with a few of my own. > > It should be possible to restore individual files rather than rolling back > the snapshot and I guess that's what was meant here. I think the terminology > in the original post may not be too clear. > > However, my impression reading this is that this is an application that runs > directly on the machine. If so, we're missing an opportunity here. Solaris > isn't really an end user OS, it's more of a server OS. If you are going to > implement a nice GUI for restoring files from a snapshot, you really want > that to work over a network as well as on the local machine. > > Ironically, if you're a windows user you already have that ability over the > network with Solaris. Run ZFS and Samba and windows users can use > Microsoft's Shadow Copy Client to right-click any file and easily restore it > from a snapshot: > http://helpdesk.its.uiowa.edu/windows/instructions/shadowcopy.htm > > What's really needed is a way to do that on Solaris and Linux machines over > the network. Integration with Apple's time machine would be great too > (especially as it sounds like they may be making it compatible with ZFS), but > unless somebody high up in Sun speaks to Apple I don't see that happening. > > So you need two UI's: > > - On the server side a simple UI is needed for creating and scheduling > snapshots of the filesystem. Tim Foster's service would be a good starting > point for that: http://blogs.sun.com/timf/entry/zfs_automatic_for_the_people > > - On the client side a simple UI is needed that allows users to easily see > previous versions of files and folders, and either restore them in place or > copy old versions to a new location. > > And the client side of this would want to be capable of running either > locally or over the network. > > I think you could probably bodge this by virtue of the fact that you can > browse the files in a snapshot. Performance would probably be slow however > and I've no doubt that far better performance could be achieved with hooks > into ZFS (which incidentally would benefit apple if they want to move time > machine to ZFS). > > That kind of thing is way outside my experience however, but it would be good > if somebody at Sun could think about it. > > > This message posted from opensolaris.org > ___ > zfs-discuss mailing list > zfs-discuss@opensolaris.org > http://mail.opensolaris.org/mailman/listinfo/zfs-discuss > > __ > This email has been scanned by the MessageLabs Email Security System. > For more information please visit http://www.messagelabs.com/email > __ > > -- Regards Russell ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] ZFS snapshot GUI
hi there, On Thu, 2007-11-22 at 00:53 -0800, Ross wrote: > It should be possible to restore individual files rather than rolling > back the snapshot and I guess that's what was meant here. I think the > terminology in the original post may not be too clear. Yep, I agree. > However, my impression reading this is that this is an application > that runs directly on the machine. If so, we're missing an > opportunity here. Solaris isn't really an end user OS, it's more of a > server OS. If you are going to implement a nice GUI for restoring > files from a snapshot, you really want that to work over a network as > well as on the local machine. Definitely - you can do that now over NFS. Here space/timf is a ZFS dataset on my desktop machine, haiiro. On "spoon" a client machine, I browse to haiiro's NFS shares: [EMAIL PROTECTED] cd /net/haiiro/space/timf/.zfs [EMAIL PROTECTED] cd snapshot [EMAIL PROTECTED] ls -1 total 56 3 backup-2007-09-25-16-21-05/ 3 backup-2007-09-25-16-49-42/ 3 backup-2007-09-25-16-53-37/ 3 backup-2007-09-25-17-35-07/ . . etc. Those are all snapshots taken on the filesystems on haiiro. You can also mkdir inside a remote directories to create snapshots, assuming you've been delegated that permission: [EMAIL PROTECTED] mkdir new [EMAIL PROTECTED] ssh [EMAIL PROTECTED] /usr/sbin/zfs list space/[EMAIL PROTECTED] NAME USED AVAIL REFER MOUNTPOINT space/[EMAIL PROTECTED] 0 - 5,18G - > - On the client side a simple UI is needed that allows users to > easily see previous versions of files and folders, and either restore > them in place or copy old versions to a new location. And that's what this is all about - trying to find a cleaner way than http://blogs.sun.com/timf/entry/zfs_on_your_desktop to tie the client and server sides together. cheers, tim -- Tim Foster, Sun Microsystems Inc, Solaris Engineering Ops http://blogs.sun.com/timf ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] Home Motherboard
If you want a board that is a steal look at this one: http://www.ascendtech.us/itemdesc.asp?ic=MBTAS2882G3NR Tyan S2882, Dual Socket 940 Opteron, 8 DDR slots, 2 PCI-X 133 busses with 2 slots each, Dual Core support. $80. Pair is with a couple of Opteron 270's from ebay for $195: http://cgi.ebay.com/MATCH-PAIR-AMD-Opteron-270-64Bit-DualCore-940pin-2Ghz_W0QQitemZ290182379420QQihZ019QQcategoryZ80142QQssPageNameZWDVWQQrdZ1QQcmdZViewItem Granted, you need an E-ATX case but those are not that expensive and an EPS12V power supply. For less than $600 you can have a hell of a base server grade system with 4 cores and 2-4G of ram. - Original Message - From: "Rob Logan" <[EMAIL PROTECTED]> To: zfs-discuss@opensolaris.org Sent: Wednesday, November 21, 2007 11:17:19 PM (GMT-0600) America/Chicago Subject: [zfs-discuss] Home Motherboard grew tired of the recycled 32bit cpus in http://www.opensolaris.org/jive/thread.jspa?messageID=127555 and bought this to put the two marvell88sx cards in: $255 http://www.supermicro.com/products/motherboard/Xeon3000/3210/X7SBE.cfm http://www.supermicro.com/manuals/motherboard/3210/MNL-0970.pdf $195 1333FSB 2.6GHz Xeon 3075 (basicly a E6750) Any Core 2 Quad/Duo in LGA775 will work, including 45nm dies: http://rob.com/sun/x7sbe/45nm-pricing.jpg $270 Four 1G PC2-6400 DDRII 800MHz 240-pin ECC Unbuffered SDRAM $ 55 LOM (IPMI and Serial over LAN) http://www.supermicro.com/manuals/other/AOC-SIMSOLC-HTC.pdf # /usr/X11/bin/scanpci pci bus 0x cardnum 0x00 function 0x00: vendor 0x8086 device 0x29f0 Intel Corporation Server DRAM Controller pci bus 0x cardnum 0x01 function 0x00: vendor 0x8086 device 0x29f1 Intel Corporation Server Host-Primary PCI Express Bridge pci bus 0x cardnum 0x1a function 0x00: vendor 0x8086 device 0x2937 Intel Corporation USB UHCI Controller #4 pci bus 0x cardnum 0x1a function 0x01: vendor 0x8086 device 0x2938 Intel Corporation USB UHCI Controller #5 pci bus 0x cardnum 0x1a function 0x02: vendor 0x8086 device 0x2939 Intel Corporation USB UHCI Controller #6 pci bus 0x cardnum 0x1a function 0x07: vendor 0x8086 device 0x293c Intel Corporation USB2 EHCI Controller #2 pci bus 0x cardnum 0x1c function 0x00: vendor 0x8086 device 0x2940 Intel Corporation PCI Express Port 1 pci bus 0x cardnum 0x1c function 0x04: vendor 0x8086 device 0x2948 Intel Corporation PCI Express Port 5 pci bus 0x cardnum 0x1c function 0x05: vendor 0x8086 device 0x294a Intel Corporation PCI Express Port 6 pci bus 0x cardnum 0x1d function 0x00: vendor 0x8086 device 0x2934 Intel Corporation USB UHCI Controller #1 pci bus 0x cardnum 0x1d function 0x01: vendor 0x8086 device 0x2935 Intel Corporation USB UHCI Controller #2 pci bus 0x cardnum 0x1d function 0x02: vendor 0x8086 device 0x2936 Intel Corporation USB UHCI Controller #3 pci bus 0x cardnum 0x1d function 0x07: vendor 0x8086 device 0x293a Intel Corporation USB2 EHCI Controller #1 pci bus 0x cardnum 0x1e function 0x00: vendor 0x8086 device 0x244e Intel Corporation 82801 PCI Bridge pci bus 0x cardnum 0x1f function 0x00: vendor 0x8086 device 0x2916 Intel Corporation Device unknown pci bus 0x cardnum 0x1f function 0x02: vendor 0x8086 device 0x2922 Intel Corporation 6 port SATA AHCI Controller pci bus 0x cardnum 0x1f function 0x03: vendor 0x8086 device 0x2930 Intel Corporation SMBus Controller pci bus 0x cardnum 0x1f function 0x06: vendor 0x8086 device 0x2932 Intel Corporation Thermal Subsystem pci bus 0x0001 cardnum 0x00 function 0x00: vendor 0x8086 device 0x0329 Intel Corporation 6700PXH PCI Express-to-PCI Bridge A pci bus 0x0001 cardnum 0x00 function 0x01: vendor 0x8086 device 0x0326 Intel Corporation 6700/6702PXH I/OxAPIC Interrupt Controller A pci bus 0x0001 cardnum 0x00 function 0x02: vendor 0x8086 device 0x032a Intel Corporation 6700PXH PCI Express-to-PCI Bridge B pci bus 0x0001 cardnum 0x00 function 0x03: vendor 0x8086 device 0x0327 Intel Corporation 6700PXH I/OxAPIC Interrupt Controller B pci bus 0x0003 cardnum 0x02 function 0x00: vendor 0x11ab device 0x6081 Marvell Technology Group Ltd. MV88SX6081 8-port SATA II PCI-X Controller pci bus 0x000d cardnum 0x00 function 0x00: vendor 0x8086 device 0x108c Intel Corporation 82573E Gigabit Ethernet Controller (Copper) pci bus 0x000f cardnum 0x00 function 0x00: vendor 0x8086 device 0x109a Intel Corporation 82573L Gigabit Ethernet Controller pci bus 0x0011 cardnum 0x04 function 0x00: vendor 0x1002 device 0x515e ATI Technologies Inc ES1000 # cfgadm -a Ap_Id Type Receptacle Occupant Condition pcie5 etherne/hp connectedconfigured ok pcie6 etherne/hp connectedconfigured ok sata0/0::dsk/c0t0d0disk connectedconfigured ok sata0/1::dsk/c0t1d0disk connectedconfigured ok sata0/2::ds
Re: [zfs-discuss] Home Motherboard
The new Intel D201GLY2 looks quite good. Fanless 64-bit CPU, low-power consumption from what I have read. Awaiting first substantive review from SilentPCReview.com before ordering one. This message posted from opensolaris.org ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
[zfs-discuss] cifs server?
So now that cifs has finally been released in b77, anyone happen to have any documentation on setup. I know the initial share is relatively simple... but what is the process after that for actually getting users authenticated? I see in the idmap service there's some configurations for authenticating against an AD server, but is there anyway to get it to authenticate with a local database? In my test environment I only have a couple users, and no AD server, so it seems a bit silly to set one up simply to get authentication to the cifs shares. Docs/links, anyone? This message posted from opensolaris.org ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] cifs server?
so apparently you need to use smbadm, but when I got to create the group: smbadm create wheel failed to create the group (NOT_SUPPORTED) This message posted from opensolaris.org ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] cifs server?
On Thu, Nov 22, 2007 at 10:27:18AM -0800, Tim Cook wrote: > So now that cifs has finally been released in b77, anyone happen to It hasn't been released. It was integrated into build 77. > have any documentation on setup. I know the initial share is The documentation will be available in the first SX release to officially have the SMB server. > relatively simple... but what is the process after that for actually > getting users authenticated? I see in the idmap service there's some > configurations for authenticating against an AD server, but is there > anyway to get it to authenticate with a local database? There's been a lot of bugfixes, so I higly recommend waiting at least until build 79 closes. What you would do is enable the idmap and smb/server services, then run "smbadm join" and go from there. But like I said, I recommend waiting for build 79. > In my test environment I only have a couple users, and no AD server, > so it seems a bit silly to set one up simply to get authentication to > the cifs shares. Docs/links, anyone? There is a workgroup mode (this is controlled by a property of the smb/server service), but I'm not sure how well it works as I've not used it. Perhaps the CIFS server project team can follow up here (but keep in mind it's Thanksgiving week...). Nico -- ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] Home Motherboard
here is a simple layout for 6 disks toward "speed" : /dev/dsk/c0t0d0s1 - - swap- no - /dev/dsk/c0t1d0s1 - - swap- no - root/snv_77 - / zfs - no - z/snv_77/usr - /usr zfs - yes - z/snv_77/var - /var zfs - yes - z/snv_77/opt - /opt zfs - yes - root @test[2:25pm]/root/boot/ grub 27 % zpool iostat -v capacity operationsbandwidth pool used avail read write read write - - - - - - root 4.39G 15.1G179 1 3.02M 16.0K mirror 4.39G 15.1G179 1 3.02M 16.0K c0t1d0s0 - - 62 1 3.35M 22.3K c0t0d0s0 - - 61 1 3.35M 22.3K - - - - - - z 319G 421G 1.15K 17 81.9M 83.8K mirror 113G 163G418 8 28.8M 40.8K c0t0d0s7 - -272 2 29.4M 48.2K c0t1d0s7 - -272 2 29.5M 48.2K mirror 103G 129G376 4 26.5M 21.4K c0t2d0- -250 3 27.1M 28.9K c0t3d0- -250 2 27.1M 28.9K mirror 104G 128G380 4 26.6M 21.6K c0t4d0- -253 2 27.1M 29.0K c0t5d0- -252 2 27.1M 29.0K - - - - - - ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] Home Motherboard
I actually have a related motherboard, chassis, dual power-supplies and 12x400 gig drives already up on ebay too. If I recall Areca cards are supported in OpenSolaris... http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&item=300172982498 On 11/22/07, Jason P. Warr <[EMAIL PROTECTED]> wrote: > If you want a board that is a steal look at this one: > > http://www.ascendtech.us/itemdesc.asp?ic=MBTAS2882G3NR > > Tyan S2882, Dual Socket 940 Opteron, 8 DDR slots, 2 PCI-X 133 busses with 2 > slots each, Dual Core support. > > $80. > > Pair is with a couple of Opteron 270's from ebay for $195: > > http://cgi.ebay.com/MATCH-PAIR-AMD-Opteron-270-64Bit-DualCore-940pin-2Ghz_W0QQitemZ290182379420QQihZ019QQcategoryZ80142QQssPageNameZWDVWQQrdZ1QQcmdZViewItem > > Granted, you need an E-ATX case but those are not that expensive and an > EPS12V power supply. For less than $600 you can have a hell of a base server > grade system with 4 cores and 2-4G of ram. > > > - Original Message - > From: "Rob Logan" <[EMAIL PROTECTED]> > To: zfs-discuss@opensolaris.org > Sent: Wednesday, November 21, 2007 11:17:19 PM (GMT-0600) America/Chicago > Subject: [zfs-discuss] Home Motherboard > > grew tired of the recycled 32bit cpus in > http://www.opensolaris.org/jive/thread.jspa?messageID=127555 > > and bought this to put the two marvell88sx cards in: > $255 http://www.supermicro.com/products/motherboard/Xeon3000/3210/X7SBE.cfm > http://www.supermicro.com/manuals/motherboard/3210/MNL-0970.pdf > $195 1333FSB 2.6GHz Xeon 3075 (basicly a E6750) > Any Core 2 Quad/Duo in LGA775 will work, including 45nm dies: > http://rob.com/sun/x7sbe/45nm-pricing.jpg > $270 Four 1G PC2-6400 DDRII 800MHz 240-pin ECC Unbuffered SDRAM > $ 55 LOM (IPMI and Serial over LAN) > http://www.supermicro.com/manuals/other/AOC-SIMSOLC-HTC.pdf > > # /usr/X11/bin/scanpci > pci bus 0x cardnum 0x00 function 0x00: vendor 0x8086 device 0x29f0 > Intel Corporation Server DRAM Controller > > pci bus 0x cardnum 0x01 function 0x00: vendor 0x8086 device 0x29f1 > Intel Corporation Server Host-Primary PCI Express Bridge > > pci bus 0x cardnum 0x1a function 0x00: vendor 0x8086 device 0x2937 > Intel Corporation USB UHCI Controller #4 > > pci bus 0x cardnum 0x1a function 0x01: vendor 0x8086 device 0x2938 > Intel Corporation USB UHCI Controller #5 > > pci bus 0x cardnum 0x1a function 0x02: vendor 0x8086 device 0x2939 > Intel Corporation USB UHCI Controller #6 > > pci bus 0x cardnum 0x1a function 0x07: vendor 0x8086 device 0x293c > Intel Corporation USB2 EHCI Controller #2 > > pci bus 0x cardnum 0x1c function 0x00: vendor 0x8086 device 0x2940 > Intel Corporation PCI Express Port 1 > > pci bus 0x cardnum 0x1c function 0x04: vendor 0x8086 device 0x2948 > Intel Corporation PCI Express Port 5 > > pci bus 0x cardnum 0x1c function 0x05: vendor 0x8086 device 0x294a > Intel Corporation PCI Express Port 6 > > pci bus 0x cardnum 0x1d function 0x00: vendor 0x8086 device 0x2934 > Intel Corporation USB UHCI Controller #1 > > pci bus 0x cardnum 0x1d function 0x01: vendor 0x8086 device 0x2935 > Intel Corporation USB UHCI Controller #2 > > pci bus 0x cardnum 0x1d function 0x02: vendor 0x8086 device 0x2936 > Intel Corporation USB UHCI Controller #3 > > pci bus 0x cardnum 0x1d function 0x07: vendor 0x8086 device 0x293a > Intel Corporation USB2 EHCI Controller #1 > > pci bus 0x cardnum 0x1e function 0x00: vendor 0x8086 device 0x244e > Intel Corporation 82801 PCI Bridge > > pci bus 0x cardnum 0x1f function 0x00: vendor 0x8086 device 0x2916 > Intel Corporation Device unknown > > pci bus 0x cardnum 0x1f function 0x02: vendor 0x8086 device 0x2922 > Intel Corporation 6 port SATA AHCI Controller > > pci bus 0x cardnum 0x1f function 0x03: vendor 0x8086 device 0x2930 > Intel Corporation SMBus Controller > > pci bus 0x cardnum 0x1f function 0x06: vendor 0x8086 device 0x2932 > Intel Corporation Thermal Subsystem > > pci bus 0x0001 cardnum 0x00 function 0x00: vendor 0x8086 device 0x0329 > Intel Corporation 6700PXH PCI Express-to-PCI Bridge A > > pci bus 0x0001 cardnum 0x00 function 0x01: vendor 0x8086 device 0x0326 > Intel Corporation 6700/6702PXH I/OxAPIC Interrupt Controller A > > pci bus 0x0001 cardnum 0x00 function 0x02: vendor 0x8086 device 0x032a > Intel Corporation 6700PXH PCI Express-to-PCI Bridge B > > pci bus 0x0001 cardnum 0x00 function 0x03: vendor 0x8086 device 0x0327 > Intel Corporation 6700PXH I/OxAPIC Interrupt Controller B > > pci bus 0x0003 cardnum 0x02 function 0x00: vendor 0x11ab device 0x6081 > Marvell Technology Group Ltd. MV88SX6081 8-port SATA II PCI-X > Controller > > pci bus 0x000d cardnum 0x00 function 0x00: vendor 0x8086 device 0x108c > Intel Corporation 82573E Gigabit Ethernet Controller (Copper) > > pci bus 0x000f cardnum 0x00 function 0x00: vendor 0x8086 device 0x109a > Intel Corporation 82573L Gigabit Ethernet Controller > > pci bus 0x0011 cardnum 0x04 function 0x
[zfs-discuss] ZFS & Lustre
First reported by a Sun blogger (http://blogs.sun.com/simons/), most of the HPC Consortium meeting in Reno presentations are now onlone: https://events-at-sun.com/hpcreno/presentations.html Some people on this list may be interested in this: "Lustre – CFS Update, Peter Braam, CEO, Cluster File System" In Lustre 1.8, ZFS will be available as an option for the disk filesystem - Lustre servers will be in user space - will have user space ZFS code – DMU I don't know enough about Lustre to comment on how this is done -- any volunteers?? :-) Rayson ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] Home Motherboard
> with 4 cores and 2-4G of ram.
not sure 2G is enough... at least with 64bit there are no kernel space
issues.
6 % echo '::memstat' | mdb -k
Page SummaryPagesMB %Tot
Kernel 692075 2703 66%
Anon33265 1293%
Exec and libs8690331%
Page cache 1143 40%
Free (cachelist) 3454130%
Free (freelist)307400 1200 29%
Total 1046027 4086
Physical 1046026 4086
this tree on the 64bit v20z box:
Page SummaryPagesMB %Tot
Kernel 668799 2612 85%
Anon38477 1505%
Exec and libs4881191%
Page cache 5363201%
Free (cachelist) 7566291%
Free (freelist) 59052 2308%
Total 784138 3063
Physical 784137 3063
and the same tree on a 32bit box:
Page SummaryPagesMB %Tot
Kernel 261359 1020 33%
Anon52314 2047%
Exec and libs 12245472%
Page cache 9885381%
Free (cachelist) 6816261%
Free (freelist) 7408819027247 240518168576 593093167776006144%
Total 784266 3063
Physical 784265 3063
from http://www.opensolaris.org/jive/message.jspa?messageID=173580
8 % ./zfs-mem-used
checking pool map size [B]: root
358424
checking pool map size [B]: z
4162512
9 % cat zfs-mem-used
#!/bin/sh
echo '::spa' | mdb -k | grep ACTIVE \
| while read pool_ptr state pool_name
do
echo "checking pool map size [B]: $pool_name"
echo "${pool_ptr}::walk metaslab|::print -d struct metaslab
ms_smo.smo_objsize" \
| mdb -k \
| nawk '{sub("^0t","",$3);sum+=$3}END{print sum}'
done
___
zfs-discuss mailing list
zfs-discuss@opensolaris.org
http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
[zfs-discuss] Questions from a windows admin - Samba, shares & quotas
Hey folks, This may sound a little crazy, but I'm a long time windows admin planning on rolling out a Solaris server to act as our main filestore, and I could do with a bit of advice. The main reason for switching is so we can use snapshots. With Samba and Microsoft's Shadow Copy Client we can backup everybody's files and give users the power to restore files themselves. That plus the other benefits of ZFS mean we're seriously looking into this. However, there are one or two side effects... I'm more than a little concerned about how we go about creating user profiles and managing quotas. On a windows server this is easy. As you create a user account the appropriate folders are created for you. So long as you have the right permissions on the parent folder, all those folders inherit their permissions automatically, and since quotas work per user, that's automatic too. The question is, can I do anything to automate all this if I move to Solaris? We've got to use ZFS to get the benefits, but that doesn't have user quotas so I'll have to script the creation of a filesystem for each user. First of all, if I have multiple filesystems, can I still share those out under one path? ie: each user has a home folder of \\server\share\username, can I still do that when every 'username' is a separate filesystem? Then, if that's possible, is there any way I can make the creation of these filesystems automatic? What I'm thinking is that I'll need a parent filesystem for these to inherit quota settings from, and that will be the 'share' location above. Now, when windows creates user accounts, it will automatically create subfolders in that parent filesystem to act as home directories. Would I be able to write a script to watch that filesystem for new subfolders and have it automatically delete the subfolder and create a filesystem in it's place? And can I set permissions on a filesystem? Could the script to that too? If it works, while it may be a bit messy, there will actually be some advantages over the windows quota system. In windows, the quota is set by ownership of the file. If users move files between each other for any reason, or an admin has to take ownership to change permissions that can get messed up, and it's nigh on impossible to find out where all your space has gone if that happens. This message posted from opensolaris.org ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] Home Motherboard
I was interested in that one till I read: One 240-pin DDR2 SDRAM Dual Inline Memory Module (DIMM) sockets Support for DDR2 667 MHz, DDR2 533 MHz and DDR2 400 MHz DIMMs (DDR 667 MHz validated to run at 533 MHz only) Support for up to 1 GB of system memory Boo!!! :) Nathan. Vincent Fox wrote: > The new Intel D201GLY2 looks quite good. > > Fanless 64-bit CPU, low-power consumption from what I have read. Awaiting > first substantive review from SilentPCReview.com before ordering one. > > > This message posted from opensolaris.org > ___ > zfs-discuss mailing list > zfs-discuss@opensolaris.org > http://mail.opensolaris.org/mailman/listinfo/zfs-discuss ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] Home Motherboard
mike wrote: > I actually have a related motherboard, chassis, dual power-supplies > and 12x400 gig drives already up on ebay too. If I recall Areca cards > are supported in OpenSolaris... At the moment you can download the Areca "arcmsr" driver from areca.com.tw, but I'm in the process of integrating it into OpenSolaris http://bugs.opensolaris.org/view_bug.do?bug_id=6614012 6614012 add Areca SAS/SATA RAID adapter driver James C. McPherson -- Senior Kernel Software Engineer, Solaris Sun Microsystems http://blogs.sun.com/jmcp http://www.jmcp.homeunix.com/blog ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] ZFS + DB + "fragments"
> OK, I'll bite; it's not like I'm getting an answer to > my other question. Did I miss one somewhere? > > Bill, please explain why deciding what to do about > sequential scan > performance in ZFS is urgent? It's not so much that it's 'urgent' (anyone affected by it simply won't use ZFS) as that it's a no-brainer. > > ie why it's urgent rather than important (I agree > that if it's bad > hen it's going to be important eventually). It's bad, and it's important now for anyone who cares whether ZFS is viable for such workloads. > > ie why it's too urgent to work out, first, how to > measure whether > we're succeeding. You don't have to measure the *rate* at which the depth of the water in the boat is rising in order to know that you've got a problem that needs addressing. You don't have to measure *just how bad* sequential performance in a badly-fragmented file is to know that you've got a problem that needs addressing (see both Anton's and Roch's comments if you don't find mine convincing). *After* you've tried to fix things, *then* it makes sense to measure just how close you got to ideal streaming-sequential disk bandwidth in order to see whether you need to work some more. Right now, the only reason to measure precisely how awful sequential scanning performance can get after severely fragmenting a file by updating it randomly in small chunks is to be able to hand out "Attaboy!"s for how much changing it improved things - even though this by itself *still* won't say anything about whether the result attained offered reasonable performance in comparison with what's attainable (which is what should *really* be the basis for handing out any "Attaboy!"s). Rather than make a politically-incorrect comment about the Special Olympics here, I'll just ask whether common sense is no longer considered an essential attribute in an engineer: given the nature of the discussions about this and about RAID-Z, I've really got to wonder. - bill This message posted from opensolaris.org ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] Questions from a windows admin - Samba, shares & quotas
Well, it looks like I've solved the question of whether you can auto-create the folders. There's a nice little samba script that you can add to the share to do it for you: >From http://www.edplese.com/samba-with-zfs.html Samba's root preexec share parameter can really come in handy when setting up user home directories. Here we tell it to automatically create a ZFS filesystem for every new user, set the owner, and set the quota to 1 GB. This can easily be expanded to other filesystem properties as well. Create a file /usr/bin/createhome.sh: #!/usr/bin/tcsh if ( ! -e /tank/home/$1 ) then zfs create tank/home/$1 chown $1 tank/home/$1 zfs set quota=1G tank/home/$1 endif Modify smb.conf and modify [homes] to resemble: [homes] comment = User Home Directories browseable = no writable = yes root preexec = /usr/bin/createhome.sh '%U' This message posted from opensolaris.org ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
