Re: Partitioning between SDD and HDD
2012/12/21 Javier Perez : > Hi, Roberto, I understand your point. One is more likely to be shuffling > data than programs, therefore it is better to put the data on the faster > storage medium. > But the way I understand it, there is a limited number of times that one can > write to a SSD due to the intrinsic nature of the media. Therefore I would > not like to put write intensive files like data, caches, etc on this drive. > > Makes sense? > > Javier > Well, both yes, and no. There is a limit to how much data you can write to a SSD drive, so it is reasonable to reduce number of writes (e.g. mounting /tmp on tmpfs which is coming to Fedora helps). On the other hand current drives are designed to last several years under heavy load, e.g. Intel's SSD 330 is rated for 3 years of 20GB writes per day. I'd suggest you read warranty terms of drives you are interested in. If you find a drive that is covered by 3 or 5 year warranty under 15 GB/day it will last you a long time, and if it won't, you'll get a replacement from the vendor. As for the data placement scheme, what will be best for you depends heavily on how you expect to use the system. If I were to build a hybrid (SSD/HDD) system, I would put most of my /home on SSD, I would be strongly tempted to put the system on SSD too, and relegate HDD to storing media files (I do not edit audio/video, I listen to some music, and HDD is more than adequate for that), .iso images and virtual machine images (because SSDs are too expensive for me for bulk storage). If you are going to shell out for an SSD, then it's worth to put it, where it will make the biggest difference, i.e. where there are many random accesses, which kill HDDs. This means the documents you edit and the binaries you start, in this order. If you spend 15% of the system price on a part, that will give you 10-fold speed-up of a task, that constitutes 5% of the time you wait for the computer, you will not notice it much. If it gives you 10-fild speed-up of a task that constitutes 70% of the time you spend waiting for the computer, you'll notice where did your money go ;). Paweł -- users mailing list users@lists.fedoraproject.org To unsubscribe or change subscription options: https://admin.fedoraproject.org/mailman/listinfo/users Guidelines: http://fedoraproject.org/wiki/Mailing_list_guidelines Have a question? Ask away: http://ask.fedoraproject.org
Re: Swapping HDD....
2012/12/27 Eddie G. O'Connor Jr. : > WOW!so much information to digest! Thanks to one and all, but after > reading the responses, I think the Clonezilla way might be for me, I'm not > too familiar with the Terminal and the command lines and suchalthough I > think it would be AWESOME to be able to do such tings strictly from a > terminalI'm also afraid since this is the only "working" laptop I have > that connects me to the outside world...and I would hate to lose all the > info and files on here.maybe I'll find some old drives and do a "test" > run firstjust to be sure I've gotten the hang of it! Thanks again > everyone! > > > EGO II > Backup. You *need* backup. Now. From what I understand, you are going to perform an operation on a disk containing the sole copy of data which you'd "hate to loose". Backup this data now. Any rearrangement of a disk bears an inherent risk of thrashing the disk. A backup is easy to make with Fedora (e.g. using deja-dup), will reduce your stress level during the operation and making your first backup may get you into a good habit ;). Besides, I agree with your conclusion, that Clonezilla may be the best way for you. I used it and I like it. Paweł -- users mailing list users@lists.fedoraproject.org To unsubscribe or change subscription options: https://admin.fedoraproject.org/mailman/listinfo/users Guidelines: http://fedoraproject.org/wiki/Mailing_list_guidelines Have a question? Ask away: http://ask.fedoraproject.org
Re: iptables is like alchemy
2013/1/3 Jorge Fábregas : > Ok, I've posted a similar setup I've used in the past that worked like a > charm. The script is the actual /etc/sysconfig/iptables. You'll notice > the syntax there is somehow different than when you manually create the > rules (or put in a script) but you get the idea. Those rules WERE THE > MINIMUM required in order to let a machine on internet reach a machine > on the internal network (port 8,555). Ask me any question if you don't > understand a line (please specify line number in question). > > eth0 is WAN, eth1 is LAN. Notice how I use "-i" and "-o" for the NICs. > > http://fpaste.org/sdPF/ > > Hi, Could you, please, post the actual config? If the configh which you posted worked for you, then we know there's no error in it. We need to see the one which has some, you know ;). It may be something really minor, like a typo, which is very difficult to spot if you read the text for the umpteenth time, and you know what you expect to read. It may be, that all it needs is just another pair of eyes. As far as what you posted goes, if that script worked for you, then replacing the port numbers and the destination IP with the ones you need for ssh would give you ssh connectivity. However, there are no holes punched in the INPUT chain for DNS or IMAP services, I suppose that the relevant lines are present in the file which we haven't seen yet. HTH, Paweł -- users mailing list users@lists.fedoraproject.org To unsubscribe or change subscription options: https://admin.fedoraproject.org/mailman/listinfo/users Guidelines: http://fedoraproject.org/wiki/Mailing_list_guidelines Have a question? Ask away: http://ask.fedoraproject.org
Re: How to Set/Get UUID for a NIC
2013/1/9 Alan Cox : > On Wed, 9 Jan 2013 09:37:05 -0600 (CST) > Michael Hennebry wrote: > >> On Wed, 9 Jan 2013, Khemara Lyn wrote: >> >> > Ok, thank you; it's that simple! I've thought about it in a harder way. >> >> Actually, it's even easier. >> NICs come with built-in six-byte MAC adddresses >> that are supposed to be unique. > > They are supposed to be unique *per machine* - you can have two nics on > the same machine with the same MAC although this is rare. > > Alan Alan, please, verify information before dissemination. MAC address assigned to a physical NIC ought to be globally unique. Ethernet frames are sent to MAC address, so two identical addresses present in any broadcast domain would awfully confuse network switches. See e.g. Wikipedia article on MAC https://en.wikipedia.org/wiki/MAC_address. Hardware producers reserve prefixes (first 3 bytes of the address) and are responsible for not manufacturing devices with identical suffixes (last 3 bytes of the address). This should result in globally unique MAC addresses, however I did encounter NICs with identical ones (cheap stuff of unknown and dubious provenance). Khemara, If you use NetworkManager, then it will present you with graphical tool for configuration of NICs and it will handle identification for you. I think it will also generate a UUID for the card and put it into its config file. However, the required and sufficient entry mapping physical device to logical one is HWADDR="xx:yy:zz:aa:bb:cc" line, where xx:yy:zz:aa:bb:cc from this example would be replaced by the real MAC address of your card. UUID line is for NetworkManager's benefit, see also discussion in the list here: https://lists.fedoraproject.org/pipermail/users/2011-May/396591.html. You can determine NIC's MAC address by running ip addr show In my KVM F18 guest it produces: $ ip addr show eth0 2: eth0: mtu 1500 qdisc pfifo_fast state UP qlen 1000 link/ether 52:54:00:42:1d:c6 brd ff:ff:ff:ff:ff:ff inet 192.168.122.150/24 brd 192.168.122.255 scope global eth0 inet6 fe80::5054:ff:fe42:1dc6/64 scope link valid_lft forever preferred_lft forever $ The card's MAC address is also present in the ifcfg-eth0 file: $ cat ifcfg-eth0 UUID="06693902-df40-49f6-8f0e-c7bac49531c7" NM_CONTROLLED="yes" HWADDR="52:54:00:42:1D:C6" BOOTPROTO="dhcp" DEVICE="eth0" ONBOOT="yes" $ HTH, Paweł -- users mailing list users@lists.fedoraproject.org To unsubscribe or change subscription options: https://admin.fedoraproject.org/mailman/listinfo/users Guidelines: http://fedoraproject.org/wiki/Mailing_list_guidelines Have a question? Ask away: http://ask.fedoraproject.org
Re: powerdown restarts
2012/7/2 Richard Vickery : > On Mon, Jul 2, 2012 at 11:14 AM, Joe Zeff wrote: > How do we boot up after "halt" or "sleep"? I quit using these commands years > ago for lack of knowledge about booting back up. The man pages never gave me > what I needed to know, and now that it has been brought up, I thought that I > would ask and get my curiosity satisfied. > > Thanks. > I'm not using "sleep", but "halt" (stop the system, leave it powered up) is very handy for systems with an UPS. On prolonged power loss UPS will tell the system to halt, then wait a bit to give it time, then cut the power. When sweet electricity starts flowing from the socket in the wall once again, UPS gives power back to the system, which then restores itself to the last power state, i.e. powered up. This results in a nice boot and the machine is alive and kicking once again. If the system was shut down, then it would just remain shut down after power was restored. Regards, Paweł -- users mailing list users@lists.fedoraproject.org To unsubscribe or change subscription options: https://admin.fedoraproject.org/mailman/listinfo/users Guidelines: http://fedoraproject.org/wiki/Mailing_list_guidelines Have a question? Ask away: http://ask.fedoraproject.org
Re: How to debug high system load?
2012/7/15 suvayu ali : > Hi Heinz, > > Sorry for the late response. > > On Thu, Jul 12, 2012 at 7:31 PM, Heinz Diehl wrote: >> On 12.07.2012, Suvayu Ali wrote: (...) >> If you're using cfq as your scheduler, try this in rc.local: >> >> echo "32" > /sys/block/sda/queue/iosched/quantum >> echo "0" > /sys/block/sda/queue/iosched/slice_idle >> echo "1" > /sys/block/sda/queue/iosched/low_latency >> echo "51200" > /sys/block/sda/queue/nr_requests >> >> Together with this in /etc/sysctl.conf: >> >> vm.dirty_ratio = 10 >> vm.dirty_background_ratio = 5 >> > > I tend not to try things that I don't understand. Could please outline > briefly what the above suggestions do? I would like to understand before > I try them out. > > Thanks, > When vm.dirty_ratio percent of total system memory is taken up by dirty pages (data waiting to be saved to disk), the process which is generating disk writes will itself start writing out dirty data. It means, that program will be made to stop using system buffers to hide cost of writes, and to write the data to the disk. When vm.dirty_background_ratio of total system memory is taken up by dirty pages, the pdflush background writeback daemon will start writing out dirty data. If you keep these lower, the system will try to prevent accumulation of large amounts of data to write. Therefore, when sync comes, you will not have to wait for the accumulated data to be written to disk. vm tunables are explained here: https://www.kernel.org/doc/Documentation/sysctl/vm.txt. -- users mailing list users@lists.fedoraproject.org To unsubscribe or change subscription options: https://admin.fedoraproject.org/mailman/listinfo/users Guidelines: http://fedoraproject.org/wiki/Mailing_list_guidelines Have a question? Ask away: http://ask.fedoraproject.org
Re: Memory on new computer
2012/7/16 Martín Marqués : > I had a problem with my main board and CPU and had to buy a new one. > Also memory, as the old computer had DDR2 memory. > > Th problem is that I see only 3.5Gb of memory, while I had a 4Gb bank > put in the Motherboard (the motherboard says it can handle up to 32Gb > or ram). > > I saw this in dmesg: > > [0.00] Memory: 3491896k/4702208k available (6177k kernel code, > 1053240k absent, 157072k reserved, 7004k data, 1004k init) > > Why do I get this: > > $ free > total used free sharedbuffers cached > Mem: 35134122591976 921436 0 8332 914584 > -/+ buffers/cache:16690601844352 > Swap:0 0 0 > There are several possible factors: 1) Integrated video card, which uses system memory. 2) Peripheral devices I/O regions are mapped below 4 GB, shadowing the physical ram. Here is (old, but still valid) explanation given by IBM for their servers: https://www-947.ibm.com/support/entry/portal/docdisplay?lndocid=MIGR-4E4RRF and here's the same issue addressed by Microsoft: http://support.microsoft.com/kb/978610 I did not manage to find Red Hat's take on the issue. In short: Works as designed. You may try to disable devices you do not use to salvage some address space, but you won't reach 4096 MB available. > > -- > Martín Marqués > select 'martin.marques' || '@' || 'gmail.com' > DBA, Programador, Administrador > -- > users mailing list > users@lists.fedoraproject.org > To unsubscribe or change subscription options: > https://admin.fedoraproject.org/mailman/listinfo/users > Guidelines: http://fedoraproject.org/wiki/Mailing_list_guidelines > Have a question? Ask away: http://ask.fedoraproject.org -- users mailing list users@lists.fedoraproject.org To unsubscribe or change subscription options: https://admin.fedoraproject.org/mailman/listinfo/users Guidelines: http://fedoraproject.org/wiki/Mailing_list_guidelines Have a question? Ask away: http://ask.fedoraproject.org
Re: Memory on new computer
2012/7/16 Martín Marqués : > 2012/7/16 Paweł Brodacki : >> 2012/7/16 Martín Marqués : >>> I had a problem with my main board and CPU and had to buy a new one. >>> Also memory, as the old computer had DDR2 memory. >>> >>> Th problem is that I see only 3.5Gb of memory, while I had a 4Gb bank >>> put in the Motherboard (the motherboard says it can handle up to 32Gb >>> or ram). >>> >>> I saw this in dmesg: >>> >>> [0.00] Memory: 3491896k/4702208k available (6177k kernel code, >>> 1053240k absent, 157072k reserved, 7004k data, 1004k init) >>> >>> Why do I get this: >>> >>> $ free >>> total used free sharedbuffers cached >>> Mem: 35134122591976 921436 0 8332 914584 >>> -/+ buffers/cache:16690601844352 >>> Swap:0 0 0 >>> >> >> There are several possible factors: >> 1) Integrated video card, which uses system memory. >> 2) Peripheral devices I/O regions are mapped below 4 GB, shadowing the >> physical ram. > > Could it be that the Motherboard uses normal ram for the integrated > ati video device? There are 500Mb of video memory. Does it use normal > RAM? Are does the 500Mb less I have? > > $ lspci -vv > 00:00.0 Host bridge: Advanced Micro Devices [AMD] Family 12h Processor > Root Complex > Subsystem: ASUSTeK Computer Inc. Device 84c8 > Control: I/O- Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop- > ParErr- Stepping- SERR- FastB2B- DisINTx- > Status: Cap- 66MHz+ UDF- FastB2B- ParErr- DEVSEL=medium >>TAbort- SERR- Latency: 0 > > 00:01.0 VGA compatible controller: ATI Technologies Inc Device 9645 > (prog-if 00 [VGA controller]) > Subsystem: ASUSTeK Computer Inc. Device 84c8 > Control: I/O+ Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop- > ParErr- Stepping- SERR- FastB2B- DisINTx+ > Status: Cap+ 66MHz- UDF- FastB2B- ParErr- DEVSEL=fast >TAbort- > SERR- Latency: 0, Cache Line Size: 64 bytes > Interrupt: pin A routed to IRQ 42 > Region 0: Memory at c000 (32-bit, prefetchable) [size=256M] > Region 1: I/O ports at f000 [size=256] > Region 2: Memory at fef0 (32-bit, non-prefetchable) [size=256K] > Expansion ROM at [disabled] > Capabilities: > Kernel driver in use: radeon > Kernel modules: radeon > I'm not an expert on reading lspci output, but ATI Device 9645 seems to be the graphics engine in AMD A4 APUs. If it indeed is, then part of system memory is used as graphics buffer. size=256M seems to suggest that it has 256 MB allocated. Therefore, it might be responsible for a large chunk of that missing 512 MB. > >> Here is (old, but still valid) explanation given by IBM for their >> servers: >> https://www-947.ibm.com/support/entry/portal/docdisplay?lndocid=MIGR-4E4RRF >> and here's the same issue addressed by Microsoft: >> http://support.microsoft.com/kb/978610 >> I did not manage to find Red Hat's take on the issue. > > I now about this. I had, 6 years ago an intel board which mapped > memory below the 4Gb so I had about 800Mb less. I was really > disappointed with intel, as it had no fix, so I decided to never by an > intel product ever. > >> In short: Works as designed. You may try to disable devices you do not >> use to salvage some address space, but you won't reach 4096 MB >> available. > > As I said, if that's the design, I will just look for another provider > of Motherboards and ditch ASUS with Intel. > Seems that Asus has a bad day, as I just recently been bitten by the PCI bridge bug on their Asus E45M1-M PRO board. I had better luck with Gigabyte. > -- > Martín Marqués > select 'martin.marques' || '@' || 'gmail.com' > DBA, Programador, Administrador > -- > users mailing list > users@lists.fedoraproject.org > To unsubscribe or change subscription options: > https://admin.fedoraproject.org/mailman/listinfo/users > Guidelines: http://fedoraproject.org/wiki/Mailing_list_guidelines > Have a question? Ask away: http://ask.fedoraproject.org -- users mailing list users@lists.fedoraproject.org To unsubscribe or change subscription options: https://admin.fedoraproject.org/mailman/listinfo/users Guidelines: http://fedoraproject.org/wiki/Mailing_list_guidelines Have a question? Ask away: http://ask.fedoraproject.org
