Re: raid recomendation
On 12/7/2012 5:48 PM, Aaron Toponce wrote: A RAID-1 will outperform a parity-based RAID using the same disks every time, due to calculating the parity. This hasn't been true for almost a decade. Even the slowest of modern single core x86 CPUs have plenty of excess horsepower for soft RAID parity, as do ASICs on hardware RAID solutions. There are two big problems today with parity arrays. The first is the IO load and latency of read-modify-write cycles which occur when partial stripes are updated. Most write IO is small and random, typically comprising over 95% of writes for most typically workloads, mail, file, LDAP, SQL servers for example. Streaming applications such as video are an exception as they write full stripes. Thus for every random write, you first must, at a minimum, read two disks (data chunk/strip and parity chunk/strip) and write back to both with the new data chunk and parity chunk. This is with an optimized algorithm. The md/RAID driver has some optimizations to cut down on RMW penalties. Many hardware RAID solutions read then write the entire stripe for scrubbing purposes (i.e. write all disks frequently so media errors are caught sooner rather than later). This is a data integrity feature of higher end controllers. This implementation is much slower due to all the extra IO and head movement, but more reliable. The second is that failed drive rebuilds take FOREVER as all disks are being read in parallel and parity calculated for every stripe, just to rebuild one disk. Even a small count 2TB drive RAID6 array can take 12-24 hours to rebuild. The recommended max array drive count for RAID5/6 are 4 and 8 drives respectively. One of the reasons for this BCP is rebuild time. With RAID10 rebuild time is a constant, as you're simply copying all the sectors from one drive to another. A 60x2TB drive RAID10 rebuild will take about 5 hours with low normal workload IO hitting the array. Further, striping across two mirrors will give increased performance that parity-based RAID cannot achieve. A parity array actually has superior read speed vs a RAID10 array of the same total spindle count because there are more data spindles. An 8 drive RAID6 has 6 data spindles, whereas an 8 drive RAID10 only has 4. Write performance, however, as I mentioned, is an order of magnitude slower due to RMW. Lastly, you can suffer any sort of disk failures, provided all mirrors in the stripe remains in tact. You mean any number not sort. Yes, with RAID10 you can lose half the drives in the array as long as no two are in the same mirror pair. I wouldn't bank on this though. Many drive dropouts are not due to problems with the drives, but with the backplanes and cabling. When that happens, if you've not staggered your mirrors across HBAs, cables, and cabinets (which isn't possible with RAID HBAs), you may very well lose two drives in the same mirror. 1: http://zfsonlinux.org Just my $.02. And that sums up the value of your ZFS on Linux recommendation, quite well. Being a fanboy is fine. Run it yourself. But please don't recommend unsupported, out of tree, difficult for the average Debian user to install, software, for a general purpose storage solution. Good hardware RAID is relatively cheap, Linux has md/RAID which isn't horrible for most workloads, and there are plenty of high quality Linux filesystems to meet most needs, with EXT4 for casual stuff, JFS and XFS for heavy duty, though XFS is a much better choice for many reasons; the big one being that it's actively developed, whereas JFS is mostly in maintenance only mode. -- Stan -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/50c30205.2050...@hardwarefreak.com
RE : Re: raid recomendation
Envoyé depuis un mobile SamsungStan Hoeppner s...@hardwarefreak.com a écrit :On 12/7/2012 5:48 PM, Aaron Toponce wrote: A RAID-1 will outperform a parity-based RAID using the same disks every time, due to calculating the parity. This hasn't been true for almost a decade. Even the slowest of modern single core x86 CPUs have plenty of excess horsepower for soft RAID parity, as do ASICs on hardware RAID solutions. There are two big problems today with parity arrays. The first is the IO load and latency of read-modify-write cycles which occur when partial stripes are updated. Most write IO is small and random, typically comprising over 95% of writes for most typically workloads, mail, file, LDAP, SQL servers for example. Streaming applications such as video are an exception as they write full stripes. Thus for every random write, you first must, at a minimum, read two disks (data chunk/strip and parity chunk/strip) and write back to both with the new data chunk and parity chunk. This is with an optimized algorithm. The md/RAID driver has some optimizations to cut down on RMW penalties. Many hardware RAID solutions read then write the entire stripe for scrubbing purposes (i.e. write all disks frequently so media errors are caught sooner rather than later). This is a data integrity feature of higher end controllers. This implementation is much slower due to all the extra IO and head movement, but more reliable. The second is that failed drive rebuilds take FOREVER as all disks are being read in parallel and parity calculated for every stripe, just to rebuild one disk. Even a small count 2TB drive RAID6 array can take 12-24 hours to rebuild. The recommended max array drive count for RAID5/6 are 4 and 8 drives respectively. One of the reasons for this BCP is rebuild time. With RAID10 rebuild time is a constant, as you're simply copying all the sectors from one drive to another. A 60x2TB drive RAID10 rebuild will take about 5 hours with low normal workload IO hitting the array. Further, striping across two mirrors will give increased performance that parity-based RAID cannot achieve. A parity array actually has superior read speed vs a RAID10 array of the same total spindle count because there are more data spindles. An 8 drive RAID6 has 6 data spindles, whereas an 8 drive RAID10 only has 4. Write performance, however, as I mentioned, is an order of magnitude slower due to RMW. Lastly, you can suffer any sort of disk failures, provided all mirrors in the stripe remains in tact. You mean any number not sort. Yes, with RAID10 you can lose half the drives in the array as long as no two are in the same mirror pair. I wouldn't bank on this though. Many drive dropouts are not due to problems with the drives, but with the backplanes and cabling. When that happens, if you've not staggered your mirrors across HBAs, cables, and cabinets (which isn't possible with RAID HBAs), you may very well lose two drives in the same mirror. 1: http://zfsonlinux.org Just my $.02. And that sums up the value of your ZFS on Linux recommendation, quite well. Being a fanboy is fine. Run it yourself. But please don't recommend unsupported, out of tree, difficult for the average Debian user to install, software, for a general purpose storage solution. Good hardware RAID is relatively cheap, Linux has md/RAID which isn't horrible for most workloads, and there are plenty of high quality Linux filesystems to meet most needs, with EXT4 for casual stuff, JFS and XFS for heavy duty, though XFS is a much better choice for many reasons; the big one being that it's actively developed, whereas JFS is mostly in maintenance only mode. -- Stan -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/50c30205.2050...@hardwarefreak.com
Re: raid recomendation
On Thu, Dec 06, 2012 at 01:18:38PM -0300, Roberto Scattini wrote: hi, i have a new dell r720 server with 5 600gb disks. his function will be a postgresql server (the size of the databases is really small with 600gb we should be fine for a long time). which raid configuration would you recommend? i was thinking in raid 5 with all five disks but i am not a expert. i prefer redundandcy against size (i mean, i can sacrifice space). and i dont want performance degradation for doing raid with an incorrect number of disks. I'll be the first one in this thread to recommend ZFS [1]. With 5 disks, I would personally do a RAID-1+0, with a hot spare. A RAID-1 will outperform a parity-based RAID using the same disks every time, due to calculating the parity. Further, striping across two mirrors will give increased performance that parity-based RAID cannot achieve. Lastly, you can suffer any sort of disk failures, provided all mirrors in the stripe remains in tact. 1: http://zfsonlinux.org If you must absolutely do a parity-based RAID, then I would suggest a 5-disk RAIDZ-1 without a hot spare. It's best practice to use the power of two, plus parity for your number of disks. In this case, it will give you the best performance, decent space, and allow for 1 disk failure. Further, I would recommend the investment in two Intel 300-series SSDs. You can then partition the SSDs giving 1 GB on each in a mirrored ZIL, and the rest to a striped L2ARC. For a PostgreSQL DB, you will see immensive performance gains that you cannot achieve with Linux-based software RAID and filesystems. And, because ZFS is also a volume manager, there is no need for LVM and the cache troubles it's plagued with [2]. 2: http://serverfault.com/questions/279571/lvm-dangers-and-caveats If interested, I've been blogging on this very topic. You can see the relevent posts to your setup here: * Installing ZFS on Debian: http://pthree.org/?p=2357 * The ZIL: http://pthree.org/?p=2592 * The ZFS ARC: http://pthree.org/?p=2659 Just my $.02. -- . o . o . o . . o o . . . o . . . o . o o o . o . o o . . o o o o . o . . o o o o . o o o pgpDxZ39JLKBv.pgp Description: PGP signature
raid recomendation
hi, i have a new dell r720 server with 5 600gb disks. his function will be a postgresql server (the size of the databases is really small with 600gb we should be fine for a long time). which raid configuration would you recommend? i was thinking in raid 5 with all five disks but i am not a expert. i prefer redundandcy against size (i mean, i can sacrifice space). and i dont want performance degradation for doing raid with an incorrect number of disks. thanks in advance! -- Roberto Scattini
Re: raid recomendation
On Thu, Dec 6, 2012 at 2:18 PM, Roberto Scattini roberto.scatt...@gmail.com wrote: hi, i have a new dell r720 server with 5 600gb disks. his function will be a postgresql server (the size of the databases is really small with 600gb we should be fine for a long time). which raid configuration would you recommend? i was thinking in raid 5 with all five disks but i am not a expert. Hi Roberto, A RAID 5 volume including the 5 drives should work fine for you. A question you should consider when you are taking this decision is: how fast can you replace a failed drive? You must know that in the meanwhile you'd be vulnerable, since an additional failure could lead to data loss. If you can't exchange the drive fast enough, maybe you should use an spare drive or thinking about RAID 6 (I've never used it though). i prefer redundandcy against size (i mean, i can sacrifice space). and i dont want performance degradation for doing raid with an incorrect number of disks. Another think you could consider is creating a separate volume for the OS. Particularly, I'd go for a 2 disks RAID 1 for the OS and a 3 disks RAID 5 for the database, since capacity isn't a problem. This configuration ensures that the database I/O traffic does not competes with the OS (when swapping and stuff). But that also depends on your hardware configuration. Best, thanks in advance! -- Roberto Scattini -- Pedro Eugênio Rocha
Re: raid recomendation
Le 15680ième jour après Epoch, Roberto Scattini écrivait: hi, i have a new dell r720 server with 5 600gb disks. his function will be a postgresql server (the size of the databases is really small with 600gb we should be fine for a long time). which raid configuration would you recommend? i was thinking in raid 5 with all five disks but i am not a expert. i prefer redundandcy against size (i mean, i can sacrifice space). and i dont want performance degradation for doing raid with an incorrect number of disks. In this case, avoid RAID5 and choose RAID1 or RAID10. Use 4 disks on the RAID10 array and the 5th as a spare one, for example. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/8738zjrur7@tourde.org
Re: raid recomendation
Hello, for databases its recommended to use raid10, cause of the speed regards Stefan
Re: raid recomendation
On 06/12/12 13:18, Roberto Scattini wrote: hi, i have a new dell r720 server with 5 600gb disks. his function will be a postgresql server (the size of the databases is really small with 600gb we should be fine for a long time). which raid configuration would you recommend? i was thinking in raid 5 with all five disks but i am not a expert. i prefer redundandcy against size (i mean, i can sacrifice space). and i dont want performance degradation for doing raid with an incorrect number of disks. thanks in advance! -- Roberto Scattini If you want performance and redundancy at cost to reduce your storage capacity think in RAID 10. Regards Federico -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/50c0d39a.2030...@uncu.edu.ar
Re: raid recomendation
Am Donnerstag, 6. Dezember 2012 schrieb Roberto Scattini: hi, i have a new dell r720 server with 5 600gb disks. his function will be a postgresql server (the size of the databases is really small with 600gb we should be fine for a long time). which raid configuration would you recommend? i was thinking in raid 5 with all five disks but i am not a expert. i prefer redundandcy against size (i mean, i can sacrifice space). and i dont want performance degradation for doing raid with an incorrect number of disks. Additionally to what has been written by the other posters: http://baarf.com :) -- Martin 'Helios' Steigerwald - http://www.Lichtvoll.de GPG: 03B0 0D6C 0040 0710 4AFA B82F 991B EAAC A599 84C7 -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/201212062022.04928.mar...@lichtvoll.de