Smartmontools + Highpoint hptrr?
I'm trying to monitor SATA drives connected to a HighPoint RocketRAID 2340 card (hptrr driver) using smartmontools, but it doesn't seem to work. Has anyone managed to get this to work? I'm using RELENG_8 / amd64 if that makes any difference. I've tried both the smartmontools port and a direct checkout from their SVN, and I'm getting very similar error messages. From the changelog and the manpage it looks like some work has been done to make this work. /Daniel Eriksson ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "freebsd-questions-unsubscr...@freebsd.org"
RE: ZFS and DMA read error
Mark Stapper wrote: > So I see. Could this be why I haven't had any read errors anymore? > (After the zpool scrub that is) Possibly, but in that case the SMART selftest should pass also. Have you tried a selftest after you did the scrub? /Daniel Eriksson ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "freebsd-questions-unsubscr...@freebsd.org"
RE: ZFS and DMA read error
Mark Stapper wrote: > Yeah, i did the long SMART selftest three times now, each of which it > failed on the same LBA address. I assume 'smartctl -a /dev/adX' reports that the read test failed at LBA XXX something? > Why would I want to clear my driver before I run these tests? In this case it's not really clearing the drive you are aiming for, it is to write to every sector. If you have a failed sector (which you do), writing to it will force the drive firmware to remap the sector. As far as I know, most drives will not remap an unreadable sector until it is written to. /Daniel Eriksson ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "freebsd-questions-unsubscr...@freebsd.org"
RE: ZFS and DMA read error
Mark Stapper wrote: > People are REALLY pushing spinrite lately... I did get it though, just to try it. SpinRite is OK but it hasn't been updated in ages. It does not work on large drives. 250GB works, 1TB does not. Haven't tried it on 500GB drives. If I were you I would 'zpool offline ...' the offending drive, rewrite the entire drive with 'dd if=/dev/zero ...' and then run a SMART selftest on it using smartmontools ('smartctl -t long /dev/adX'). When you 'zpool online ...' the drive ZFS will resilver it for you. After doing all of this I would then run a 'zpool scrub ...'. If the scrub finishes without checksum errors and without any ATA-related errors the drive is probably in good enough condition to keep using, but watch out for more ATA errors. If the drive is dying it won't be long before it starts to generate more ATA errors. /Daniel Eriksson ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "freebsd-questions-unsubscr...@freebsd.org"
RE: No Device Node assigned for HD?
> You have two options: > 1. Use the RR2310 BIOS screen (or hptraidconf from inside FreeBSD) to > initialize the drive and create a single drive JBOD array with it. > 2. Connect the drive to a header on your motherboard and create a > partition table on it, then reconnect it to your RR2310 card. I would suggest doing #2 above if you don't plan on using the drive as part of a RR2310-controlled array. If you add a partition table and let your RR2310 card treat it as "Legacy" then the drive can be moved around freely between motherboard connectors and RR2310 connectors. If you do #1 then you need a RocketRAID card to access the data on the drive. /Daniel Eriksson ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "freebsd-questions-unsubscr...@freebsd.org"
RE: No Device Node assigned for HD?
Rob wrote: > The only difference I've found is that in the RocketRAID BIOS, the 3 > 500GB drives are recognized with a Legacy Status, whereas the 1TB is > recognized as New Status. Not sure what that means or how to > change it. Your "problem" is that the old drives you have hooked up to the RocketRAID card all have a partition table. When the RR BIOS sees that partition table it assumes it is a "Legacy" drive and exposes it to the OS as a single drive JBOD array. Your brand new 1 TB drive has nothing on it, and your RocketRAID card is waiting for you to initialize it and create an array with it before exposing it to the OS. You have two options: 1. Use the RR2310 BIOS screen (or hptraidconf from inside FreeBSD) to initialize the drive and create a single drive JBOD array with it. 2. Connect the drive to a header on your motherboard and create a partition table on it, then reconnect it to your RR2310 card. Partition table example: # gpart create -t GPT adXX # gpart add -b YY -s ZZ -t freebsd-ufs adXX /Daniel Eriksson ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "freebsd-questions-unsubscr...@freebsd.org"
RE: "Fixing" a RAID
Ryan Coleman wrote: > Jun 4 23:02:28 testserver kernel: ar0: 715425MB RAID5 (stripe 64 KB)> status: READY > Jun 4 23:02:28 testserver kernel: ar0: disk0 READY using ad13 at ata6-slave > Jun 4 23:02:28 testserver kernel: ar0: disk1 READY using ad16 at ata8-master > Jun 4 23:02:28 testserver kernel: ar0: disk2 READY using ad15 at ata7-slave > Jun 4 23:02:28 testserver kernel: ar0: disk3 READY using ad17 at ata8-slave > Jun 4 23:05:35 testserver kernel: g_vfs_done():ar0s1c[READ(offset=501963358208, length=16384)]error = 5 > ... My guess is that the rebuild failure is due to unreadable sectors on one (or more) of the original three drives. I recently had this happen to me on an 8 x 1 TB RAID-5 array on a Highpoint RocketRAID 2340 controller. For some unknown reason two drives developed unreadable sectors within hours of each other. To make a long story short, the way I "fixed" this was to: 1. Used a tool I got from Highpoint tech-support to re-init the array information (so the array was no longer marked as broken). 2. Unplugged both drives and hooked them up to another computer using a regular SATA controller. 3. One of the drives was put through a complete "recondition" cycle(a). 4. The other drive was put through a partial "recondition" cycle(b). 5. I hooked up both drives to the 2340 controller again. The BIOS immediately marked the array as degraded (because it didn't recognize the wiped drive as part of the array), and I could re-add the wiped drive so a rebuild of the array could start. 6. I finally ran a "zpool scrub" on the tank, and restored the few files that had checksum errors. (a) I tried to run a SMART long selftest, but it failed. I then completely wiped the drive by writing zeroes to the entire surface, allowing the firmware to remap the bad sectors. After this procedure the long selftest succeeded. I finally used a diagnostic program from the drive vendor (Western Digital) to again verify that the drive was working properly. (b) The SMART long selftest failed the first time, but after running a surface scan using the diagnostic program from Western Digital the selftest passed. I'm pretty sure the diagnostic program remapped the bad sector, replacing it with a blank one. At least the program warned me to back up all data before starting the surface scan. Alternatively I could have used dd (with offset) to write to just the failed sector (available in the SMART selftest log). If I were you I would run all three drives through a SMART long selftest. I'm sure you'll find that at least one of them will fail the selftest. Use something like SpinRite 6 to recover the drive, or use dd / dd_rescue to copy the data to a fresh drive. Once all three of the original drives pass a long selftest the array should be able to finish a rebuild using a fourth (blank) drive. By the way, don't try to use SpinRite 6 on 1 TB drives, it will fail halfway through with a division-by-zero error. I haven't tried it on any 500 GB drives yet. /Daniel Eriksson ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
RE: FreeBSD + ZFS on a production server?
Anders Häggström wrote: > I plan to install a web server for production use and ZFS looks very > interesting, especially since it has built-in support for RAID and > checksum. ZFS is very nice, but slightly over-hyped imho. However, some of the hype is warranted and for some use cases ZFS is a much better fit than UFS. Despite what Wojciech Puchar says, ZFS checksumming can be very useful. I recently had two drives in a hardware RAID-5 array (8 x 1 TB on a Highpoint RocketRAID 2340) develop unreadable sectors seemingly at the same time. I'm not sure what caused it but the end result was a broken/unavailable array. To make a long story short I managed to get the drives to remap the bad sectors and bring the array back online. Since I had ZFS on the array I didn't have to wait for fsck to run (takes a very long time on a 7 TB array and requires a LOT of memory to even work), and after the pool had been scrubbed I had a list of files with bad checksums that I could restore from backup. With UFS I would have had silent data corruption. Beware, there have been reports of mmap not working properly together with ZFS. I'm not sure if this is still a problem and if it would affect a typical web server. It does not seem to affect any of my fileservers (exporting NFS). /Daniel Eriksson ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
RE: raid6 on freebsd7 only showing 61GB instead of 4TB
Oliver Howe wrote: > I bought a new storage server and installed freebsd7 onto it. > it came with two raid partitions, one of 32GB which i used > for the o/s and one of 4.7TB which i am planning to use as a > nfs partition. everything went fine during the install, fdisk > said that there was 4.7TB on the second partition which i > labelled "/export". but when the machine booted up and i did > df -h it said that that partition only has 61GB and not 4.7TB As others have pointed out, fdisk is not able to handle partitions this big. You need to: 1. umount /export 2. dd if=/dev/zero of=/dev/da1 bs=64k count=1 3. gpt create /dev/da1 4. gpt add /dev/da1 5. newfs -O2 -U /dev/da1p1 6. edit your /etc/fstab and change /dev/da1s1d to /dev/da1p1 7. mount /export 8. be happy! The GENERIC kernel comes with GEOM_PART_GPT support so there is no need to load any kernel modules or recompile your kernel to get this to work. (Step #2 above is probably overkill. It erases the old disklabel so that your /dev/da1?? devices disappear.) Beware that running fsck on a 4.7TB partition will take a REALLY long time. If you run FreeBSD 7 in 64 bit mode (amd64), and you really should with 16 GB of memory, then I would recommend using ZFS instead of UFS. For ZFS you would do something like this: 1. umount /export 2. dd if=/dev/zero of=/dev/da1 bs=64k count=1 3. zpool create tank /dev/da1 4. edit /boot/loader.conf and add something like this: vm.kmem_size="1024M" vm.kmem_size_max="1024M" vfs.zfs.arc_max="512M" vfs.zfs.prefetch_disable=1 5. edit /etc/rc.conf and add zfs_enable="YES" 6. reboot 7. be happy! (With 16 GB memory you can probably use larger values for slightly better performance in step #4 above.) /Daniel Eriksson ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
RE: Promise RAID array and mounting questions
Trevor Hearn wrote: > I can see the promise array, it shows us as DA0. The problem > is what is > listed in the Dmesg stream shows the right amount of storage > space on the > array, but when I use sysinstall, I cannot mount that space > as the full > volume. I don't know if I am missing information on what to put for > geometry, but anything I put does not get accepted as usable. You will not be able to use sysinstall for this, the array is too big. You should use gpt instead. Do you want the entire array as one partition? # gpt create /dev/da0 # gpt add /dev/da0 # newfs -O2 -U /dev/da0p1 (or 'newfs -O2 -U -i 524288 /dev/da0p1' if you have mostly multi-megabyte files) If you want multiple partitions you will have to pass a size (in sectors) as a parameter to 'gpt add'. Beware that you might not be able to fsck the filesystem because of its size. I'm not sure how well gjournal handles fsck of large filesystems. Personally I'm going with ZFS for my next large array (8x750GB). /Daniel Eriksson ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
RE: smart_host on sendmail min config
Why are you not using the supplied scripts and configuration knobs to start Sendmail? It works just fine for 99% of FreeBSD users (my guess), so what makes your environment so special that your needs cannot be met without custom scripts? /Daniel Eriksson ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
RE: smart_host on sendmail min config
Aryeh Friedman wrote: > What else do I need to add to this to make it work (i.e. send all mail > via mx1.optonline.net): If your hostname is myhostname.mydomain.tld, then do the following: * cd /etc/mail * make all * edit myhostname.mydomain.tld.mc (it was created by the previous step) and change the smart host line like you already have done: define(`SMART_HOST', `mx2.optonline.net')dnl * make all install restart * try to send mail again /Daniel Eriksson ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
Migrating a file system with minimal downtime
I want to migrate a file system containing multiple jails from a small drive to a large (RAID-1) array. I want to do this with minimal downtime. Simply shutting down the jails and using dump/restore to move the file system takes too long, but what if I do it in several steps like this: 1. "dump -0 -L -f - /usr/jails | restore -rf -" to dump the live file system at level 0 2. shut down the jails 3. unmount the original file system 4. "dump -1 -f - /usr/jails | restore -rf -" to dump any changes since the first dump 5. remount the new file system in the proper location 6. restart the jails This should work, right? Or am I missing something? (One of the jails is a mysql server, the rest are www servers.) /Daniel Eriksson ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
RE: php5-extensions and xorg libraries
Zbigniew Szalbot wrote: > I will appreciate any help. When trying to install php5-extensions, I > get the following error: If you really don't need the xorg-libraries (true for most web servers for example) then there's an easy way to avoid them: 1. install all your ports with "make -DWITHOUT_X11 install clean" 2. add "'*' => 'WITHOUT_X11=1'," to the make args in pkgtools.conf (or whatever the format is in the file) to prevent portupgrade from adding the xorg-libraries in the future This works just fine for a typical webserver (analog, webalizer, php with the gd extension, ...). I just installed a new webserver with multiple jails to keep the different sites compartmentalized, and avoiding the xorg-libs saved me a ton of space and compile time. /Daniel Eriksson ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
RE: Upper limit on mount points?
Bill Moran wrote: > We have some systems with a lot of jails. We're using ezjail, which results > in a lot of nullfs mounts. In combination with some other nullfs tricks I'm > using on this system, I'm a bit concerned that we're going to hit some sort > of limit on the number of mountpoint. > > Google hasn't been much help, and I thought I'd ask here before I dug in to > the sourcecode. I'm wondering if there's an upper limit (either hard or > practical) on the number of mountpoints on a system? I have a box running 6-STABLE that has 700+ nullfs mounts, and it runs just fine. /Daniel Eriksson ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
RE: Why is 'disklabel'ng a new drive so difficult?
I've started using GPT for everything but the system disk. To add a GPT table to a new disk and create a partition in that table that spans the entire disk, all you need to do is this: # gpt create /dev/adX # gpt add /dev/adX # newfs -O2 -U /dev/adXp1 Done! /Daniel Eriksson ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
RE: gjournal and zfs, questions
Ivan Voras wrote: > GJournal aready is: > http://www.freebsd.org/cgi/cvsweb.cgi/src/sys/geom/journal/g_j > ournal.c?rev=1.9&content-type=text/x-cvsweb-markup Thank you for the info! zfs looks more interesting, but having gjournal in the tree makes it more convenient. Any guess on when zfs will hit the regular CVS tree? /Daniel Eriksson ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
gjournal and zfs, questions
Is gjournal and/or zfs stable enough for production-like usage yet? I've got a 3.5TB filesystem that I'd rather not have to fsck in case of a crash or power outage. I can live with running CURRENT on this box as long as I know the filesystem is stable. (I'd rather run RELENG_6, but zfs is only for current afaik.) When will gjournal and zfs be committed to the CVS tree? Will either of them be merged to STABLE? /Daniel Eriksson ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
RE: PHP Session Errors
Beecher Rintoul wrote: > I'm trying to get a new php install configured, but I'm > getting the following > error when calling session_start () What you probably want is all of the common extensions. Check out /usr/ports/lang/php5-extensions for a meta-port that will install what you need. /Daniel Eriksson ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
RE: Forwarding mail to another server
Pat Maddox wrote: > I wanted to try Daniel's method, using SMART_HOST. So I removed the > ssmtp port and, made the changes back to /etc/mail/mailer.conf, and > followed his instructions. None of the messages go through, they > always end up on my local machine with the message "Failed to route > address." Any ideas on what to do? Did you actually uncomment the line? If it starts with "dnl" it is considered a comment. Did you preserve the `' around the names? Can you "telnet mailrelay.myisp.com 25" and have it reply properly? You can also try to bump up the log level and see if /var/log/maillog gives you any more hints. There are several ways to bump the loglevel, but I usually just add the following line to the .mc file (and "make cf install restart"): define(`confLOG_LEVEL', `25') Another thing you could try is to use square brackets around the mailrelay, like so: define(`SMART_HOST', `[mailrelay.myisp.com]') This will force Sendmail to connect directly to that host, instead of doing a DNS lookup for MX pointers. /Daniel Eriksson ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
RE: Forwarding mail to another server
Pat Maddox wrote: > For example, if I did > mail [EMAIL PROTECTED] > > And sent a message, my machine would send that to mail.bresnan.net and > have them pass it on. Is it possible to do that? If so, how? cd /etc/mail make all this will create a FQDN.mc file (f.e. myhost.mydomain.com.mc) edit this file and uncomment the SMART_HOST line and insert an SMTP server that will relay mail for you example: define(`SMART_HOST', `mailrelay.myisp.com') make cf install restart Done! /Daniel Eriksson ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
Polling mode vs. net.isr.enable
Does the value of net.isr.enable (0 or 1) have any effect when running an interface in polling mode? If I understand correctly, with net.isr.enable set to 1 the processing of incoming packets are handled directly in the ISR (interrupt service routine?) instead of being deferred to the ithread. However, polling mode means no interrupts so the ISR is never called(?). /Daniel Eriksson ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
Dummynet traffic shaping question (TCP-ACK prioritization)
(question at the end) I have a server that sits on a medium speed link (10Mbit, full duplex) that under certain network loads starts to show what looks like TCP-ACK delay problems. At full upstream saturation the downstream speed is reduced. I modded the firewall rules to prioritize TCP-ACKs into one queue and all other outgoing traffic into another queue. Something like this: ${fwcmd} pipe 1 config ${fwcmd} queue 1 config pipe 1 weight 100 ${fwcmd} queue 2 config pipe 1 weight 1 # Route all outgoing TCP traffic with the ACK flag through the high priority queue ${fwcmd} add queue 1 tcp from any to any out via ${ext_if} tcpflags ack iplen 0-80 # Route all other (established) outgoing TCP traffic through the low priority queue ${fwcmd} add queue 2 tcp from any to any out via ${ext_if} established Looking at the output of 'ipfw show' seems to indicate the queues are getting the packets they should get: 00100 1738731 69778250 queue 1 tcp from any to any out via em0 tcpflags ack iplen 0-80 00200 5133634 7689253633 queue 2 tcp from any to any out via em0 established Even though everything looks OK, the results have not been what I hoped for (same problem with downstream speed during full upstream saturation). My question is: Do I need to tell the pipe how fat it is (${fwcmd} pipe 1 config bw 10Mbit/s) to get the queue prioritization to work properly, or is it OK to leave out the speed and just let it run full tilt? /Daniel Eriksson ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
RE: Sharing directories with jails
Emanuel Strobl wrote: > You can also use nullfs (man (8) mount_nullfs). It's slow and > not certified to be bugfree but I never had any problems and > especially for centralized ports very useful. What has given you the idea that nullfs is slow? I'm using it extensively and have not noticed any significant slowdown. Under what usecase(s) is it slow? (My usage is mainly for medium to large files, with <200 files per directory.) /Daniel Eriksson ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
ciss and SmartArray 5300, performance expectations?
Hi! I have been offered to buy a Compaq SmartArray 5300 (not sure exactly what version, I think it's the 5302/64) and some discs (14 x 36GB) that is no longer being used after an upgrade. I'm not getting it for free, so I'm only interested if it actually provides decent performance. What is the word on this controller and the support for it in FreeBSD 5.3/6.0? I understand it is supported through the ciss driver, but I have been unable to find any recent detailed discussions on how well this driver works. I did find a discussion about the lack of support for TCQ from 2002, but that's about it as far as detailed discussions go. Also, can I expect the 5300 to play nice with my non-Compaq server? It uses a Tyan Tiger MPX motherboard so it has 64bit/66MHz PCI, but I'm worried that maybe the card wants to live in a Compaq Proliant machine and won't work properly in my server (with regards to building arrays from BIOS and similar tasks). How about firmware upgrades of the card and/or the discs? Can that be done without a Proliant server? I would appreciate any information! /Daniel Eriksson ___ [EMAIL PROTECTED] mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
RE: cue images
Jeremy Faulkner wrote: > What international standard describes their format? > > Windows is not a standard. Many times a de-facto standard is just as important/valid as a "real" standard. /Daniel Eriksson ___ [EMAIL PROTECTED] mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
RE: fsck'ing a Vinum RAID5 volume (and a stale drive)
Benjamin P. Keating wrote: > Is > this approach correct? does this do anything productive or just forces > the state label to change and do nothing to the drives? I don't feel > confident that it did anything and Im having a VERY hard time finding > documentation on this. Let me give you an example of a valid setstate use case: One of my servers has a LOT of discs. For some reason I suffer from interrupt storms during device probing (started after I added a second Highpoint RocketRAID 454 to the machine). These storms sometime prevent the ata code from detecting all the discs. If this happens to be a disc that is part of a RAID-0 array, then when vinum starts up it detects that one of the discs have disappeared and (correctly) marks the array as crashed. There is no "proper" way to recover from a crashed RAID-0 array - your data is normally lost forever. However, in this specific case I _know_ that nothing has been written to the discs, so once I get the missing disc back online I can use setstate to change the array status from crashed to up, confident that no data has actually been lost. There are a few other use cases for setstate, but they are all (?) outside of the normal procedures for using and maintaining RAID arrays. /Daniel Eriksson ___ [EMAIL PROTECTED] mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
RE: fsck'ing a Vinum RAID5 volume (and a stale drive)
Benjamin P. Keating wrote: > I've found on the net that I can switch the state by doing: > > $ vinum setstate up backup.p0 backup.p0.s3 Ouch, this is a bad move. You just told vinum to start using the stale (=out of date data) disc as if it was up to date and nothing was wrong with it. Basically you have trashed your data, possibly beyond repair. Why was the disc in a stale state? If it developed bad blocks that could not be remapped, then vinum marked it as stale and continued to use the other discs in degraded mode (just as it should). Even if the disc did not break (maybe just connection problems or something), once vinum marked it as stale any further writing to the array would immediately invalidate the data on the disc; which means the only way to bring it back up would be to go throug a proper rebuild of the data. > I rebooted, the state is "up" so I unmounted the volume to fsck it. Is > this approach correct? does this do anything productive or just forces > the state label to change and do nothing to the drives? I don't feel > confident that it did anything and Im having a VERY hard time finding > documentation on this. Your approach is not correct. You should have paid attention to the vinum manpage which says this about setstate: "This bypasses the usual consistency mechanism of vinum and should be used only for recovery purposes. It is possible to crash the system by incorrect use of this command." It is unfortunate that the manpage says "for recovery", since people can misunderstand and think you can recover from a crashed disc. setstate should not need to be used during normal operation, even if a disc in a RAID-5 array crashes. > Im assuming I'll want to answer yes to at least some of those. Can I > say yes to all of them? What errors should I say no to? I have no idea > whats bad bad and whats correctable. An fsck of a degraded RAID-5 array should not normally have any errors. The errors you are seeing is because you have forced out-of-date (stale) data into the middle of the filesystem, messing up pretty much the entire filesystem. > Now. because this is a raid5 volume, are some of these fsck prompts > false positives? ie; fsck is giving a error but really it's fine as > it's raid5? No, that is not how RAID-5 works. RAID-5 protects the integrity of the filesystem by ensuring that the stored data can be read/written even if one disc fails. RAID arrays work at a level below the filesystem, and they generally don't know anything about the actual filesystem. Your only hope now is if you haven't actually allowed anything to be written to the array. If fsck changed things around then you are probably out of luck. IF nothing has been written, then reset the failed disc to the "stale/down" state (which should put the array in degraded mode hopefully) and then try fsck again. If you are lucky you should see no errors at all. If the disc isn't physically broken, then the proper way to get your array back to the original "all up" state, you should run a "vinum start backup.p0.s3". I'm not sure if this will rebuild all the data automatically (it does for RAID-1 arrays), but if it doesn't then I guess you also need to run a "vinum rebuildparity backup.p0". /Daniel Eriksson ___ [EMAIL PROTECTED] mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
Multiple choices for RAID-0, best performance?
With the recent addition of a geom-enabled vinum, and the discussed (but not yet functional?) geom stripe module, we now have 4 different choices for software RAID-0 on ATA devices (the other two are regular vinum and ataraid). I'm currently using a mix of vinum and ataraid on one of my servers, but I haven't been in a position to do any proper performance measurements yet (it's a production server). Are there any obvious performance differences between the different ways of doing RAID-0? Any difference in resource usage? Also, will the RAID-0 performance of ataraid be affected a lot depending on what controller is used? For example would there be any performance difference (transfer speed and/or system resource usage) between ataraid on a HPT374-based card (such as the Highpoint RocketRAID 454) and one of the new Promise cards (which we have really good drivers for afaik). Daniel Eriksson ___ [EMAIL PROTECTED] mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"