RE: RAID Gone Wild - One Array Split Into Two
What branch is this patch supposed to apply against? I've tried 7.0-RELEASE and 7.1-RELEASE, and judging by the dates involved, it's somewhere in between the two. Or should I be asking this on the freebsd-hackers list, where that patch was originally posted? Alex Alex, This is known problem with FreeBSD and ICH7..ICH8..ICH9 chipsets. There is a patch for it: http://www.nabble.com/Vital-Patches-for-ataraid-with-Intel-Matrix-RAID-(ICH7)-td16179257.html I though didn't test the patch, and funny thing, I posted earlier today asking if anyone had tried out that patch. Hope this helps, keep me posted if you're able to bring it online. Thanks, Tamouh -Original Message- From: owner-freebsd-questi...@freebsd.org [mailto:owner-freebsd-questi...@freebsd.org] On Behalf Of Alex Kirk Sent: March 1, 2009 8:20 PM To: Jamie Cc: questi...@freebsd.org Subject: Re: RAID Gone Wild - One Array Split Into Two Does anyone have a clue how I can fix this, preferably while retaining my data? I could wipe the box if necessary, but I'd really prefer not to, as that would be a huge pain in the butt. Thanks, Alex Kirk I would begin by going into the raid BIOS at bootup to see what containers are now configured. If everything is hosed up in there the OS isn't going to be able to fix anything. - Jamie Sorry, should have already gone over this. The RAID BIOS is terrible - my options are Create Array, Delete Array, Reset Disk States, and Exit. It shows only the one array, but all four disks show as Offline Member in red there. I'm just concerned that if I reset the array or delete it, the state table (or whatever other magic is involved in making RAID work) will get hosed up and the data will be unrecoverable. Alex This message was sent using IMP, the Internet Messaging Program. ___ 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 This message was sent using IMP, the Internet Messaging Program. ___ 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
RAID Gone Wild - One Array Split Into Two
First off, I realize that this may be more of a lower-level hardware question than is appropriate to ask here, but I'm at a real loss, and have no idea who else to ask...so I apologize in advance if I'm being a pest. That said: I've got a FreeBSD 7.0/stable box that is used as the development server for a live system I administer. It recently crapped out on me (the dev box), and I realized that its power supply had kicked the bucket. After going out and replacing the power supply, it booted right back up, I ssh'd in, and when I ran my first userland command - w, FWIW - it froze up solid. I got one more SSH session in attempting to figure out WTF was going on before it wouldn't even log me in any more. After a couple of hard reboots, I decided to attach a monitor to it to see what was going on. It turns out that the RAID5 array on the system had really lost its mind - all four devices that were part of the array were listed as being offline, which of course meant that the system could no longer boot (as it was booting off of the RAID). The controller is an integrated Intel Matrix DHC7R, built onto the motherboard. I looked around the web a bit to try to figure out how to fix this, and ran across a couple of forum posts (which I can unfortunately no longer seem to find) suggesting that this particular controller was prone to an issue where hard power-downs would sometimes make the drives go offline, and that I needed to boot from CD to re-initialize them into their previous state. I tried first with an Ubuntu Linux CD I had handy - which promptly freaked out and dropped me into an emergency shell - and then the FreeBSD 7.0 boot-only disc. The latter was a bit more helpful, because I got this diagnostic: ar0: WARNING - parity protection lost, RAID5 array in DEGRADED mode ar0: 715418MB Intel MatrixRAID RAID5 (stripe 64KB) status: DEGRADED ar0: disk0 READY using ad4 at ata2-master ar0: disk1 READY using ad8 at ata4-master ar0: disk2 READY using ad6 at ata3-master ar0: disk3 DOWN no device found for this subdisk ar1: 715418MB Intel MatrixRAID RAID5 (stripe 64KB) status: BROKEN ar1: disk0 DOWN no device found for this subdisk ar1: disk1 DOWN no device found for this subdisk ar1: disk2 DOWN no device found for this subdisk ar1: disk3 READY using ad10 at ata5-master Now I can see that my problem is that I've somehow got *two* RAID devices, both improperly configured, whereas I'd only had one before. Does anyone have a clue how I can fix this, preferably while retaining my data? I could wipe the box if necessary, but I'd really prefer not to, as that would be a huge pain in the butt. Thanks, Alex Kirk This message was sent using IMP, the Internet Messaging Program. ___ 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: RAID Gone Wild - One Array Split Into Two
Does anyone have a clue how I can fix this, preferably while retaining my data? I could wipe the box if necessary, but I'd really prefer not to, as that would be a huge pain in the butt. Thanks, Alex Kirk I would begin by going into the raid BIOS at bootup to see what containers are now configured. If everything is hosed up in there the OS isn't going to be able to fix anything. - Jamie Sorry, should have already gone over this. The RAID BIOS is terrible - my options are Create Array, Delete Array, Reset Disk States, and Exit. It shows only the one array, but all four disks show as Offline Member in red there. I'm just concerned that if I reset the array or delete it, the state table (or whatever other magic is involved in making RAID work) will get hosed up and the data will be unrecoverable. Alex This message was sent using IMP, the Internet Messaging Program. ___ 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: RAID Gone Wild - One Array Split Into Two
This is why I love open-source mailing lists - you never know what sort of awesome you'll find! My question at this point, though, is how in the world I could actually apply this patch, seeing as how the system is in a non-bootable state. Is this something that's already been included in a development branch that I could go download? Or do I need to do something else? Thanks, Alex Alex, This is known problem with FreeBSD and ICH7..ICH8..ICH9 chipsets. There is a patch for it: http://www.nabble.com/Vital-Patches-for-ataraid-with-Intel-Matrix-RAID-(ICH7)-td16179257.html I though didn't test the patch, and funny thing, I posted earlier today asking if anyone had tried out that patch. Hope this helps, keep me posted if you're able to bring it online. Thanks, Tamouh -Original Message- From: owner-freebsd-questi...@freebsd.org [mailto:owner-freebsd-questi...@freebsd.org] On Behalf Of Alex Kirk Sent: March 1, 2009 8:20 PM To: Jamie Cc: questi...@freebsd.org Subject: Re: RAID Gone Wild - One Array Split Into Two Does anyone have a clue how I can fix this, preferably while retaining my data? I could wipe the box if necessary, but I'd really prefer not to, as that would be a huge pain in the butt. Thanks, Alex Kirk I would begin by going into the raid BIOS at bootup to see what containers are now configured. If everything is hosed up in there the OS isn't going to be able to fix anything. - Jamie Sorry, should have already gone over this. The RAID BIOS is terrible - my options are Create Array, Delete Array, Reset Disk States, and Exit. It shows only the one array, but all four disks show as Offline Member in red there. I'm just concerned that if I reset the array or delete it, the state table (or whatever other magic is involved in making RAID work) will get hosed up and the data will be unrecoverable. Alex This message was sent using IMP, the Internet Messaging Program. ___ 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 This message was sent using IMP, the Internet Messaging Program. ___ 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
portmanager looping on libtool on 6.2 - 6.3 upgrade
Hello All, I'm in the process of bringing a production web/mail server up to FreeBSD 7.0 from 6.2. After practicing the process on a non-production box set up in essentially the same manner, I discovered that the only major issue to look out for was the fact that I needed to hold back the upgrade of Python, since the CMS system running on the box will die if it doesn't have Python 2.4 specifically. So last night, I got the actual OS brought from 6.2 - 6.3, and then looked for a different method of port upgrading than I'd used on the development system, since even specifying ports to exclude via portmanager's -ip option didn't seem to work well when going from 6.3 - 7.0. I ran across some articles that told me that I could do ports one at a time, confirming upgrades of dependencies, with a command such as: portmanager editors/emacs -l -ui -f I was able to get some ports, such as bash, upgraded using this method. However, I've now reached a point where virtually everything I try to upgrade is now failing, because it's dependent on libtool, which is stuck in some sort of infinite upgrade loop. Attempting to upgrade it manually via the same command line as above (replacing editors/emacs with devel/libtool15), I end up with output like this: ok to update/rebuild /devel/libtool15 libtool-1.5.26 (yes/no/auto yes to all) [y/n/a] [y]? y [lots of trimmed-out building activity, followed by two more times asking me if it's OK to update/rebuild libtool] MGPMrUpdate 0.4.1_9 command: #1 of 14 cd /usr/ports/devel/libtool15 make -V OPTIONS reverting bsd.port.mk patch -=cd /usr/ports/Mk; patch -R /usr/local/share/portmanager/patch-bsd.port.mk-0.3.6; Hmm... Looks like a unified diff to me... The text leading up to this was: -- |--- /usr/ports/Mk/bsd.port.mk Tue Nov 8 01:02:51 2005 |+++ bsd.port.mkWed Nov 16 02:16:57 2005 -- Patching file bsd.port.mk using Plan A... Hunk #1 failed at 2049. 1 out of 1 hunks failed--saving rejects to bsd.port.mk.rej done rCreateInstalledDbVerifyContentsFile 0.4.1_9 error: @comment ORIGIN: not found in /var/db/pkg/bsdpan-MIME-tools-5.420/+CONTENTS bsdpan-MIME-tools-5.420 installation is corrupt! recomend running pkg_delete -f bsdpan-MIME-tools-5.420 then manually reinstalling this port Port Status Report forced mode 1 :libtool-1.5.26 /devel/libtool15 MISSING skipping libtool-1.5.26 /devel/libtool15 marked IGNORE reason: looping, 3rd attempt at make portmanager 0.4.1_9 INFO: finished with some ports not updated if --log was used see /var/log/portmanager.log I've tried simply going into /usr/ports/devel/libtool15 and running suod make install clean, and I end up with this output: === Installing for libtool-1.5.26 === Generating temporary packing list === Checking if devel/libtool15 already installed === libtool-1.5.26 is already installed You may wish to ``make deinstall'' and install this port again by ``make reinstall'' to upgrade it properly. If you really wish to overwrite the old port of devel/libtool15 without deleting it first, set the variable FORCE_PKG_REGISTER in your environment or the make install command line. *** Error code 1 Stop in /usr/ports/devel/libtool15. I'm very nervous about deinstalling it and reinstalling, or doing a forced reinstall, because so much appears to be dependent on libtool, and the last thing I want to do is bork a busy production box so bad that I have to physically go to the data center and hit up the console to fix it (especially since I'm not sure I could get in there over the weekend, and I'm flying out of town for the holidays on Monday morning). Does anyone have any advice on how to get this fixed? The system appears to be fully functional in the meantime, so it's not a huge rush to get things upgraded, but obviously I'd rather progress sooner rather than later. Alex Kirk This message was sent using IMP, the Internet Messaging Program. ___ 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
CARP-Like Solution With Machines On Different Networks?
Hello All, I'm attempting to put a redundant fail-over system in place for a machine that I manage for a non-profit organization of modest budget. For the time being, I'm most interested in having MySQL and HTTP connections roll over to a backup system in the event that the primary machine goes down for some reason, and then return control to the primary box once it returns - nothing particularly fancy. After doing some research on the matter, it looks like CARP would be a winning solution - but only if the backup system was on the same network segment as the primary box. Given that there's no money to colocate a second backup system at the same facility as the main machine (and protection against failure at the colo facility is one of the primary drivers for the failover setup), however, it looks like CARP wouldn't be useful. That said, are there any solutions which behave similarly to CARP that I could use for a pair of machines connected solely via the Internet? For now, I'd even be happy if there was some way to simply do TCP port-level proxying, so to speak (i.e. connections come in to a given machine, and are proxied to the main system if it's up, but go to the backup box if not)? Thanks in advance for any advice you can provide. Alex Kirk This message was sent using IMP, the Internet Messaging Program. ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
