Re: Raid disk delimma
Let's keep this on the list... also sprach Jack Schneider p...@dp-indexing.com [2009.09.21.1857 +0200]: Results from :mdadm -Asayes Just returns a prompt: # vgchange -a (gives argument needed message) vgchange -a y :gives 7 logical volume(s) in volume group Speeduke now active 'first seems to complete without error and the second (with my change) seems to indicate that all volume groups are found...hmmm! I'll take a shot into the dark: does it work if you add rootdelay=30 to the linux kernel boot line? E.g. kernel /boot/vmlinuz ro root=/dev/md0 rootdelay=30 This will incur a 30 second delay during boot. If that makes it work, you can experiment with lowering this value, but you will need to have some sort of sleep in there like that, because apparently your controllers need more time to make the disks available, and they do not properly convey to the kernel when they're ready. Cheers, -- .''`. martin f. krafft madd...@d.o Related projects: : :' : proud Debian developer http://debiansystem.info `. `'` http://people.debian.org/~madduckhttp://vcs-pkg.org `- Debian - when you have better things to do than fixing systems geld ist das brecheisen der macht. - friedrich nietzsche digital_signature_gpg.asc Description: Digital signature (see http://martin-krafft.net/gpg/)
Re: Raid disk delimma
On Mon, 21 Sep 2009 19:03:45 +0200 martin f krafft madd...@debian.org wrote: Let's keep this on the list... OK with me.. I normally just hit reply in Claws 3.5. You may get 2 if I use reply-all.. also sprach Jack Schneider p...@dp-indexing.com [2009.09.21.1857 +0200]: Results from : mdadm -Asayes Just returns a prompt: # vgchange -a (gives argument needed message) vgchange -a y :gives 7 logical volume(s) in volume group Speeduke now active 'first seems to complete without error and the second (with my change) seems to indicate that all volume groups are found...hmmm! I'll take a shot into the dark: does it work if you add rootdelay=30 to the linux kernel boot line? E.g. kernel /boot/vmlinuz ro root=/dev/md0 rootdelay=30 This will incur a 30 second delay during boot. If that makes it work, you can experiment with lowering this value, but you will need to have some sort of sleep in there like that, because apparently your controllers need more time to make the disks available, and they do not properly convey to the kernel when they're ready. Cheers, You mean in the grub boot edit function? At boot-up ? Where I put it, /boot/grub/menu.lst It didn't help... Same fault. Sorry about that.. It was not a problem until the 2.6.30 update I did about 9/14- 9/15 time frame... the problem came up after the next boot. I had to shut down for a storm... I don't trust UPS's for protection..8-( TIA, Jack -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: Raid disk delimma
also sprach Jack Schneider p...@dp-indexing.com [2009.09.20.0044 +0200]: I get to the Give root password for maintenance etc. place and when I do it will not allow me to find any logs. I can run dmesg which looks normal. Anything that that tries to use /var -locks. etc. fails, so it's unable to find /dev/md1 and contents..it seems to me... df -h finds /md0 but not /md1 Have a look at the output of mdadm -Es and compare that to the ARRAY lines /etc/mdadm/mdadm.conf. Anything different? What do you see when you run mdadm -Asayes vgchange -a ? -- .''`. martin f. krafft madd...@d.o Related projects: : :' : proud Debian developer http://debiansystem.info `. `'` http://people.debian.org/~madduckhttp://vcs-pkg.org `- Debian - when you have better things to do than fixing systems vulgarity is simply the conduct of other people. -- oscar wilde digital_signature_gpg.asc Description: Digital signature (see http://martin-krafft.net/gpg/)
Re: Raid disk delimma
Jack Schneider wrote: Hi, All Somewhere (about 9/15) in my enthusiasm for Debian, Testing up2date kernel 2.6.30-1-amd64,--updating I said (y) where I should have said (q). My system will not boot correctly. It halts with: fsck died with exit status 8 I have 4 disks, 2 in raid1, ie. 2 in use - 2 unused Disks are setup as /dev/md0 --/root (small-ish) partition and /dev/md1 much bigger containing /tmp /opt /var and several VirtualBox DVI partitions all under lvm2. I get to the Give root password for maintenance etc. place and when I do it will not allow me to find any logs. I can run dmesg which looks normal. Anything that that tries to use /var -locks. etc. fails, so it's unable to find /dev/md1 and contents..it seems to me... df -h finds /md0 but not /md1 I have tried to google/linux and man on mdadm etc. but ... nothing close... I don't know where to go from here... HELP! or directions, Please! TIA Jack Hi, really don't know if it's relevant in your situation, but I was once stuck with such a situation on a raid1 system (which worked correctly for a while). Turned out that the filesystem size was incorrect, and was overlapping with the superblocks (don't know how it happened, must have been at the raid creation time). I ran a e2fsck -cc (from a live-cd, takes a while), then resize2fs and finally fsck on every raid group. It does the trick, and the system has been running fine ever since (was ± one year ago.) Hope it helps, Tom -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: Raid disk delimma
On Sun, 20 Sep 2009 08:37:51 +0200 martin f krafft madd...@debian.org wrote: also sprach Jack Schneider p...@dp-indexing.com [2009.09.20.0044 +0200]: I get to the Give root password for maintenance etc. place and when I do it will not allow me to find any logs. I can run dmesg which looks normal. Anything that that tries to use /var -locks. etc. fails, so it's unable to find /dev/md1 and contents..it seems to me... df -h finds /md0 but not /md1 Have a look at the output of mdadm -Es and compare that to the ARRAY lines /etc/mdadm/mdadm.conf. Anything different? What do you see when you run mdadm -Asayes vgchange -a ? Hi,Martin Thanks for your reply... A little more info /home is also on a /dev/md1 partition. Re:mdadm -Es and compare that to the ARRAY lines /etc/mdadm/mdadm.conf. Anything different? There is 'no' difference in the array info UUID's match for both md0 md1. I have yet to try the other commands. need to see what they mean...8-) Jack -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Raid disk delimma
Hi, All Somewhere (about 9/15) in my enthusiasm for Debian, Testing up2date kernel 2.6.30-1-amd64,--updating I said (y) where I should have said (q). My system will not boot correctly. It halts with: fsck died with exit status 8 I have 4 disks, 2 in raid1, ie. 2 in use - 2 unused Disks are setup as /dev/md0 --/root (small-ish) partition and /dev/md1 much bigger containing /tmp /opt /var and several VirtualBox DVI partitions all under lvm2. I get to the Give root password for maintenance etc. place and when I do it will not allow me to find any logs. I can run dmesg which looks normal. Anything that that tries to use /var -locks. etc. fails, so it's unable to find /dev/md1 and contents..it seems to me... df -h finds /md0 but not /md1 I have tried to google/linux and man on mdadm etc. but ... nothing close... I don't know where to go from here... HELP! or directions, Please! TIA Jack -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org