Re: no devices listed in mdstat
/proc/mdstat exists. I can't create md arrays, as the utilities say that /dev/md0 is a bad argument. Just to be sure: what's the output of uname -a mkraid --version ? Egon Eckert
Re: raid 1 and kernel upgrade 2.2.12-2.2.14
this is quite interesting, since it shouldn't work : And, apropos, are any (stability) problems known with the famous raid-2.2.14-B1? BTW: which raidtools to use with it? Dangerous-0.90-2116 or the "ordinary" 19990824-0.90? The latter I guess.. Egon Eckert
Re: [new release] raidreconf utility
There's a new version of the raidreconf utility out. I call it 0.0.2. Isn't this what supposed 'LVM' to be about? (BTW there seem to be 2 different implementations of LVM on the web -- one included in 0.90 raid patch and one on http://linux.msede.com/lvm/) Can someone clarify this? A few months ago I asked what's the 'translucent' feature as well, but no reply.. :( Thanks, Egon Eckert
Re: Root RAID and unmounting /boot
syslogd 378 root5w REG9,18548 192396 /var/log/boot.log klogd 389 root2r REG8,1 191102 12 /boot/System.map-2.2.12-20 Is it safe to kill these? These are loggers, so I guess nothing terrible would happen. But I wouldn't kill them anyway.. I would fix the problem, not make it worse. Also i would be quite grateful if someone could explain to me why I must unmount /boot inorder for the lilo -r /mnt/newroot to work? I don't unmount anything before running lilo. What happens when i run lilo -r /mnt/newroot? You didn't try it? My 'extra' small non-raided boot partition usually contains 1. /boot with some files used by lilo (boot.*, chain.b, map...) -- copied from 'raided' partition 2. the kernel - 'zImage' 3. /dev with just few needed special files like 'hda', 'hda7' etc. 4. lilo.conf Egon
Re: any disk monitoring software for raids?
Anybody know of any disk rate monitoring software out there that will monitor disk ios and disk transfer rates on Linux systems? Do you know this? vmstat 1 Egon Eckert
Re: Kernel probs...
Where do I get the pre-2.2.13 patches? Looked on kernel.org but they weren't there. http://www.us.kernel.org/pub/linux/kernel/alan/2.2.13pre/ Egon Eckert
Re: Lilo configuration, device 0900 not known
If i try and put root=/dev/md1 in lilo.conf then i get the error device 0x0900 not known. There's my lilo.conf line: root=0x900 It works. 0x901 should work too, I think. :-) Egon
Re: raidhotremove
you will have to use raidsetfaulty first, then raidhotremove... Ugh? How? There's no such binary on my system. Echo it somewhere to /proc? Egon Eckert
Re: archive of this mailing list ?
is there an archive of this mailing list ? if yes: where can i find it ? There are probably more of them. The one I use is accessible from http://linuxwww.db.erau.edu/mail_archives/ Egon
Re: Need recommendations
tolerant free system. Currently I have 1 server with 4 disks on it (3 9.OGB and 1 4GB on external disk bay). On the 9 GB disks I would probably do 1x 5 GB partition 1x almost 4 GB partition 1x small partition for swap (30 MB or so, depending on total space) On the 4 GB disk I would do 1x almost 4 GB partition 1x small partition (the same size as on the 9 GB disks) for kernel ('boot') Then I would create: - RAID 5 volume on 3x 5 GB partitions on the 9 GB disks - RAID 5 volume on 4x 4 GB partitions on the remaining space on the 9's and on the 4 GB. - swap area on the 3 9 GB disks -- very fast, the kernel distributes the load on them You would have almost 22 GB of protected space. When the 4 GB disk would fail, you could boot from the diskette using a kernel parameter like root=0x900 to tell the kernel where's the root filesystem. Egon
Translucent mode, LVM?
Just curious: 1. What's the Translucent mode? The ability to do linear over RAID-5 etc.? 2. How about LVM (in alpha RAID patches)? Is it the same / incompatible / competition with LVM on http://linux.msede.com/lvm/? Thanks, Egon Eckert
Re: Recovering from 'Fake Crash'
that does for you. Perhaps those are the correct steps for a brand new unformatted drive. I did not have to do anything with the partitions. Nono.. With new (unformatted) drive you have to create the partition(s) first, that's for sure. I explained the reason for the strange, complicated method in my msg to Sergei. Egon
Re: Problems building raidtools 0.90
/usr/include/linux/errno.h:4: asm/errno.h: No such file or directory You must have at least kernel headers ready. Unpack recent kernel's sources to /usr/src/linux and make links /usr/include/{asm,linux,scsi} as suggested in /usr/src/linux/README. You may get further then. :) Egon
Re: Recovering from 'Fake Crash'
How do I re-sync these RAID devices? You have to add the partitions on sdb (if this is your 2nd SCSI disk) to the arrays again. 1. Do 'raidstop /dev/md0' .. 'raidstop /dev/md8'. 2. Change types of the sdb's partitions to 83 (if they have type fd) -- this will exclude them from autodetection on boot. 2. Reboot (the status will be still the same, degraded mode, but you will be able to add them 'raid hot' in next step). 3. Do 'raidhotadd /dev/md0 /dev/sdb1' .. 'raidhotadd /dev/md8 /dev/sdb8'. 4. Eventually wait until the resync finishes. 5. 'Raidstop' them all again. 6. Change the partitions' types back to fd. 7. Reboot. It should be completed. Is it necessary to have a 'spare disk' to be able to recovery my array? (ie, a partition on the IDE drive could be the spare disk..) Absolutely not. There is documentation on How-to-Setup a RAID, but I can't find much on one of the more important aspects, how to recover. :) Any URLs I'm missing? I prepared something (but it's a bit rough, needs cleanup) - look there: http://www.chance.cz/~egon/tinyr1rhowto.html Egon
Tiny RAID-1 recovery HOWTO proposal
Hi all, after some hours playing with RAID-1 recovery I wrote a tiny HOWTO-proposal, suitable mainly for newbies (I hope!). Please, comment it, send me all the corrections and flames, so it eventually can help constitute a needed section in the "real" HOWTO. The URL: http://www.chance.cz/~egon/tinyr1rhowto.html (Am I too bold/confident/daring/presumptuous?) (need corrections in the english as well) :-) Thanks, Egon
Nice: real disk failure
This is fascinating. It reminds me of Star Trek calm conversation in control center during laser shooting on enemy spaceship. :) hda: read_intr: status=0x59 { DriveReady SeekComplete DataRequest Error } hda: read_intr: error=0x40 { UncorrectableError }, LBAsect=1521738, sector=1501515 end_request: I/O error, dev 03:05 (ide), sector 1501515 raid1: Disk failure on hda5, disabling device. Operation continuing on 1 devices kmalloc called nonatomically from interrupt 002c raid1: md0: rescheduling block 750757 dirty sb detected, updating. md: updating md0 RAID superblock on device hdb5 [events: 0004](write) hdb5's sb offset: 798208 md: recovery thread got woken up ... md0: no spare disk to reconstruct array! -- continuing in degraded mode md: recovery thread finished ... (skipping faulty hda5 ) . raid1: md0: redirecting sector 750757 to another mirror Since this failure (a few days) it runs perfectly without reboot. Marvellous! Egon
RAID-1 resync hanging
Hello, may be I missed a thread about this... I'm using 2.0.36+raid990421, RAID-1 over two cheap (WD) similar IDE discs, booting root from it (partition types 0xfd). After an unclean reboot it starts resyncing, but stops immediately or after a while, raising indefinitely estimated completion time :) (and starts resyncing again next time, obviously). When I boot the same kernel "elsewhere" - not mounting this RAID-1 device as root - the recovery runs finishes OK, but this is only my amateurish observation, probably nothing very important.. Thanx for help, Egon Eckert, Prague, Czech Republic
Re: linux RAID on 2.0.38 Kernel
I've got one machine running this (2.0.37) and it works fine. I've got several Pentium I/II machines using 2.0.37 with raid 990824 patches and for me it's ROCK SOLID, even on SMP boards. I consider this a most stable Linux with SW-RAID setup, but your mileage may vary. :-) Egon Eckert