On Wed, 28 Jul 2010, Krzysztof Mazurek wrote:
Hi everone,
I managed to configure the multipathing devices with
device-mapper-multipathing rdac (active-passive), however while system is
booting I can see a lot of messages like:
Buffer I/O error on device sdaa, logical block 0
end_request: I/O error, dev sdab, sector 0
end_request: I/O error, dev sdac, sector 0
[...]
Yes we get them too.
Probably nash or something else during startup is checking all the devices
(including the passive ones) that cause the very long system startup (we
have like 26 FC devices).
After startup device multipathig works fine, /dev/mapper is full of
aliased devices, but still, from time to time, there are some error
messages appering on the console. Like system was checking the passive
nodes.
Is there any way to prevent such behavior ?
I asked something similar to this question back on 8 July 2010 (see:
http://www.mail-archive.com/[email protected]/msg07203.html) where I'd
potentially discovered that the scsi_dh_rdac module might not have been
being loaded until later in the boot process and therefore before that point
the kernel isn't sure how to handle multiple connections/paths to the same
devices.
I asked about what the 'correct' way was to add the module into the initrd
as a preload but have since discovered that it was apparently put into
initrd in kernels from kernel-2.6.18-166.el5 onwards (see
https://bugzilla.redhat.com/show_bug.cgi?id=518496).
As a result of that and the fact that no-one responded to my email I'm a bit
stumped myself. I can't try the mkinitrd options I list in my previous mail
because we only have one system I could try it on right now and that's in
live/production use. If you make any progress please post it to the list!
Ben
--
Unix Support, MISD, University of Cambridge, England
Plugger of wire, typer of keyboard, imparter of Clue
Life Is Short. It's All Good.
_______________________________________________
rhelv5-list mailing list
[email protected]
https://www.redhat.com/mailman/listinfo/rhelv5-list
