[DRBD-user] filesystem corruptions

bro zxr at lanparty.lv
Mon Oct 3 10:17:53 CEST 2005

Note: "permalinks" may not be as permanent as we would like,
direct links of old sources may well be a few messages off.


Todd Denniston wrote:
[snip]
> 
> 
> as Brandon indicated DRBD is doing the correct thing by panicking, when the
> scsi device goes nuts.
> 
> This looks very much like a hardware problem.  Assuming no hardware has
> intentionally changed on the scsi bus I would [no warranties on my
> procedures]
> 1) disconnect and reconnect each connection in the scsi bus (too often I
> have seen corrosion muck up a scsi bus) , including the terminator(s) if
> they are not built into the device logically.
> 
> 2) verify the last device on the buss is either a proper terminator (that is
> working) or the last device has internal termination turned on.
> 
> seeing the 'ata3:' above the scsi error makes me want to think you have SATA
> hardware, which IIRC current linux is presenting as scsi devices.
> 
> 3) in either case, backup any data and run badblocks in full write mode and
> Brandon's suggestions on the physical device.
> 

Thanks for the ideas, ive been testing the underlying with `badblocks 
-sw /dev/sdb1` and `sdc1` all weekend long, and havent found nothing at all.
About that SATA thing - that's true, ive got two sata devices on which 
drbd resides, so i pressume that there aint problems with connectors, 
and corrosion is the second thing we can exclude from problem causers ;)
Actually it seems like some weird HW problem, cuz the secondary node is 
running fine for about 4 days w/o problems instead of the primary which 
crashed 8h after bootup.
Any ideas and thoughts about this will be greatly appreciated ;)



More information about the drbd-user mailing list