Hi,
The reason is found out, as you've predicted.
> is it because of slow read operation (i.e. some raid arrays are known to
> wake-up slowly)
The udev worker is blocked until timeout before killed, cause the
"blkid" in /usr/lib/udev/rules.d/13-dm-disk.rules is too slow because
of high IO load on the thin pool.
After remove this rule:
IMPORT{builtin}="blkid"
for testing, the problem cannot be reproduced any more.
But, we tried to increate udevd timeout to 5 min, the problem still
happens; with 10 min timeout, can workaround this problem.
>From my understanding, the blkid will only read the fs superblock, why
it takes so long for a small IO to finish, even under high IO load?
Regards,
Eric
_______________________________________________
linux-lvm mailing list
linux-lvm@xxxxxxxxxx
https://www.redhat.com/mailman/listinfo/linux-lvm
read the LVM HOW-TO at http://tldp.org/HOWTO/LVM-HOWTO/
