Thread (1 message) 1 message, 1 author, 2012-09-03
  • (off-list ancestor, not in this archive)
  • RE: hang · James Harper <hidden> · 2012-09-03

RE: hang

From: James Harper <hidden>
Date: 2012-09-03 11:28:48

quoted
Hmm.

If it's related to lvm snapshots, there's a workaround I just added to
the
bcache-3.2 branch - I'd give that a try, looks plausible from your log.
Is that the default writeback_percent = 10 in commit
113740372e2244951fb34235b30a2ac646a5d8c2?

Can I effect the same change by "echo 10 >writeback_percent"? If so then it
would save the hassle of a rebuild just to test. I assume it's sleep time in your
timezone at the moment so I'll test it anyway and let you know - it was failing
pretty consistently, within 30 seconds of starting my restore job.
Setting writeback_percent=10 manually resulted in no change - crash still occurred within a few minutes of the restore job starting.

Deleting the snapshot and running on the original LV resulted in no problems, so it's definitely snapshot related.

James
Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help