devsk
2010-Jul-11 00:43 UTC
[zfs-discuss] zpool scrub is clean but still run in checksum errors when sending
''zpool scrub mypool'' comes out clean. zfs send -Rv mypool at blah | ssh ... reports IO error. And indeed, ''zpool status -v'' shows errors in some files in an older snapshot. Repeat scrub without errors and clear the pool. And send now fails on a different set of files. How can this happen? What would be the best way to troubleshoot this kind of error? -- This message posted from opensolaris.org
devsk
2010-Jul-11 01:01 UTC
[zfs-discuss] zpool scrub is clean but still run in checksum errors when sending
This is the fourth time its happened. I had a clean scrub before starting send 15 minutes ago. And now, I see permanent errors in one of the files of one of the snapshots. Different file from last time. How do I find what''s going on? ''dmesg'' in guest or kernel does not have any relevant entries. BTW: This is inside VirualBox but I have the host cache disabled. -- This message posted from opensolaris.org