zfs corruption due to dual import
edited Jun 7, 2012 1:15PM in Oracle Solaris File Systems and Disk Management (MOSC) 3 commentsAnswered
Hello,
I had an issue where a zpool somehow got imported on two nodes at the same time. This lead to a panic
( assertion failed: 0 == dmu_buf_hold_array ) Luckily the pool could be cleared no harm done. However
I would like to prevent this in the future. Because it happened in a cluster, cluster was obviously a suspect,
however there were absolutely no cluster related actions, events of any sorts at the time. We go to great
extents in cluster to guarantee data integrity so the resource group manager does not allow this to happen.
The question is if you know of any either open zfs issues or some that already had been fixed and patch is
I had an issue where a zpool somehow got imported on two nodes at the same time. This lead to a panic
( assertion failed: 0 == dmu_buf_hold_array ) Luckily the pool could be cleared no harm done. However
I would like to prevent this in the future. Because it happened in a cluster, cluster was obviously a suspect,
however there were absolutely no cluster related actions, events of any sorts at the time. We go to great
extents in cluster to guarantee data integrity so the resource group manager does not allow this to happen.
The question is if you know of any either open zfs issues or some that already had been fixed and patch is
0