Hello all, I recently noticed that my storage pool has started to take a lot of time finishing a scrub, approximately the final 10% takes 30m to finish while the previous 90 are done is as many minutes. The 'zpool status' command does however not change its estimated remaining time. Currently 61% of the capacity is in use and I didn't observe this behaviour when it was ~58%. Monitoring the output of 'zpool iostat' I've learnt that during the later part of a scrub read speed drops to a tenth of what it is when the scrub starts. Sure, disks have different performance on different parts of the platter and then there is seek times but a tenth?? According to 'iostat' all looks good, no errors or congestion.
I'm considering moving data out of the pool to see if that affects the scrub behaviour but first I would like to ask the following. Has anyone here experienced anything similar and if so what caused it? And if it is possible to monitor which file is currently being scrubbed during a scrub? Thank you! -- This message posted from opensolaris.org _______________________________________________ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
