After reading some posts on Server Room I’ve decided to do some scrubbing on my zfs pools … just in case ( done it a few times some long time ago )
The I’ve noticed that one pool takes ages to scrub:
state: ONLINE
scrub: scrub in progress for 0h59m, 40.72% done, 1h26m to go
turns out one of the drives is under-performing :
extended device statistics
device   r/s   w/s  kr/s  kw/s wait actv svc_t %w %b
lofi1Â Â Â Â 0.0Â Â Â 0.0Â Â Â 0.0Â Â Â 0.0Â 0.0Â 0.0Â Â Â 0.0Â Â 0Â Â 0
lofi2Â Â Â Â 0.0Â Â Â 0.0Â Â Â 0.0Â Â Â 0.0Â 0.0Â 3.0Â Â Â 0.0Â Â 0 100
lofi3Â Â Â Â 0.0Â Â Â 0.0Â Â Â 0.0Â Â Â 0.0Â 0.0Â 0.0Â Â Â 0.0Â Â 0Â Â 0
sd0Â Â Â Â Â Â 0.0Â Â Â 0.0Â Â Â 0.0Â Â Â 0.0Â 0.0Â 0.0Â Â Â 0.0Â Â 0Â Â 0
sd1Â Â Â Â Â Â 0.0Â Â Â 0.0Â Â Â 0.0Â Â Â 0.0Â 0.0Â 0.0Â Â Â 0.0Â Â 0Â Â 0
sd2Â Â Â Â Â Â 0.0Â Â Â 0.0Â Â Â 0.0Â Â Â 0.0Â 0.0Â 0.0Â Â Â 0.0Â Â 0Â Â 0
sd3Â Â Â Â 226.0Â Â Â 0.0 28742.7Â Â Â 0.0Â 1.5Â 0.8Â Â 10.3Â 63Â 85
sd4Â Â Â Â 227.0Â Â Â 0.0 29056.7Â Â Â 0.0 19.0Â 1.0Â Â 88.1 100Â 99
That’s bad, it’s my only Seagate which survived in the box … looks like I will have to start hunting for cheap 1TB hitachi again …
by Tony Hoyle
21 Mar 2010 at 21:56
I’ve tended to stick to deathstars since Hitachi started doing them – they’ve been utterly reliable for me.
I’ve had no luck with samsung – my record being failure in less than a week.
by seeder
21 Mar 2010 at 22:01
Yeah, same here however after scrub I’ve noticed that Hitachi had one h/w error and this in dmesg
Mar 21 18:55:47 host ahci: [ID 777486 kern.warning] WARNING: ahci0: ahci port 2 has interface fatal error
Mar 21 18:55:47 host ahci: [ID 687168 kern.warning] WARNING: ahci0: ahci port 2 is trying to do error recovery
Mar 21 18:55:47 host ahci: [ID 551337 kern.warning] WARNING: ahci0: Transient Data Integrity Error (T)
Mar 21 18:55:47 host 10B to 8B Decode Error (B)
Mar 21 18:55:47 host CRC Error (C)
will check later if connectors are good …