Time Nick Message 00:34 pdurbin prologic: people here might know about ZFS 00:36 prologic ahh 00:36 prologic anyone here a zfs expoert and/or SMARRT wizard? :) 00:39 pdurbin prologic: SMARRT with 2 Rs? 00:43 bene it's verruh SMARRT 00:44 pdurbin wicked smaht 00:44 prologic eh? 00:45 pdurbin prologic: what's your question? 00:46 prologic Okay 00:47 prologic my question is: 00:48 prologic Should I just replace the disk /dev/sdf that last night reported an increased error count from 7 to 8? I'm doing a scrub atm (normally I do once/month via crond). It's a RAID-Z2 so I'm not that worried; but just what's best practice here? smartctl -a /dev/sdf (https://gist.github.com/39c9365994e4e98a7371) seems to indicate all 8 errors are CRC errors; so the drive is producing CRC errors? 00:49 prologic Also: 00:49 prologic $ zpool status 00:49 prologic pool: data 00:49 prologic state: ONLINE 00:49 prologic scan: scrub repaired 0 in 0h27m with 0 errors on Sat Aug 22 10:35:01 2015 00:49 prologic So I'm not sure really what to do here 00:49 prologic the pool is not reporting any issues despite the CRC error count increases 00:52 pdurbin I guess replace it. If it's cheap. 00:54 pdurbin or it's enough to convince the vendor to send you a new one. NetApp was always good about this 00:57 pdurbin they also cost an arm and a leg, of course :) 01:28 prologic well these drives aren't cheap 01:28 prologic and they are still under warrenty 02:05 prologic someone just mentioned in #zfsonlinux that CRC errors usually mean a bad SATA cable 02:06 prologic and that the drive is probably fine (not reporting any critical sector/read errors) 02:08 pdurbin hmm. ok 02:10 prologic hey this is all new to me :) 02:10 prologic I'm just trying to find practical sound advice 02:10 prologic from more experience persons :) 02:11 prologic rung the shop I bought them from and he said to bring it in for replacement (seems to know about SMART stuff); said it'd take 3-4 weeks to RMA 02:11 prologic so I'd have to buy a new driver anyway (which they have none of that even closely match the other 9 I have!) 11:18 pdurbin prologic: cool. so you were able to convince them to RMA it 14:27 bene in the US most of the drive manufacturers offer advance RMAs for a $5-10 fee 14:27 bene takes 2-3 days in my experience 14:27 bene they send you a new drive and prepaid return mailer for the dead one 14:27 bene well worth it from my POV 14:28 bene but that drive doesn't look like it's in bad shape to me 14:28 bene the cable swap might be worth trying though 14:28 bene similar advice here: https://ata.wiki.kernel.org/index.php/Libata_error_messages 14:28 bene (search for ICRC) 15:56 pdurbin I wonder if this will be any good: http://www.devopsdays.org/events/2015-boston/ 16:50 bene sold out 17:54 pdurbin huh. I couldn't even tell what the talks are 20:34 bene if you're on the boston devops meetup list you get the announcements about that conference when it's still open 20:38 pdurbin oh. I think I heard about it at http://theshipshow.com/2015/08/managing-the-communityops/ 20:38 pdurbin bene: is there a list of talks?