mejutty Posted March 25, 2017 Share Posted March 25, 2017 (edited) I decided to replace my oldest disk a 3tb wd green. I had a new 8tb red ready to go all pre-cleared so I stopped the array assigned the new drive to the slot I wanted to use and started array and all looks good rebuild underway. Then this morning I wake up to see the array is now rebuilding at 345kb/sec I also noted that mover started and started to move some files. The first file it moved seemed to go as per normal in terms of speed but after that things went bad. What I did notice is that mover was moving data to the new currently being rebuilt disk I though the free space on a larger disk was not available until it was rebuilt and then even after a stop and start of the array?? Am I in trouble here?? It rebuilt 2.5tb in 11 hours now sayd it will finish in 256 days. tower-diagnostics-20170326-0754.zip Edited March 26, 2017 by mejutty spelling Quote Link to comment
Squid Posted March 26, 2017 Share Posted March 26, 2017 (edited) Maybe a problem, or a reporting bug in the preclear script Mar 25 20:04:42 Tower emhttp: WDC_WD80EFZX-68UW8N0_VKHZ2DPX (sdi) 7814026532 Mar 25 20:04:42 Tower emhttp: WDC_WD80EFZX-68UW8N0_VKHYKGVX (sdn) 7814026532 Mar 25 20:04:42 Tower kernel: md: import disk1: (sdn) WDC_WD80EFZX-68UW8N0_VKHYKGVX size: 7814026532 erased . . . Mar 25 20:05:12 Tower preclear.disk: Pausing preclear of disk 'sdn' Mar 25 20:05:12 Tower preclear.disk: Pausing preclear of disk 'sdi' . . . Mar 25 20:12:52 Tower kernel: md: recovery thread: recon D1 ... . . . Mar 25 20:12:54 Tower preclear.disk: Resuming preclear of disk 'sdn' Mar 25 20:12:54 Tower preclear.disk: Resuming preclear of disk 'sdi' Personally, if I saw on my system that it was resuming a preclear of a disk that has already been precleared, and is in the process of being rebuilt I'd be very worried.... If this is indeed what is happening, then it could quite possibly explain the very slow speed. Edited March 26, 2017 by Squid Quote Link to comment
mejutty Posted March 26, 2017 Author Share Posted March 26, 2017 I did see that i killed the preclear script before i replaced the disk. Also it initialy started rebuilding at over 100mb/sec so it only started happening. I also verfied with stats that there wasnt any other disk reads going on. Quote Link to comment
Squid Posted March 26, 2017 Share Posted March 26, 2017 (edited) is dd still running though? Depending upon how the preclear script was made, any processes spawned by it may not themselves be killed alongside the parent ps -aux | grep dd Beyond that, and if the stats are correct, then there is nothing particularly wrong in the syslog to indicate anything wrong. Edited March 26, 2017 by Squid Quote Link to comment
mejutty Posted March 26, 2017 Author Share Posted March 26, 2017 I hope dd is not running will check when i get home Quote Link to comment
mejutty Posted March 26, 2017 Author Share Posted March 26, 2017 (edited) nope no dd running. root@Tower:~# ps aux | grep dd root 2 0.0 0.0 0 0 ? S Mar23 0:00 [kthreadd] root 11 0.0 0.0 0 0 ? S< Mar23 0:00 [lru-add-drain] root 3683 0.0 0.0 273020 3948 ? S 00:35 0:00 /usr/sbin/winbindd -D root 11800 0.0 0.0 273020 7168 ? Ss Mar25 0:00 /usr/sbin/winbindd -D root 11802 0.0 0.0 272588 6804 ? S Mar25 0:00 /usr/sbin/winbindd -D root 13732 11.2 0.0 0 0 ? S Mar25 133:58 [unraidd] root 17046 0.0 0.0 9640 1832 pts/2 S+ 15:54 0:00 grep dd so what about it writing to the disk before it's finished a rebuild Edited March 26, 2017 by mejutty Quote Link to comment
Squid Posted March 26, 2017 Share Posted March 26, 2017 Disabling mover & ca appdata backup will stop most large writes to the array. You can try to avoid incidental writes to the array. I've never particularly worried about them. So long as my checks / rebuilds happen in what I consider to be a reasonable time, I don't care about the actual raw values. But should it drop down to kb/s then I'd be ripping everything apart and just start playing to see what fixes it. Unfortunately, I don't see anything particularly worrisome outside of the preclear within the syslog, so I'm not quite sure where to go from here beyond stopping everything and restarting it all again. Maybe wait for someone else to pipe in here... Quote Link to comment
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.