shire Posted April 15, 2018 Share Posted April 15, 2018 Hello, a few weeks ago I changed the file system from rfs to xfs. That worked great...almost. In the process I moved data from a rfs drive to an already reformated xfs drive. The xfs drive turned red (drive7) while writing to it. However reading form that drive and writing to another one worked without any problem before in that process. To solve the problem I deleted all data on drive7 and replaced that with a new one (the data was still on that rfs drive I was copying from). I rebuilt the array with a parity check and all looked good. I restarted copying the data to that new drive and so on. I also checked if I can access the data on that drive without problems (manly video files) which worked. This night the mover tried to move files to the array and the same drive bay with the new drive red balled again. I don't know what is going on. I added the diagnostic. Maybe somebody could take a look. Bye. tower-diagnostics-20180415-0810.zip Quote Link to comment
JorgeB Posted April 15, 2018 Share Posted April 15, 2018 Disk dropped offline so there's no SMART, but possibly a bad cable/connection issue.. Quote Link to comment
shire Posted April 15, 2018 Author Share Posted April 15, 2018 I changed the SATA cable. The drive is still read (of course). I activated SMART on that drive and started the self test which discovered no errors. Just now I started a telnet session. I'm copying a big file form another disk to disk7 via command line. Let's see if there are any errors... Furthermore I don't know what to do. I would probably try to rebuild that disc and see what's happening. Quote Link to comment
shire Posted April 15, 2018 Author Share Posted April 15, 2018 Hm... is copying possible? Looks like the drive is emulated and you can't copy to it via the command line. Quote Link to comment
JorgeB Posted April 15, 2018 Share Posted April 15, 2018 After a disk is disable it needs to be rebuilt, to the old or a new disk, but replace the cables first to rule them out. Quote Link to comment
trurl Posted April 15, 2018 Share Posted April 15, 2018 2 hours ago, shire said: Hm... is copying possible? Looks like the drive is emulated and you can't copy to it via the command line. Emulated disks can be read or written so without further information it's not clear what your problem is with copying from command line. But, it's really much better to rebuild the disk before trying to do anything else with the array, since everything has to work a lot harder to emulate the disk (all disks are involved in the emulation), and the array is unprotected unless you have more parity disks than disabled disks. Quote Link to comment
shire Posted April 28, 2018 Author Share Posted April 28, 2018 Yeah...it's not the drive and not the cable. I changed the cable and tried rebuilding the drive. As soon as it started the system reported: unRAID Disk 7 SMART health [199] Warning [TOWER] - udma crc error count is 2 1524860822 WDC_WD30EZRZ-00Z5HB0_WD-xxx (sdg) warning To test further I changed the SATA-port (onboard to my Supermicro aoc-saslp-mv8) using the same place in my XCase drive bay (which also could be responsible). I was able to rebuild the drive and I now try to copy files from the cache drive into the array. Looks like I should get a second Supermicro card. Funny....the first thing that died on my old Supermicro mainboard was a SATA-port too before I bought the ASRock. Quote Link to comment
trurl Posted April 28, 2018 Share Posted April 28, 2018 https://lime-technology.com/forums/topic/66327-unraid-os-version-650-stable-release-update-notes/ A little ways down under under Solutions to Common Problems it explains these CRC warnings. There may be some other things in that post worth looking at for you. Quote Link to comment
trurl Posted April 28, 2018 Share Posted April 28, 2018 24 minutes ago, shire said: I now try to copy files from the cache drive into the array. Not sure what this is about. The normal way is to let mover move files from cache to array instead of doing it yourself. It is very important that you don't mix disks and user shares when copying or moving files. Only disk to disk or user share to user share. And cache is a disk. Quote Link to comment
shire Posted April 28, 2018 Author Share Posted April 28, 2018 10 minutes ago, trurl said: Not sure what this is about. The normal way is to let mover move files from cache to array instead of doing it yourself. It is very important that you don't mix disks and user shares when copying or moving files. Only disk to disk or user share to user share. And cache is a disk. That's exactly what I'm doing. I started the mover manually. Quote Link to comment
shire Posted April 28, 2018 Author Share Posted April 28, 2018 (edited) 23 minutes ago, trurl said: https://lime-technology.com/forums/topic/66327-unraid-os-version-650-stable-release-update-notes/ A little ways down under under Solutions to Common Problems it explains these CRC warnings. There may be some other things in that post worth looking at for you. Thanks. I used the "Common Problems Plugin" to check. There was nothing which needed a reconfiguration. Quite interesting was that: Marvel Hard Drive Controller Installed That Supermicro controller is still the first card recommended in the thread "SATA controller card". That's the reason I bought it in the first place. Also I hadn't had a problem with it since I installed it a few years ago. Edited April 28, 2018 by shire Quote Link to comment
John_M Posted April 28, 2018 Share Posted April 28, 2018 It seems that the mvsas driver has got more and more flaky with newer releases. I used to have success with my AOC-SAS-MV8 and AOC-SAS2-MV8 as long as I didn't try to enable IOMMU but recently they had become a liability (dropping disks for no obvious reason) so I replaced them with Dell H310s, which I'm finding more reliable and also faster. 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.