October 24, 20169 yr I finally upgraded from 6.1.9 to 6.2.2 yesterday. Everything went well as I had some concerns about my XFS file system from 6.1.9 possibly having issues based on the upgrade thread. Anyway, server was running fine for most of the day and then I lost access to the webgui. I have a keyboard and monitor hooked up so when I logged in as root, I was able to get the WebUI back. I could see the cache drive was unmountable and consulted the wiki for my options. First, I ran a check using the -nv option and within 2 seconds it reported "Metadata corruption detected at xfs_agf block 0x1/0x200". I'm running a Sandisk Ultra II 480GB drive as my cache where I store all of my appdata and docker image. I was able to repair the drive using XFS_Repair and everything is back to normal with no obvious data loss (lost+found folder has one file with 0kB size) but I'm wondering if there's anything else I should be concerned about or need to do? On an unrelated topic of my cache drive, should I be using a plugin to perform a trim operation with this drive? Diagnostics log from before my reboot is attached. Any insight appreciated. tower-diagnostics-20161023-2206.zip
October 24, 20169 yr Community Expert You might consider editing the subject line since it say "XFS Parity" and you are really talking about your "XFS Cache". Parity doesn't have a filesystem.
October 24, 20169 yr Author You might consider editing the subject line since it say "XFS Parity" and you are really talking about your "XFS Cache". Parity doesn't have a filesystem. Good point. Done
October 27, 20169 yr Author Just bumping this to see if there's any recommendation on further action required? Server has been stable since using XFS_Repair. I'm just not sure if this is something I need to monitor or take further action with or chalk it up to a one-time upgrade issue? Also, any insight on whether I should be "trimming" this SSD cache drive is appreciated. The trim plugin available wasn't working at first because I had my SSD cache on a SAS/SATA card. I moved the drive to one of the SATA ports on my MB but that didn't seem to fix it. I wanted to avoid further investigation if it's not warranted in this case.
October 27, 20169 yr Community Expert Just bumping this to see if there's any recommendation on further action required? Server has been stable since using XFS_Repair. I'm just not sure if this is something I need to monitor or take further action with or chalk it up to a one-time upgrade issue? Should be fine, if you don't have one consider buying a UPS to avoid power cuts, in my experience they are the cause of most file system corruption cases. Also, any insight on whether I should be "trimming" this SSD cache drive is appreciated. The trim plugin available wasn't working at first because I had my SSD cache on a SAS/SATA card. I moved the drive to one of the SATA ports on my MB but that didn't seem to fix it. I wanted to avoid further investigation if it's not warranted in this case. You should be trimming it, or write performance will degrade (it will also wear out faster), fstrim should work on most onboard SATA ports, it will work on any Intel onboard SATA port in AHCI mode.
October 27, 20169 yr Author Just bumping this to see if there's any recommendation on further action required? Server has been stable since using XFS_Repair. I'm just not sure if this is something I need to monitor or take further action with or chalk it up to a one-time upgrade issue? Should be fine, if you don't have one consider buying a UPS to avoid power cuts, in my experience they are the cause of most file system corruption cases. Thanks. I actually have a UPS hooked up to my UnRAID server and it's functioning correctly. Also, any insight on whether I should be "trimming" this SSD cache drive is appreciated. The trim plugin available wasn't working at first because I had my SSD cache on a SAS/SATA card. I moved the drive to one of the SATA ports on my MB but that didn't seem to fix it. I wanted to avoid further investigation if it's not warranted in this case. You should be trimming it, or write performance will degrade (it will also wear out faster), fstrim should work on most onboard SATA ports, it will work on any Intel onboard SATA port in AHCI mode. So I have the SSD Dynamix Plugin installed but I can't seem to configure anything?! Unlike other plugins, hovering over the icon isn't a hyperlink to anything. Looks like the only parameter to set would be the duration between trims but how would I know if a trim operation is ever being performed? Should I just add a line to my go script? Same question though, how would I know it completed?
October 27, 20169 yr Looks like the only parameter to set would be the duration between trims Settings - Scheduler Should I just add a line to my go script?no Same question though, how would I know it completed? Something like this appears in the syslog Oct 27 00:02:00 Server_A root: /mnt/cache: 101.5 GiB (108977012736 bytes) trimmed
October 27, 20169 yr Author Looks like the only parameter to set would be the duration between trims Settings - Scheduler Should I just add a line to my go script?no Same question though, how would I know it completed? Something like this appears in the syslog Oct 27 00:02:00 Server_A root: /mnt/cache: 101.5 GiB (108977012736 bytes) trimmed Awesome, thanks for the tips!
Archived
This topic is now archived and is closed to further replies.