Ockingshay Posted March 2, 2013 Share Posted March 2, 2013 checked my server this morning and my 4 plugins aren't working. If i look under settings i can see that they are all in a "stopped" status. I then check the data directories and they are all missing crucial files. each plugin says "install" as the option and this obviously will put the config files back in but loses all my customisation. The only thing i can think of is the mover running and doing something i don't want it to do/or i don't understand the principle of how to set things up. So my question is...what is the way to install and have them persistent? I have all plugins installed to /mnt/cache/apps/... It manually created a share called apps and it's set to use cache "only" mover runs 03:40 each day as per default. I also notice that with simpleFeatures the shares with cache are all still in a pending state for a couple of days now. I've not used cache drive before. Can someone please give me their example of how they would set this all up? Thanks Link to comment
Automatic Posted March 2, 2013 Share Posted March 2, 2013 Mar 2 09:00:01 Tower logger: mover started Mar 2 09:00:01 Tower logger: skipping Appdata/ Mar 2 09:00:01 Tower logger: mover finished That's the log from my mover script, seems to skip right over the cache-only share. Any chance of uploading your log? Link to comment
Helmonder Posted March 2, 2013 Share Posted March 2, 2013 The mover will not take action if that share is set to cache only, so that cannot be it. How long ago did you install the plugins ? Recently ? Is the physical data still in the /mnt/cache/apps directory or is it gone ? Have you rebooted and what happens then ? Aso please attach a syslog Link to comment
Ockingshay Posted March 2, 2013 Author Share Posted March 2, 2013 Had a chance to look at logs and i see a lot of this: Mar 2 03:55:00 Ruby logger: rsync error: some files/attrs were not transferred (see previous errors) (code 23) at main.c(1042) [sender=3.0.7] Mar 2 03:55:00 Ruby logger: ./apps/Plex/Application Support/Plex Media Server/Metadata/Movies/b/224a15fe5089733dd3a8fc376580c3ed6441015.bundle/Contents/com.plexapp.agents.themoviedb/posters/4c0e1fbbccb7c7d1a721de4eb9d2d3d14f764cf7 Mar 2 03:55:00 Ruby shfs/user0: shfs_mkdir: assign_disk: apps/Plex/Application Support/Plex Media Server/Metadata/Movies (28) No space left on device Mar 2 03:55:00 Ruby logger: cd+++++++++ apps/Plex/Application Support/Plex Media Server/Metadata/Movies/ Mar 2 03:55:00 Ruby logger: rsync: recv_generator: mkdir "/mnt/user0/apps/Plex/Application Support/Plex Media Server/Metadata/Movies" failed: No space left on device (28) Mar 2 03:55:00 Ruby logger: *** Skipping any contents from this failed directory *** Mar 2 04:01:23 Ruby logger: rsync: recv_generator: mkdir "/mnt/user0/apps/DropboxDB" failed: No space left on device (28) Mar 2 04:01:23 Ruby logger: *** Skipping any contents from this failed directory *** Mar 2 04:01:23 Ruby logger: rsync error: some files/attrs were not transferred (see previous errors) (code 23) at main.c(1042) [sender=3.0.7] Mar 2 04:01:23 Ruby logger: mover finished That leads to 2 questions, why is their no space left on device (28)? - what is device (28)? Why are they being moved when the share is set to cache only? Interestingly enough it didn't touch a 5 and 6th plugin (transmission & Dropbox), which are also installed in /mnt/cache/apps/ Mar 2 03:40:01 Ruby logger: mover started Mar 2 03:40:01 Ruby logger: moving Dropbox/ Mar 2 03:40:01 Ruby logger: ./Dropbox/sickbeard Mar 2 03:40:01 Ruby logger: .d..t...... ./ Mar 2 03:40:01 Ruby logger: .d..t...... Dropbox/ Mar 2 03:40:01 Ruby logger: cd+++++++++ Dropbox/sickbeard/ Mar 2 03:40:01 Ruby logger: ./Dropbox/Transmission/ Log is 9mb, cannot attach to this post Very frustrating... Link to comment
Ockingshay Posted March 2, 2013 Author Share Posted March 2, 2013 I also tried running the mover again manually and it gets the same errors: failed: No space left on device (28) and the share still says it has pending cache. Is there something wrong with the share? is it worth recreating it? I stopped the array and removed the cache drive from the assigned disks, started it up and then did the same to re-add it and i still get those errors. dropbox this time: 2 12:37:26 Ruby logger: cd+++++++++ apps/DropboxDB/ Mar 2 12:37:26 Ruby logger: rsync: recv_generator: mkdir "/mnt/user0/apps/DropboxDB" failed: No space left on device (28) Mar 2 12:37:26 Ruby logger: *** Skipping any contents from this failed directory *** Mar 2 12:37:26 Ruby logger: rsync error: some files/attrs were not transferred (see previous errors) (code 23) at main.c(1042) [sender=3.0.7] Mar 2 12:37:26 Ruby logger: mover finished Mar 2 12:38:07 Ruby emhttp: shcmd (438): /usr/local/sbin/set_ncq sdb 1 &> /dev/null Mar 2 12:38:07 Ruby emhttp: shcmd (439): /usr/local/sbin/set_ncq sda 1 &> /dev/null Mar 2 12:38:07 Ruby emhttp: shcmd (440): /usr/local/sbin/set_ncq sde 1 &> /dev/null Mar 2 12:38:07 Ruby emhttp: shcmd (441): /usr/local/sbin/set_ncq sdd 1 &> /dev/null Mar 2 12:38:07 Ruby kernel: mdcmd (35): set md_num_stripes 1280 Mar 2 12:38:07 Ruby kernel: mdcmd (36): set md_write_limit 768 Mar 2 12:38:07 Ruby kernel: mdcmd (37): set md_sync_window 384 Mar 2 12:38:07 Ruby kernel: mdcmd (38): set spinup_group 0 0 Mar 2 12:38:07 Ruby kernel: mdcmd (39): set spinup_group 1 0 Mar 2 12:38:07 Ruby kernel: mdcmd (40): set spinup_group 2 0 I know there is enough space on the drive as it's 500GB, which webgui shows as 500GB free Link to comment
Automatic Posted March 2, 2013 Share Posted March 2, 2013 I also tried running the mover again manually and it gets the same errors: failed: No space left on device (28) and the share still says it has pending cache. Is there something wrong with the share? is it worth recreating it? I stopped the array and removed the cache drive from the assigned disks, started it up and then did the same to re-add it and i still get those errors. dropbox this time: 2 12:37:26 Ruby logger: cd+++++++++ apps/DropboxDB/ Mar 2 12:37:26 Ruby logger: rsync: recv_generator: mkdir "/mnt/user0/apps/DropboxDB" failed: No space left on device (28) Mar 2 12:37:26 Ruby logger: *** Skipping any contents from this failed directory *** Mar 2 12:37:26 Ruby logger: rsync error: some files/attrs were not transferred (see previous errors) (code 23) at main.c(1042) [sender=3.0.7] Mar 2 12:37:26 Ruby logger: mover finished Mar 2 12:38:07 Ruby emhttp: shcmd (438): /usr/local/sbin/set_ncq sdb 1 &> /dev/null Mar 2 12:38:07 Ruby emhttp: shcmd (439): /usr/local/sbin/set_ncq sda 1 &> /dev/null Mar 2 12:38:07 Ruby emhttp: shcmd (440): /usr/local/sbin/set_ncq sde 1 &> /dev/null Mar 2 12:38:07 Ruby emhttp: shcmd (441): /usr/local/sbin/set_ncq sdd 1 &> /dev/null Mar 2 12:38:07 Ruby kernel: mdcmd (35): set md_num_stripes 1280 Mar 2 12:38:07 Ruby kernel: mdcmd (36): set md_write_limit 768 Mar 2 12:38:07 Ruby kernel: mdcmd (37): set md_sync_window 384 Mar 2 12:38:07 Ruby kernel: mdcmd (38): set spinup_group 0 0 Mar 2 12:38:07 Ruby kernel: mdcmd (39): set spinup_group 1 0 Mar 2 12:38:07 Ruby kernel: mdcmd (40): set spinup_group 2 0 I know there is enough space on the drive as it's 500GB, which webgui shows as 500GB free Twenty eight (I would presume) is the error code, is there enough space on disks inside of the array? Also, from googling it looks like "/tmp" is full, try restarting the server (Clears /tmp/ out)? Source of the /tmp/ idea:- http://www.webhostingtalk.com/showthread.php?t=655982 Link to comment
Ockingshay Posted March 2, 2013 Author Share Posted March 2, 2013 i've gone for a full reboot and parity check, have to wait 7 hours now. didn't check tmp before reboot but now it empty, so i'll run file mover after that. I have rebooted previously, so i don't think this is the issue. Link to comment
Automatic Posted March 2, 2013 Share Posted March 2, 2013 i've gone for a full reboot and parity check, have to wait 7 hours now. didn't check tmp before reboot but now it empty, so i'll run file mover after that. I have rebooted previously, so i don't think this is the issue. Everything other than /boot/* and /mnt/* is cleared out 100% upon a restart, so, yes, tmp will be empty (Unless a program places tmp files there). Also, just a note, you can write to your drive during a parity check (I wouldn't recommend if it if you think a drive is failing, but, if it's just a scheduled parity check it'll be fine), it will be slower, but, it'll still work, so, just run the mover now. Link to comment
Ockingshay Posted March 2, 2013 Author Share Posted March 2, 2013 ok, just running mover and getting the same... Mar 2 13:20:09 Ruby shfs/user0: shfs_mkdir: assign_disk: apps/sickbeard/sickbeard (28) No space left on device Mar 2 13:20:09 Ruby logger: cd+++++++++ apps/sickbeard/sickbeard/ Mar 2 13:20:09 Ruby logger: rsync: recv_generator: mkdir "/mnt/user0/apps/sickbeard/sickbeard" failed: No space left on device (28) Mar 2 13:20:09 Ruby logger: *** Skipping any contents from this failed directory *** Mar 2 13:20:10 Ruby logger: rsync error: some files/attrs were not transferred (see previous errors) (code 23) at main.c(1042) [sender=3.0.7] Mar 2 13:20:10 Ruby logger: ./apps/sickbeard/sickbeard/postProcessor.pyc Mar 2 13:20:10 Ruby shfs/user0: shfs_mkdir: assign_disk: apps/sickbeard/sickbeard (28) No space left on device Mar 2 13:20:10 Ruby logger: cd+++++++++ apps/sickbeard/sickbeard/ Mar 2 13:20:10 Ruby logger: rsync: recv_generator: mkdir "/mnt/user0/apps/sickbeard/sickbeard" failed: No space left on device (28) Mar 2 13:20:10 Ruby logger: *** Skipping any contents from this failed directory *** Link to comment
Automatic Posted March 2, 2013 Share Posted March 2, 2013 ok, just running mover and getting the same... Mar 2 13:20:09 Ruby shfs/user0: shfs_mkdir: assign_disk: apps/sickbeard/sickbeard (28) No space left on device Mar 2 13:20:09 Ruby logger: cd+++++++++ apps/sickbeard/sickbeard/ Mar 2 13:20:09 Ruby logger: rsync: recv_generator: mkdir "/mnt/user0/apps/sickbeard/sickbeard" failed: No space left on device (28) Mar 2 13:20:09 Ruby logger: *** Skipping any contents from this failed directory *** Mar 2 13:20:10 Ruby logger: rsync error: some files/attrs were not transferred (see previous errors) (code 23) at main.c(1042) [sender=3.0.7] Mar 2 13:20:10 Ruby logger: ./apps/sickbeard/sickbeard/postProcessor.pyc Mar 2 13:20:10 Ruby shfs/user0: shfs_mkdir: assign_disk: apps/sickbeard/sickbeard (28) No space left on device Mar 2 13:20:10 Ruby logger: cd+++++++++ apps/sickbeard/sickbeard/ Mar 2 13:20:10 Ruby logger: rsync: recv_generator: mkdir "/mnt/user0/apps/sickbeard/sickbeard" failed: No space left on device (28) Mar 2 13:20:10 Ruby logger: *** Skipping any contents from this failed directory *** I'm out of ideas, I'd recommend googling some more while waiting for someone with more unraid knowledge to help you. You might also want to post the exact amount of RAM free, along with the space left on each data disk & cache drive. Link to comment
Joe L. Posted March 2, 2013 Share Posted March 2, 2013 Type df and see what space is available. It sure looks like one of your disks is full/nearly full. Link to comment
Ockingshay Posted March 2, 2013 Author Share Posted March 2, 2013 Filesystem 1K-blocks Used Available Use% Mounted on /dev/sdc1 3813856 198816 3615040 6% /boot /dev/md1 2930177100 2205306612 724870488 76% /mnt/disk1 /dev/md2 1953454928 647672604 1305782324 34% /mnt/disk2 /dev/sdd1 488371640 38656 488332984 1% /mnt/cache shfs 4883632028 2852979216 2030652812 59% /mnt/user0 shfs 4883632028 2852979216 2030652812 59% /mnt/user One thing i do remember is that the cache drive wasn't a new drive and i don't recall it doing any sort of clearing/formatting. Could it be that it's reporting empty, when in actual fact it's not? Is there any way to low level format it in unraid? EDIT* Just noticed preclear as a sticky, will give that a go Link to comment
Helmonder Posted March 3, 2013 Share Posted March 3, 2013 Not sure what you want to preclear but remember that action is desctructive, all data on the drive will be deleted. Preclear is used to "burn in" a new drive. Link to comment
Ockingshay Posted March 3, 2013 Author Share Posted March 3, 2013 Not sure what you want to preclear but remember that action is desctructive, all data on the drive will be deleted. Preclear is used to "burn in" a new drive. I do not understand what Limetech's definition of "burn in" is, but i do know it writes the drive to all 0s, which is what i want it to do to eliminate it being a possibility in my issue of the mover running out of space. I still have the issue, but it seems the mover starts alright, but then runs out of space. What space is it running out of, ie what temp space does it use to move the files? not flash surely? left on my flash is 3.7GB. Ultimately this is a side issue because i have set my "apps" (and now created a new share called plugins, to test) to "cache only" and shouldn't be part of the mover script. Like automatic pointed out his mover does this: Mar 2 09:00:01 Tower logger: mover started Mar 2 09:00:01 Tower logger: skipping Appdata/ Mar 2 09:00:01 Tower logger: mover finished Where as my mover doesn't. I've taken out all the non pertinent info in my log file and managed to get it under the limit, so please see attached. Searching the web it appears to use /tmp, if i du -h that i get: root@Ruby:/tmp# du -h 0 ./notifications/unread 0 ./notifications 4.0K ./web 0 ./.X11-unix 0 ./.ICE-unix Where is /tmp located? on which drive? Filesystem 1K-blocks Used Available Use% Mounted on /dev/sdc1 3813856 199020 3614836 6% /boot /dev/md1 2930177100 2205306612 724870488 76% /mnt/disk1 /dev/md2 1953454928 647719252 1305735676 34% /mnt/disk2 /dev/sdd1 488371640 103328 488268312 1% /mnt/cache shfs 4883632028 2853025864 2030606164 59% /mnt/user0 shfs 4883632028 2853025864 2030606164 59% /mnt/user syslog4.txt Link to comment
Ockingshay Posted March 3, 2013 Author Share Posted March 3, 2013 I think i may be getting somewhere, i just noticed those errors are attributed to /mnt/user0/apps/xxx, where as i'm now using /plugins....i can see there is a new entry in log, which appears as though it's working: Mar 3 10:05:35 Ruby logger: mover started Mar 3 10:05:35 Ruby logger: skipping Plugins/ Mar 3 10:05:35 Ruby logger: moving apps/ a corrupt share perhaps? i'll remove it completely, as not needed anyway. *EDIT* well i've cracked it and i'd like to thank everyone involved for your inputs. A great little community here. It wasn't just one thing, but many. On a side note, using the x9scm-f mobo with IPMI has helped immensely, so i recommend that board to anyone. Link to comment
trurl Posted March 3, 2013 Share Posted March 3, 2013 Just to clear up one of your previous questions, /tmp is in RAM. unRAID has a RAM filesystem where all of its operating system files are kept. They are restored on boot by unpacking bzroot. This is why anything stored in any of the usual linux operating system folders does not survive a reboot. In general, anything not in /boot (your flash drive) or /mnt (actual disks or shares on those disks) is in RAM. Link to comment
Recommended Posts
Archived
This topic is now archived and is closed to further replies.