January 24, 20188 yr Hi, i have a problem with the docker xeoma it shows whole day:Jan 24 01:37:11 Unraid-Server shfs: error: shfs_rmdir, 1517: Directory not empty (39): rmdir: /mnt/disk4/Kamera/Kinderzimmer/2018-01-17Jan 24 01:37:11 Unraid-Server shfs: error: shfs_rmdir, 1517: Directory not empty (39): rmdir: /mnt/disk4/Kamera/Kinderzimmer/2018-01-18Jan 24 01:42:11 Unraid-Server shfs: error: shfs_rmdir, 1517: Directory not empty (39): rmdir: /mnt/disk4/Kamera/Kinderzimmer/2018-01-17Jan 24 01:42:11 Unraid-Server shfs: error: shfs_rmdir, 1517: Directory not empty (39): rmdir: /mnt/disk4/Kamera/Kinderzimmer/2018-01-18Jan 24 01:47:11 Unraid-Server shfs: error: shfs_rmdir, 1517: Directory not empty (39): rmdir: /mnt/disk4/Kamera/Kinderzimmer/2018-01-17Jan 24 01:47:11 Unraid-Server shfs: error: shfs_rmdir, 1517: Directory not empty (39): rmdir: /mnt/disk4/Kamera/Kinderzimmer/2018-01-18 It seems like no one can help me here, so i wrote to support from xeoma, but they ask now: "Could you specify how exactly this drive is mounted in your system? Can you send us the exact commands you used?" -> I use the user share (add share "kamera", cache: only) What should i respond to them? Quote Could you specify how exactly this drive is mounted in your system? Can you send us the exact commands you used? Edited January 24, 20188 yr by nuhll
January 24, 20188 yr 9 minutes ago, nuhll said: Jan 24 01:37:11 Unraid-Server shfs: error: shfs_rmdir, 1517: Directory not empty (39): rmdir: /mnt/disk4/Kamera/Kinderzimmer/2018-01-17 Those messages are informational. You must have a share called "Kamera" and some process is trying to delete the directory "Kinderzimmer/2018-01-17" on that share, but that directory is not empty (has files and/or directories still underneath it). Please post diagnostics for more info.
January 24, 20188 yr Author Yes, thats what i already know. Problem is, it shouldnt delete because max age or max size is not reached. And yes, like i said i have an share kamera with cache yes. What can i tell them to "Could you specify how exactly this drive is mounted in your system? Can you send us the exact commands you used?" Heres diag: unraid-server-diagnostics-20180124-0204.zip Thats currently my only left problem, besides this unraid runs like a cat. Im just sad i didnt found this earlier... Edited January 24, 20188 yr by nuhll
January 24, 20188 yr Doubt that it will help. From your syslog: Jan 18 17:33:18 Unraid-Server emhttpd: shcmd (66): mkdir -p /mnt/disk4 Jan 18 17:33:18 Unraid-Server emhttpd: shcmd (67): mount -t xfs -o noatime,nodiratime /dev/md4 /mnt/disk4 But /dev/md4 is not exactly the physical disk. It is disk4 including parity code in the driver.
January 24, 20188 yr Author Where does that come from? This error only appears when running the xeoma docker.
January 24, 20188 yr From your syslog as I said. That is unRAID mounting disk4 when the array starts.
January 24, 20188 yr Author Eh,so tehre is a problem with my disk 4 or what? What i can do?! "Where does that come from?" Means: Why does this happens Edited January 24, 20188 yr by nuhll
January 24, 20188 yr Author Thats all i could edit... Is this correct? Edited January 24, 20188 yr by nuhll
January 24, 20188 yr Author First is camera name, then "root" for camera files, right is the name of the directory. And its working, its creating files... Edited January 24, 20188 yr by nuhll
January 24, 20188 yr As experiment in your case, please remove or disable that Cache Dirs plugins. Next, pick one of those directories which it complains it can't remove - if you browse the share, do you see files in there?
January 24, 20188 yr Author Hmm, i try it, yes there are files, but what i wonder... the files have different names?!?!?! I wait some minutes to look if the error still pops up. Edit: All directorys have movie files in it. (or the folder which has the movie files) Jan 24 02:40:24 Unraid-Server cache_dirs: Stopping cache_dirs process 21202Jan 24 02:42:38 Unraid-Server shfs: error: shfs_rmdir, 1517: Directory not empty (39): rmdir: /mnt/disk4/Kamera/Kinderzimmer/2018-01-17Jan 24 02:42:38 Unraid-Server shfs: error: shfs_rmdir, 1517: Directory not empty (39): rmdir: /mnt/disk4/Kamera/Kinderzimmer/2018-01-18Jan 24 02:42:38 Unraid-Server shfs: error: shfs_rmdir, 1517: Directory not empty (39): rmdir: /mnt/disk4/Kamera/Kinderzimmer/2018-01-19Jan 24 02:42:38 Unraid-Server shfs: error: shfs_rmdir, 1517: Directory not empty (39): rmdir: /mnt/disk4/Kamera/Kinderzimmer/2018-01-20Jan 24 02:42:38 Unraid-Server shfs: error: shfs_rmdir, 1517: Directory not empty (39): rmdir: /mnt/disk4/Kamera/Kinderzimmer/2018-01-21Jan 24 02:42:38 Unraid-Server shfs: error: shfs_rmdir, 1517: Directory not empty (39): rmdir: /mnt/disk4/Kamera/Kinderzimmer/2018-01-22Jan 24 02:42:38 Unraid-Server shfs: error: shfs_rmdir, 1517: Directory not empty (39): rmdir: /mnt/disk4/Kamera/Kinderzimmer/2018-01-23 Edited January 24, 20188 yr by nuhll
January 24, 20188 yr Author So i guess, thats the problem. I told them i cant change that and thats how unraid works and asked them if they make their product work. Lets see what happens. They answered: Quote Hello! The mounting includes these parameters: noatime and nodiratime We recommend mounting without them, as these may confuse Xeoma’s archive system. Sincerely, Paul VIP Partners Manager, FelenaSoft Company Edited January 24, 20188 yr by nuhll
January 24, 20188 yr 2 hours ago, nuhll said: So i guess, thats the problem. I told them i cant change that and thats how unraid works and asked them if they make their product work. Lets see what happens. They answered: I doubt that's the problem. atime/diratime refer to the "last accessed" time stamp, which tells you the last time a file or directory was merely opened for reading. Another possibility is that those messages might be normal. That is, normally a program could remove a directory using this command: rmdir /some/dir/path And if the directory is not empty you get this message: rmdir: failed to remove '/some/dir/path': Directory not empty Knowing this, a program can do this instead: rmdir /some/dir/path >/dev/null which makes the shell throw the output of the command away. A lot of programs could be coded this way. In normal SMB-type access this should never happen though because Samba doesn't try to delete a directory if it knows it's not empty. Hence we coded 'shfs' to output a warning message if this ever happens. Also, we output the message because it gives a clue if a file or directory remains on the cache after the mover runs. So my question to you is: Does the application seem to be working correctly, and you are just annoyed by the syslog messages? If so we can consider inhibiting that message (and similar).
January 25, 20188 yr Author Hi, it seems the program is working fine besides this errors. But im not sure if deleting works. Ive set it to 100gb and 7 days. Today is 25. and i still have videos from 18. - so i guess i wait another day. I work with linux since over 10 years, i know this error occurs when the directory you want to delete is not empty. - but thanks. rm -r would work EDIT: deleting archives seem to work! Ive set to 5 days now, and it deleted until 20. Edited January 25, 20188 yr by nuhll
January 25, 20188 yr Same problem here. Xeoma docker looks like is running well, but I have that errors in error log.
January 25, 20188 yr 5 hours ago, nuhll said: Hi, it seems the program is working fine besides this errors. But im not sure if deleting works. Ive set it to 100gb and 7 days. Today is 25. and i still have videos from 18. - so i guess i wait another day. I work with linux since over 10 years, i know this error occurs when the directory you want to delete is not empty. - but thanks. rm -r would work EDIT: deleting archives seem to work! Ive set to 5 days now, and it deleted until 20. Ok I'll see about quieting down the messages.
January 25, 20188 yr Author Cool would be like Display the error one time, and if it comes more then X times in X minutes/hours/days just summarize it, like Jan 24 02:42:38 Unraid-Server shfs: error: shfs_rmdir, 1517: Directory not empty (39): rmdir: /mnt/disk4/Kamera/Kinderzimmer/2018-01-17 *there were 12312 more entries like this (clickable to expand and show all) Edited January 25, 20188 yr by nuhll
January 25, 20188 yr 3 minutes ago, nuhll said: Cool would be like Display the error one time, and if it comes more then X times in X minutes/hours/days just summarize it, like Jan 24 02:42:38 Unraid-Server shfs: error: shfs_rmdir, 1517: Directory not empty (39): rmdir: /mnt/disk4/Kamera/Kinderzimmer/2018-01-17 *there were 12312 more entries like this (clickable to expand and show all) That's a great way to consume additional RAM.
January 25, 20188 yr Author Because we have RAM problems thesedays. Wheres difference to now compared to store it expandable... Oo Like 1kb more ram? cmon.
January 25, 20188 yr 42 minutes ago, nuhll said: Because we have RAM problems thesedays. Wheres difference to now compared to store it expandable... Oo Like 1kb more ram? cmon. Let a process perform an attempt to remove 10000 directories, where each path is 100 characters, then the string data itself is 1 MB excluding time stamps etc. And becomes at least 2 MB with the overhead of heap management and some reasonable data structure to manage the entries. So more like 2000 times more RAM than your guess. And about 0.1% of the total amount of memory for a system with 2GB RAM. And that expansion can't take advantage of any storage space reserved for a normal log data ring buffer. Not to forget that the /var/log/ files are linear and rsyslogd must make the decision on repeated entries on-the-fly when emitting data to the files. We have to pay lots of money and electricity bills because of developers who thinks RAM is for free. These are the same developers who never spends time reading up on cache memory, memory bandwidth etc. One of the selling points of unRAID is that it is lean.
January 28, 20188 yr 1 hour ago, nuhll said: @limetech Did you suspent the messages in the new version? From 6.4.1 release notes: Quote shfs: less verbose logging
Archived
This topic is now archived and is closed to further replies.