enJOyIT Posted January 27, 2022 Share Posted January 27, 2022 (edited) Hi, a really strange problem... I have a plenty of debian linux vm machines running on proxmox which are connected to several unraid user shares. They keep unexpected losing the connection to these shares. Not every share is involved at the same time, this means sometimes share1 is losing the connection and the other time share2 is involved. I can't tell when it happens... Yesterday every machine was connected to every share and today one machine loses connection to on share. The other unraid-shares of this machine are ok as well the other machines! I moved to unraid from openmediavault and NEVER had such issues, so I pressume it's related to unraid. If this happens the mounted folder isn't readable anymore and I get "Cannot read file" and in midnight commander the folder looks like "?serien" or "?filme". I'm mounting the shares via etc/fstab: Quote //192.168.20.215/filme /mnt/filme cifs x-systemd.automount,username=plex,password=xxxxxxxx 0 0 //192.168.20.215/serien /mnt/serien cifs x-systemd.automount,username=plex,password=xxxxxxx 0 0 //192.168.20.215/musik /mnt/musik cifs x-systemd.automount,username=plex,password=xxxxxxx 0 0 Maybe there is something what I am missing? Edited January 28, 2022 by enJOyIT Quote Link to comment
enJOyIT Posted January 28, 2022 Author Share Posted January 28, 2022 (edited) Today, it happend again: No Logs in my Client?! It just dropped the connection. I transfered some files from disk to disk (both in array) minutes before (but nothing related to "serien"). Some files of "serien" were transfered to the cache dir in the same time. Is this related to that?! The only info from this timeframe i have, is from the unraid-server: Jan 28 06:04:27 unraid kernel: mdcmd (49): set md_num_stripes 1280 Jan 28 06:04:27 unraid kernel: mdcmd (50): set md_queue_limit 80 Jan 28 06:04:27 unraid kernel: mdcmd (51): set md_sync_limit 5 Jan 28 06:04:27 unraid kernel: mdcmd (52): set md_write_method Jan 28 06:09:52 unraid smbd[23687]: [2022/01/28 06:09:52.014082, 0] ../../source3/smbd/smb2_read.c:255(smb2_sendfile_send_data) Jan 28 06:09:52 unraid smbd[23687]: smb2_sendfile_send_data: sendfile failed for file xxxx/xxxxx/yyyyyyyy.zzz (Connection reset by peer) for client ipv4:192.168.20.221:36518. Terminating client ipv4:192.168.20.221 is not the client pc with the failed connection at the top, it's another! But it has the same error: strange! Edited January 28, 2022 by enJOyIT Quote Link to comment
enJOyIT Posted January 28, 2022 Author Share Posted January 28, 2022 (edited) Additional info: I reconnected the shares and then startet the mover to move some date from cache to the array and bam, the next share dropped: But no logs... on client or unraid server It must have something todo with the mover. As I reconnected (umount "filme" and mount -a) the share (which worked), the mover was still going to move files... and in the next minute "serien" dropped... "mount" on the client gives me this: //192.168.20.215/musik on /mnt/musik type cifs (rw,relatime,vers=3.1.1,cache=strict,username=plex,uid=0,noforceuid,gid=0,noforcegid,addr=192.168.20.215,file_mode=0755,dir_mode=0755,soft,nounix,serverino,mapposix,rsize=4194304,wsize=4194304,bsize=1048576,echo_interval=60,actimeo=1,x-systemd.automount) //192.168.20.215/filme on /mnt/filme type cifs (rw,relatime,vers=3.1.1,cache=strict,username=plex,uid=0,noforceuid,gid=0,noforcegid,addr=192.168.20.215,file_mode=0755,dir_mode=0755,soft,nounix,serverino,mapposix,rsize=4194304,wsize=4194304,bsize=1048576,echo_interval=60,actimeo=1,x-systemd.automount) //192.168.20.215/serien on /mnt/serien type cifs (rw,relatime,vers=3.1.1,cache=strict,username=plex,uid=0,noforceuid,gid=0,noforcegid,addr=192.168.20.215,file_mode=0755,dir_mode=0755,soft,nounix,serverino,mapposix,rsize=4194304,wsize=4194304,bsize=1048576,echo_interval=60,actimeo=1,x-systemd.automount) r What is going on here????? Edited January 28, 2022 by enJOyIT Quote Link to comment
enJOyIT Posted January 28, 2022 Author Share Posted January 28, 2022 (edited) I can reproduce this error now. It's definitely connected to the mover/cache drive. If I have files on the cache and start the mover, the connection to the share that contains the moved files drops at the end. It even drops if I just copy files (via Windows) to the mount (cache yes). Maybe it's locking the drive for a second and my linux machine thinks the mount is gone?!?! Is there any one from limetech who can analyse this behaviour? Attached the diagnostic file etc... But imho there isn't really interesting in it. Because I don't get any entry in logfiles etc... Similar threads (it's macos but the behaviour is the same): Same but on reddit: I need help! P.S. I'm getting two new SSD drives (these are my cache drives) tomrorow (my current have no trim support with my HBA - LSI SAS3008) - Maybe (but I doubt) it will help? unraid-diagnostics-20220128-0958.zip Edited January 28, 2022 by enJOyIT Quote Link to comment
enJOyIT Posted January 28, 2022 Author Share Posted January 28, 2022 Sorry for spamming, but I want to keep the posts separated for chronological order... I enabled mover-logging, but still no helpful information: Jan 28 11:48:24 unraid emhttpd: shcmd (94072): /usr/local/sbin/mover |& logger & Jan 28 11:48:24 unraid root: mover: started Jan 28 11:48:24 unraid move: move: file /mnt/cache/download/nzbget/completed/xxxxxxxxx/yyyyyyyyyy.abc Jan 28 11:49:19 unraid root: mover: finished Quote Link to comment
enJOyIT Posted January 28, 2022 Author Share Posted January 28, 2022 Is there a official support email address? I'm willing to pay for support! I'm afraid that moving to unraid was a big mistake if i can't get this problem solved Quote Link to comment
dlandon Posted January 28, 2022 Share Posted January 28, 2022 I think you are running into the stale file handle issue with CIFS mounts. What I think is happening is you are referring to a file on the cache and when it is moved the file handle has changed and you can then no longer access it. When UD (Unassigned Devices Plugin) mounts a CIFS share, it uses the 'noserverino' parameter that prevents the stale file handle. Example UD mount command: /sbin/mount -t 'cifs' -o rw,noserverino,nounix,iocharset=utf8,file_mode=0777,dir_mode=0777,uid=99,gid=100,credentials='/tmp/unassigned.devices/credentials_Public' '//MEDIASERVER/Public' '/mnt/remotes/MEDIASERVER_Public' Basically what the 'noserverino' does is use a local 'ino' and the not the server 'ino' which can change if the file is moved. Take a look at all the parameters used here and see if any others would apply to your situation. Give that a try. Let me know how it goes and we can work on it somemore. 2 Quote Link to comment
enJOyIT Posted January 28, 2022 Author Share Posted January 28, 2022 Thank you for your reply! Meanwhile I was trying the same with NFS-share... But it ended up in the same issue. Is NFS using the same mechanism? Quote Link to comment
dlandon Posted January 28, 2022 Share Posted January 28, 2022 If you use NFSv4, you won't see the stale file handles. You will have to use Unraid 6.10rc2 though for NFSv4 support. Quote Link to comment
enJOyIT Posted January 28, 2022 Author Share Posted January 28, 2022 Well... it seems that the "noserverino"-parameter does the job. But there is another issue now, i think it's related to this paramter. The unraid-CPU usage is stuck at 20% now... Is this a expected behaviour? Quote Link to comment
dlandon Posted January 28, 2022 Share Posted January 28, 2022 Glad to hear you got things working. Samba is a rather cantankerous beast and has been known to cause this sort of behavour. What version of Unraid are you using? Quote Link to comment
enJOyIT Posted January 28, 2022 Author Share Posted January 28, 2022 I am using 6.9.2 Quote Link to comment
dlandon Posted January 28, 2022 Share Posted January 28, 2022 There have been some samba changes in 6.10. One of the latest changes is the addition of samba security. Sorry I can't be more specifc. There is so much going by me in development I can't keep track of it all. If you have the stomach for it, you could run the latest 6.10rc2 or wait for 6.10 final and take a look at it then. For the moment, a wait and see approach might be best, 1 Quote Link to comment
enJOyIT Posted January 28, 2022 Author Share Posted January 28, 2022 Ok, I will have a look and maybe upgrading to the RC (depends on the release date of the final version 🙂) Thank you very much so far!!! Quote Link to comment
dlandon Posted January 28, 2022 Share Posted January 28, 2022 2 minutes ago, enJOyIT said: Thank you very much so far!!! You are very welcome. You're fairly new here, but I think you'll find the support you get on the forum is exceptional. If you ever feel you are not getting what you need, feel free to tag @Squid or myself and we'll be sure to get the right person on the job. 1 Quote Link to comment
AndrewZ Posted January 28, 2022 Share Posted January 28, 2022 Or email support if the forum isn't given you the answers you need. 1 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.