KluthR

Old backups from the older plugin version are not supported. And all newer have -failed suffix and are not available either.

Für die Interface ID bleibt die Adresse gleich. Der Rest wird mit DNS geregelt

Since this issue is known, I leave my findings here. Seems nchan can be a bit diffcult sometimes. I killed (kill -9) the main process and restarted it, using rc.nginx which worked.

Hi, my Unraid 6.12.3 stopped accepting HTTP requests this early morning. Some basic tests and I discovered, that my nginx error log is full of this: 2023/09/03 05:07:01 [alert] 2261#2261: worker process 5493 exited on signal 6 2023/09/03 05:07:01 [alert] 2261#2261: shared memory zone "memstore" was locked by 5493 ter process /usr/sbin/nginx -c /etc/nginx/nginx.conf: ./nchan-1.3.6/src/store/memory/memstore.c:705: nchan_store_init_worker: Assertion `procslot_found == 1' failed. here some lines just before this happened: 2023/09/03 04:52:21 [info] 19603#19603: Using 116KiB of shared memory for nchan in /etc/nginx/nginx.conf:161 2023/09/03 04:52:21 [info] 19603#19603: Using 131072KiB of shared memory for nchan in /etc/nginx/nginx.conf:161 2023/09/03 04:57:55 [info] 28501#28501: Using 116KiB of shared memory for nchan in /etc/nginx/nginx.conf:161 2023/09/03 04:57:55 [info] 28501#28501: Using 131072KiB of shared memory for nchan in /etc/nginx/nginx.conf:161 ter process /usr/sbin/nginx -c /etc/nginx/nginx.conf: ./nchan-1.3.6/src/store/memory/memstore.c:705: nchan_store_init_worker: Assertion `procslot_found == 1' failed. 2023/09/03 04:57:55 [alert] 2261#2261: worker process 28504 exited on signal 6 2023/09/03 04:57:55 [alert] 2261#2261: shared memory zone "memstore" was locked by 28504 Some nchan issues? Server uptime: 10:12:30 up 13 days, 20:10, 2 users, load average: 0.04, 0.08, 0.03 Never had this issue before. Memory: free -h total used free shared buff/cache available Mem: 15Gi 5.5Gi 8.0Gi 129Mi 2.1Gi 9.6Gi Swap: 0B 0B 0B The syslog has no further entries. If I should provide diags anyway, let me know. I left the server in its state for further diagnostics, if needed. EDIT: It seems, that this issue happens to some users here. Googling revelaed, that the nchan author should have fixed this already? Whoch version of nchan is Unraid using? => https://github.com/slact/nchan/issues/446#issuecomment-400874440 EDIT2: Ahh, I just see, Unraid uses latest (1.3.6) of nchan. EDIT3: Some other author comment on this: https://github.com/slact/nchan/issues/446#issuecomment-686848568 Maybe a nchan expert here could tell us more, how nchan is built?

Ich nutze nach wie vor die „—mac-address=„ option um eine feste v6 zu setzen. Was anderes ist mir nicht bekannt.

Are the xmls still existing in /boot/config/plugins/dockerman/templates-user?

But thats not needed for the macvtap solution, right? Its just needed in case Host <-> containers want to talk, like ususal?

As said in the other forum, the RAM is dead. Yes, because this feature was meant to backup not docker related files. I will see how I can change that behavior. Maybe some extra docker related extra files.

No problem. But THIS is the old thread 😅

You already posted in the right thread, or do you really use the old plugin version?

Not directly. You could set every container to ignore it. Then you end up with it only backing up USB.

Good point, have to test that.

Easy one: add /mmt/user/appdata as allowed sources as well (and remove the docker.img path). please read the hints inside the help block for the sources field.

2023.08.28 should fix the docker auto update issue.

You mapped the binhex-plex /config to "/mnt/cache/appdata/", thats why. Prove me wrong, but that seems not right. The /mnt/cache/appdata ist also listed as source directory, si its being skipped. Any other mappings are treated as external volume mappiung as discarded, resulting into no folder being backed up. While writing this, Iam on my way to find the bug. Seems, that the mentioned container is accessing data from Home-Assistant which, I guess, is running? If so, please exclude the HA mapping from that metioned container. Sidenote: Future update will implement a is-a-mapping-mapped-for-more-than-one-container? feature. That will detect multiple mappings. What the plugin does then is still open, but I feel the need for such detection. If something is cleaning up files inside /tmp, it could lead to this behavior. The log file is saved within /tmp and is being cleared usually after a reboot only. The plugin does NOT touch the docker service itself in any way! Please share a debug lof for further inspection. Yep, see above. Im on it. Seems right then. Hmm.

Could you provide a "ls -la" from inside your backup destination within the Unraid terminal?

Will looking into this the next days Please share the debug log.

You just have to set the absolute path as destination. You try to backup jellyfin within jfa-go. Please exclude the jellyfin share. "Tar said: tar: Removing leading `/' from member names; mnt/user/appdata/tdarr/logs/Tdarr_Node_Log.txt: Mod time differs; mnt/user/appdata/tdarr/logs/Tdarr_Node_Log.txt: Size differs". Something is changing the mentioned files. Some other container which has the tdarr volume attached?

I just pipe those path straight to tar. In theory, I could users define relative paths and make them absolute. However, I have to add some more love to this feature.

Please Share the debug log either by use the Share function (post UUID here) or attach the file manually.

Unpack the archive by yourself and move the folders to their correct location. Install the containers with "Previous apps" I guess.

Yea, seems so. Maybe I should check and skip those folders?

What happens if you hit „Manual backup“?

I personally have issues stopping the guacd properly at it sometimes does not respond to SIGTERM. Even the SIGKILL is ignored.

In the per container settings. Also, Please use the new thread for the new plugin instead.