Seriously Trumpy,
You close my thread without answering the question in any of your replies, you explain to me that a drive disconnect is noticed by plex, and turn the question around asking me how plex will know its reconnected.
If it is notified of its disconnect, it is only logical it can be notified of its re-appearance.
I am a paying customer, you do absolutely NO service to plex with this kind of behaviour, so many years ive been here and never have i experienced such a rude person as you.
I just saw your post come in.
May I be of help here?
@ChuckPA said:
I just saw your post come in.
May I be of help here?
I was really just looking for a short, clear answer to what i thought was a simple question.
This is the thread i am talking about;
https://forums.plex.tv/discussion/322006/ubuntu-with-windows-nfs-share-behaviour-of-trashed-items#latest
The scenario is simple, if i disconnect an NFS share, plex will notice and mark items as trashed without plex running a library scan, but if i reconnect the nfs share, plex will not notice and correct the trash icon unless i run a library scan.
The question, which i definately do not believe was answered is;
Why is plex able to mark items as trash without a library scan, when it is not able to unmark them unless a library scan is initiated.
What triggered it was the iNotify at the mount point . That’s what marked them unavailable.
Why they didn’t rescan is unknown.
I have one urgent thing to do right now. I’m almost done with that and then I’ll figure this out. If there’s an iNotify problem, I’ll deal with it immediately.
I want to point out that i do not have the scan automatically when changes are detected enabled, i scan just at intervals.
So the items are being marked as trash without a scan happening first, and they are not being unmarked again when the mount becomes available, unless i either scan manually, or wait for the scan interval to scan.
I realise inotify dont work reliably over NFS shares, this is why i scan at intervals.
Thank you for the assistence.
@RyanDungeyMX said:
I realise inotify dont work reliably over NFS shares, this is why i scan at intervals.
There is your scan happening. How often do you let it scan the library?
Keep also in mind that any access to an item, (be it a [shared] user or one of the nightly maintenance jobs) does also notice if the file is unavailable.
Normally, Plex will ignore it if a whole drive or network share is unavailable. It will then do no changes to the library.
Depending on how the media storage is mounted, this may not work under all circumstances.
But @ChuckPA is more qualified to answer detailed questions about this.
Where iNoitify is enabled or not, the mount point will get the ‘stale file handle’ flag and PMS will be forced to close it. This alone is enough to ‘trash can’ the media.
May I have your log files?
@ChuckPA said:
Where iNoitify is enabled or not, the mount point will get the ‘stale file handle’ flag and PMS will be forced to close it. This alone is enough to ‘trash can’ the media.
May I have your log files?
I checked just now by manually umounting the nfs, which then showed them as unavailable(but did not show a trash icon), when remounting it properly showed them as available.
Whats odd here, is that sometimes when i get to work and i open plex, i see trash icons on maybe half my movies library, if i check the nfs mount its available, and all movies are located on the same nfs share so it makes no sense that only some of them are trashed.
If i run a scan manually, it immediately corrects them. I dont understand why its happening.
The incident described above i noticed again this morning, around 0900 GMT ish
Logs attached.
Thanks.
Please detail how your networking is laid out. I am finding errors where PMS cannot read /var/lib/plexmediaserver which implies it’s on a network (NFS) connection.
I’m also seeing errors getting out to plex.tv, fanart.tv and others. This has been happening at least since the 28-May
com.plexapp.agents.localmedia.log.4:2018-05-30 13:21:48,682 (7fbc23fff700) : ERROR (model:205) - Cannot read model from /var/lib/plexmediaserver/Library/Application Support/Plex Media Server/Metadata/TV Shows/6/b7dc1b52f38e762653f8f4faa3dc6dceec5a866.bundle/Contents/com.plexapp.agents.localmedia
com.plexapp.agents.localmedia.log.4:2018-05-30 13:21:51,583 (7fbc23fff700) : ERROR (model:205) - Cannot read model from /var/lib/plexmediaserver/Library/Application Support/Plex Media Server/Metadata/TV Shows/f/669f21ae31bac118cd280f62ba1b6d1cde1d55e.bundle/Contents/com.plexapp.agents.localmedia
com.plexapp.agents.localmedia.log.4:2018-05-30 13:21:54,418 (7fbc23fff700) : ERROR (model:205) - Cannot read model from /var/lib/plexmediaserver/Library/Application Support/Plex Media Server/Metadata/TV Shows/4/d57c3289b01a1f3331e3bdb5364596e6d6b0f41.bundle/Contents/com.plexapp.agents.localmedia
com.plexapp.agents.thetvdb.log:2018-05-30 15:49:35,203 (7f22daffd700) : ERROR (networking:219) - Error opening URL 'https://tvdb2.plex.tv/search/series?name=Mit+Luksushus'
com.plexapp.agents.thetvdb.log:2018-05-30 15:49:35,400 (7f22daffd700) : ERROR (networking:219) - Error opening URL 'https://tvdb2.plex.tv/search/series?name=Mit+Luksushus'
com.plexapp.agents.thetvdb.log:2018-05-30 15:49:38,439 (7f22daffd700) : ERROR (networking:219) - Error opening URL 'https://meta.plex.tv/tv_e/300455/mit+plastikmareridt/2015'
com.plexapp.agents.thetvdb.log:2018-05-30 15:49:38,666 (7f22d8a06700) : ERROR (networking:219) - Error opening URL 'https://meta.plex.tv/tv_e/157421/politijagt/2009'
com.plexapp.agents.thetvdb.log:2018-05-30 15:49:40,602 (7f22d9ffb700) : ERROR (networking:219) - Error opening URL 'https://meta.plex.tv/tv_e/96011/paradise+hotel/2005'
com.plexapp.agents.thetvdb.log:2018-05-30 15:49:40,711 (7f22daffd700) : ERROR (networking:219) - Error opening URL 'https://tvdb2.plex.tv/series/300455/images?lang=en'
Plex is installed on ubuntu 16.04, locally in /var/lib/plexmediaserver
My media is mounted in /media/ from a windows server running NFS server.
So the plex install is not on an NFS share, it is completely local.
As you see from the logs, the machine can’t connect to various hosts. If I were to detail them all, it is a clear indication of host networking. That’s a root problem for you to resolve first.
As for the files being marked “Unavailable”, without scanning,
- PMS does know it’s a NFS mount because the directory handle is open (
opendir() library call).
- When the NFS mount does down, the file handle is closed.
- PMS is force-notified of this event because PMS keeps the directory’s existence monitored.
- Everything on that base inode is marked even though iNotify is not active. This is the same as pulling a USB drive when the directory listing is cached.
- It will not come back until it is manually scanned again.
I dont see any issues with plex in general despite the errors you mentioned, but ill keep that in mind.
So, i still have some questions - Why is it that only some movies in my library is trashed, and others not? The are all on the same nfs share, in the same folder.
Regarding your 5th point, how come it wont come back unless manually scanned? Is there some technical limitation thats the result of this behaviour?
Thanks for the answers!
Why only some? Caching. Give the caching a bit to catch up and they will also indicate down. Between Linux FS, PMS internal buffers, and the browser app, there are all buffers which need to expire and be reloaded.
It won’t come back because once out of local filesystem cache, they’re out of memory. Linux has a finite buffer and it doesn’t go ‘looking’ unless instructed to.
Thank you, i appreciate the answers!