@ChuckPa
I followed your steps exactly, again 
I made sure to remove every file called âHTTPCookiesâ within the agent folders.
I restarted PMS and did a Refresh All Metadata, and gave it 2 hours. It seems to be done - the condition continues.
download the logs now please. Let me see what has transpired.
Plex Media Server Logs_2019-12-16_22-16-27.zip (8.0 MB)
attached.
I canât explain it.
I see the urllib2 errors and am told they mean nothing because they point to âresourceHashesâ which were removed.
json = self._core.networking.http_request("http://127.0.0.1:32400/:/plugins/com.plexapp.system/resourceHashes"
Other than this, I have no answer. Iâm at a loss.
One last experiment if you have time.
- Stop Plex
- Rename
LibrarytoLibrary.save - Setup a âtestâ server â expressly with the Friendly Name âtestâ (avoid colliding with your main server name)
- Setup one section.
- See what this does.
If this works then we have proof the database canât handle the changes.
We can move over all your watch history.
@ChuckPa As mentioned Iâm having the same issue with my posters. They havenât been loading automatically so Iâve been assigning them manually for weeks. I deleted the file as you mentioned above and almost all my posters are blank despite showing a match when editing. Iâve attached my logs.Plex Media Server Logs_2019-12-16_22-58-27.zip (5.3 MB)
Because this problem is persisting, Iâm now compiling a list of threads where itâs occurring. I see the errors, which are purported to not be errors. I want Engineering to address them.
I will escalate this in my Thursday staff meeting.
Would you mind recreating the fault since itâs so predictable for you?
Engineering is monitoring this thread and looking for the root cause.
Your logs should be DEBUG (on), VERBOSE (off) as usual.
Thanks
Sure can.
Iâve reset the Shield to factory, and updated all apps to current via the Play Store. SD card is erased and adopted; Plex DB storage is internal, user accessible. No other PMS server settings changed from default.
Shield version: 8.0.1
PMS version: 1.18.3.2156
Behaviour is the same as previously posted. Media scan on a small test movie folder started at 12:27pm local time, and stopped scanning/updating 10 minutes later, around 12:37pm local.
Client performance is currently very sluggish, and is having trouble reading even the posters which have already been downloaded. Library browse from the client should begin around 12:52pm local in the logs. Note that during the 10 minutes the library update was progressing from 12:27pm-12:37pm, client performance was normal.
Sometime yesterday, my issues have apparently resolved. Movies without posters suddenly gained posters, and new additions get the posters automatically and correctly.
I wonder if the HTTPCookies removal was indeed the entire fix, and the sheer size of my library meant that it took a couple days to sort itself out? (I suspect my library is much larger than the average library.)
I will report back if anything changes.
@cmackOTI @ShadeRF
Ensure that you are removing EVERY copy of HTTPCookies from each subfolder. Not just from com.plexapp.system as one of the original posts seemed to indicate.
This problem is still evolving. Weâre discovering more and more at the lower levels of whatâs happening.
One I looked at was for a user that migrated from windows some months back
Note on windows end of line is CR LF
on Linux it is LF
A change recently came to one of the http code libraries in Plex Media Server that added some validation and tripping over the extra CR in the CR LF that came in from when PMS was on windows
These cookies appear to have long period before expiry - at least 6 months - may be a year - not sure
To see if there is a CRLF one needs to look at these files using a hex viewer or Notepad++ with option to view all symbols
Application Support/Plex Media Server/Plug-in Support/Data/com.plexapp.agents.thetvdb/HTTPCookies
Application Support/Plex Media Server/Plug-in Support/Data/com.plexapp.system/HTTPCookies
1 Like
After updating to Version 1.18.4.2171 any newly added movie would not pull any poster or metadata. After following ChuckPAâs instructions my issue have been resolved.
I am trying to establish if everyone that had the issue had migrated the server from windows sometime in the past months. That would explain the CR LF in the HTTPCookies files
My server has only ever been on Synology. There was no migration of any kind here.
That is concerning - because if we are getting CR LF at the end of the cookie lines on Synology - it means that the issue may come back.
Please be on the lookout for this issue coming back again - the symptoms in the logs show thetvdb agent HTTPCookies having had the \r = the âCRâ in âCR LFâ. On Linux systems end of line is just LF
In your logs this related to the HTTPCookies file for a com.plexapp.system request
ValueError: Invalid header value '__cfduid=de63f8f488bc73b8e0ca9e5fe799b42b71556421494\r'
and this related to a com.plexapp.agents.thetvdb request
2019-12-16 11:41:01,556 (7fe0c1aa4700) : DEBUG (networking:143) - Requesting 'https://artworks.thetvdb.com/banners/graphical/5c16b1ebdb2cc_t.jpg' 2019-12-16 11:41:01,567 (7fe0c1aa4700) : INFO (__init__:1079) - Invalid header value '__cfduid=dde5a96f45aed52383dd346851b6d95d61551110821\r'
Could you examine your current HTTPCookies files to see how the lines end - if they end just with Hex 0A i.e. LF or not
Iâve been on the same Windows machine for several years. The issue was only observed after updating. I donât think I had added any new shows, only movies, hence only affecting my recently added movies.
I just wanted confirm that deleting the cookies files resolved the issue. If there is any additional information that youâd want please just let me know.
So just on windows and your issue only noticed now after updating to 1.18.4.2171 which was released yesterday
Did you save any logs for when you had the problem?
For what itâs worth my issues started shortly after moving from Windows to Unraid. Removed all httpcookies files and that just made all my posters blank and they never corrected so I restored to a backup and I just have to manually set every single poster. So annoying.
Iâve been having this issue for weeks now, itâs becoming rather annoying.
One thing I have discovered is that my plex webpage does have poster image links when I inspect element:
(Link highlighted in screenshot: http://image.tmdb.org/t/p/original/mzdDEuzFOPlfIHmln32JCzNtXPx.jpg)
In this example, a link can be seen which goes to the poster for the film, then at the bottom it also has a generated screenshot from the film showing up differently. All the boxes on this page appear to have posters in but none are being shown.
For me this issue came up after I reinstalled windows (had about a 2 week downtime). I assumed the bug was my configuration but Iâve done two reinstallations since, tried plex pass and non plex pass downloads and still have it.
I also attempted any suggested fixes in this thread but ended with the same results.
Please let me know if I can provide anything to assist in debugging this issue.
Many thanks
1 Like
