also would like to know how soon after PMS launch was vmmap connected to the process - to allow me to convert relative time to actual and logs
Ok Jefe! 
I will provide you with a VMMP file once it’s above 1GB (atm it’s at 280MB something) and I started VMMP approximately 3-10 minutes after I restarted the PMS (where I adjusted the Registry to activate LogNumFiles = 10). Approximately at 17:52 o’clock 
Thanks a bunch!
Just a quick update. I started a refresh for all movies again and it got stuck on the third item again 
I hoped on another outcome, as I did some adjustments to the RAM in my BIOS, but apparently this didn’t help, unfortunately.
Hi
I’m having exactly the same problem, I originally had all the files updated but randomly for some reason, one day some meta data was missing. When I try to refresh the meta data it sticks on the first three.
+1 for me. Thought I was going crazy. My root issue is that cast photos tend to randomly disappear. I have a lot of movies in my library (over 1000), so I figured the “Refresh All Metadata” was the best option. Everytime I have done this, the first 3 movies in the library remain in a perpetual spin. I have left the PC and PMS running over multiple days, and they are still spinning. The cast photos have not been updated, so I can only assume that the metadata was not refreshed.
Side note: I realize the cast photo issue is a separate issue and I’ve been tweaking agent settings to get that resolved without luck.
If I navigate to an individual movie and attempt the refresh - all goes smoothly. It is only when using the “Refresh All” option that this happens. Attached are the logs. Not sure how far back they go, but I can tell you that I started the refresh late last night, after 11pm EST 2020-06-29.
Update: Looking at the “Alerts” page shows that PMS attempted to download metadata for all movies in the library. This took all of about 2 minutes. Not sure whether metadata was actually refreshed/re-downloaded or not. The first 3 library items are still spinning, despite the completion of this operation. So - that may just be a visual bug. More importantly, I’m fairly certain that the operation didn’t actually yield any results. There are plenty of movies where it’s clear the metadata was not refreshed.
1 Like
Could you have a look at the windows SYSTEM event log with eventvwr.exe
Select: Windows Logs → System
Then “Filter Current Log”
and in the box where it says “” type in 4227 and hit OK
See if you find any
If you do, please export the filtered events to csv file and attach here
I would like to see if the scanner hit the tcp port exhaustion issue
I think you did hit this problem - because these log lines are one of the manifestations of the problem
Jun 30, 2020 00:41:15.666 [19364] ERROR - Error issuing curl_easy_perform(handle): 7
Jun 30, 2020 00:41:15.666 [19364] WARN - HTTP error requesting GET http://127.0.0.1:49929/system/agents/update?mediaType=1&force=1&respectTags=0&guid=com%2Eplexapp%2Eagents%2Eimdb%3A%2F%2Ftt0143808%3Flang%3Den&id=11072 (7, Couldn't connect to server) (Failed to connect to 127.0.0.1 port 49929: Address already in use)
We could tweak the dynamic tcp port pool in windows (doubling its size) and adjust TIME_WAIT to try and avoid this issue with large libraries - or you could try the new movies scanner, which after a forum preview release, is now out as beta
See Release Note for beta 1.20.0 here
and the first item
- New improved movie agent. (More info here )
The new scanner runs in-line and so avoids the high use of tcp ports that the old scanner with its separate process was using up
Hey sa! Thanks for the updates and I’m happy to see a new PMS version and I’m curious whether my issues have been resolved now. For me personally, as you’re the only one who is very well aware of my situation, should I still stick to the altered Windows settings? Were there any Curl updates in the meantime? And: Do you think I can analyze my TV show library now all at once?
Have a great day!
The new version of Plex Media Server should avoid the problem for Movies libraries with the new in-line scanner.
TV Shows still use the old scanner and will therefore still have the problem.
There was curl change which reduced the number of ports used for TV Shows library analysis by 66% but this was not sufficient to resolve the problem and the tweaks you have for Windows TCP dynamic ports pool size and TIME_WAIT will need to remain in place. The curl change has been pulled back as it led to some performance issues with sftp and a new change is awaiting approval, testing and release by curl and then it would be incorporated into Plex Media Server. But as before this would not on its own resolve the issue with analysis of large TV Shows libraries and would still need the tweaking of the Windows TCP Parameters.
1 Like
Could you have a look at the windows SYSTEM event log with eventvwr.exe
Select: Windows Logs → System
Then “Filter Current Log”
and in the box where it says “” type in 4227 and hit OK
See if you find any
If you do, please export the filtered events to csv file and attach hereI would like to see if the scanner hit the tcp port exhaustion issue
Attached. It appears that you are correct.
4227-error-log.7z (5.9 KB)
This isn’t a large enough priority for me to attempt changing settings in Windows, as long as I know it’s on the table and being worked on. I’m happy to wait for a more permanent solution. (I don’t refresh all very often).
Just thinking out loud here: couldn’t there be a scheduled task that refreshes meta data of a certain portion of the library, then resumes the following day? Or better yet, maybe the UI could allow us to batch select items (say, everything starting with the letter A) and then just allow us to refresh the metadata for the selection?
As it is, I end up refreshing individual items as I find issues, but then it seems like a week or so later, the library item needs to be refreshed again. (Usually I notice when the cast photos are missing). I can’t recall whether this behavior applies to TV Show libraries, but I know for a fact it applies to Movie libraries.
This topic was automatically closed 90 days after the last reply. New replies are no longer allowed.
