Server Version#: 1.25.8.5663
Player Version#: Shield TV (Android) Current version
I have been getting a lot of server crashes recently, which seem to be related to media scanning after the DVR has recorded something. It can be used all day playing things, but in the evening when the DVRis used more it will crash, not after every recording but pretty much every day at least once.
The crash logs folder is empty. The only odd thing in the logs is the media scanner log, where there is a 408 (timeout) error and a message about not being able to allocate a changestamp.
Apr 01, 2022 19:00:11.410 [3588] DEBUG - [MI] Closing input file: "G:\Recordings\Richard Osman's House of Games (2017)\Season 05\Richard Osman's House of Games (2017) - S05E120 - Richard Osman's House of Games.ts"
Apr 01, 2022 19:00:11.410 [3588] DEBUG - Downloading document http://127.0.0.1:32400/library/changestamp
Apr 01, 2022 19:00:11.410 [3588] DEBUG - HTTP requesting GET http://127.0.0.1:32400/library/changestamp
Apr 01, 2022 19:00:16.433 [2552] DEBUG - [HttpClient] HTTP simulating 408 after curl timeout
Apr 01, 2022 19:00:16.433 [3588] ERROR - HTTP 408 downloading url http://127.0.0.1:32400/library/changestamp
Apr 01, 2022 19:00:16.464 [3588] ERROR - Exception inside transaction (inside=1) (E:\jenkins\server\3587903764\Library\MediaItem.cpp:1799): Unable to allocate a changestamp from the server
Apr 01, 2022 19:00:16.464 [3588] ERROR - Exception thrown during analysis: Unable to allocate a changestamp from the server
Apr 01, 2022 19:01:39.052 [2552] DEBUG - [HttpClient] HTTP simulating 408 after curl timeout
Apr 01, 2022 19:01:39.052 [2852] DEBUG - HTTP requesting GET http://127.0.0.1:32400/:/metadata/notify/changeItemState?librarySectionID=17&metadataItemID=254694&metadataType=4&state=-1&parentID=-1&rootID=-1&mediaState=thumbnailing
Apr 01, 2022 19:03:09.065 [2552] DEBUG - [HttpClient] HTTP simulating 408 after curl timeout
Yes we have an uploaded crash report for April 1st at 19:00
The crash reports get removed from the directory when processed
A copy gets saved in %TEMP%
Could you try out the latest beta 1.25.9.5721 which fixes a number of crashes when handling exceptions and if the server continues to crash please get me fresh debug server logs and also in case there is an issue with the crash report uploads, look for a copy of the crash dump file in %TEMP% - the crash dump file copied to %TEMP% does not have any filename extension and would be a string of random characters with dash separators. You would identify it by checking when it was written
The server is running the released 1.25.9.5721 now, and is still crashing,
It’s definitely related to the DVR recording - it’s crashed twice today after finishing a recording.
I’ve found and attached (zipped) the latest of the files from the TEMP folder. It looks like a binary of some sort. I have also attached the logs from the time it crashed last.
Thanks for the crash report. A potential fix has been produced. I will add you to the Private Message thread giving access to development build 1.26.1.5744 to confirm the fix
Thanks. I thought I’d fixed it, I redid the tuner channels and the DVR setup in Plex and recreated all the schedules, it worked for 2 days but crashed again today.
The last thing in the log is starting a transcode job. I wasn’t playing anything at the time so it must have been the DVR - I don’t know what it was doing to need a transcode.
