About start another prime + Plex scan...
Nick, that log is already huuuge!
About start another prime + Plex scan...
Nick, that log is already huuuge!
Thankfully text compresses really really well. Especially with max compression and a good dictionary size.
Scan completed without crashing this time. Mount still going strong. At midnight, my Plex server will start analyzing, so that will be another major test lasting eight hours.
No issues to report as of this morning. Build 228 has been stable while running extensive scan and media analysis all night.
Nick, if you still need the log, let me know.
Thank you @VBB and @thestigma for really doing the testing for this as I had very little luck recreating the issue
My pleasure. BTW, the log is currently 27GB Would be nice if I could delete it.
Ok, so I'm now at 90hrs continuous stress-testing (or at least something like 95% contnous) on ver 180 and 221. Still no issues to be seen from a user-perspective.
I think I will stop the focused stability testing on this now and turn my attention to some other things that require me to use other (newer) builds for testing. All I can say is this looks stable to me now.
I will still keep my debug-log going for the next week or two just to catch any surprise re-occurrence that might pop up, but I won't keep the system 24/7 testing from now.
Out of curiosity, how do you stress-test, since you don't use Plex? One thing I noticed with all the builds we've been testing up until 228 (the first one to be OK for me) is that they're all fine until I run a Plex scan. I upload close to 750GB every day, and I never encountered any issues doing so, even with the builds that would crash during Plex activities.
Good news!
Thanks for testing
It if is working I don't need the log!
That is interesting... One conjecture is that Google might be canary deploying new features so it doesn't affect all users or regions...
Stability testing on the latest beta would be much appreciated!
I've been stress-testing very differently.
I've basically just been manually hashing tons of existing files - so rclone has to read through everything to get the data - keeping it busy 24/7.
Since I already confirmed getting multiple stall occurrences doing this on earlier versions I figured it was best to stick with this same scenario for testing. It was also just convenient ;D
Running 235 now...
Build 235 is also good. No issues to report.
Thank you for testing
This topic was automatically closed 3 days after the last reply. New replies are no longer allowed.