Closed ghost closed 2 years ago
I'm seeing some groups taking a looooong time. They also fail to insert on the first run, then seem to succeed on the second run, when repairing:
Updated group alt.binaries.hdtv.x264
Attempting to repair 20,000 parts.
-
0s to download articles, 0s to process collections, 0s to insert binaries/parts, 0s for part repair, 1s total.
-
514s to download articles, 19s to process collections, 0s to insert binaries/parts, 0s for part repair, 533s total.
-
220s to download articles, 19s to process collections, 0s to insert binaries/parts, 0s for part repair, 239s total.
.
483s to download articles, 22s to process collections, 0s to insert binaries/parts, 0s for part repair, 506s total.
20,000 parts repaired.
Updated group alt.binaries.hdtv.x264
Received 20,000 articles of 20,000 requested, 0 blacklisted, 0 not yEnc.
20000 articles failed to insert! [Blacklight\Binaries::scan]
2116s to download articles, 83s to process collections, 0s to insert binaries/parts, 0s for part repair, 2199s total.
Updated group alt.binaries.x264 813257342 813277341 77
This is in comparison to other groups that take no time to download and process 20k articles:
Received 20,000 articles of 20,000 requested, 0 blacklisted, 9066 not yEnc.
0s to download articles, 0s to process collections, 0s to insert binaries/parts, 0s for part repair, 1s total.
Updated group alt.binaries.dvd 6887535424 6887555423 8
Received 20,000 articles of 20,000 requested, 0 blacklisted, 362 not yEnc.
1s to download articles, 0s to process collections, 2s to insert binaries/parts, 0s for part repair, 4s total.
Updated group alt.binaries.blu-ray 4457930149 4457950148 14
Received 20,000 articles of 20,000 requested, 0 blacklisted, 2021 not yEnc.
0s to download articles, 0s to process collections, 3s to insert binaries/parts, 0s for part repair, 4s total.
Updated group alt.binaries.multimedia 12393080493 12393100492 5
I tried to dig into the nttp code but unfortunately, I'm not very familiar with how it all works. I tracked the delay to a while loop in Net_NNTP_Protocol_Client->_getTextResponse()
but am not sure how to resolve the issue.
I noticed this behaviour, but it stopped on my end couple of days ago, so i am not sure what is the cause. Are you running latest master?
I last updated on Feb 12. I just updated now to see if the latest upstream updates fix the issue for me as well. My suspicion is that the groups got hit with a massive load of spam that is taking a while to work through. If you got through it first, it would make sense because I ended up disabling my system for a few of days while I tried to debug the issues with part failures, etc.
This happens on updating binaries a lot. Tmux binaries pane shows:
Query error occurred.
. Can't find the actual error though.The error gets thrown after update binaries hangs without output for a while. Sometimes more than a minute. No INSERT queries related to binaries shown in MariaDB when it hangs.