Kplex stopping when DNS lookup 8000 failed #49
Comments
|
Thanks! That looks like a SEGV which is disturbing and definitely a bug: thanks for finding it: I'll look into it |
|
Thanks again for reporting that: it was a bad bit of quality control on my part. I've pushed a change to the develop branch which should fix this. Please give it a go and let me know if it works. And thanks again! |
|
No problem! Glad to be able to contribute! Will update. Since this was the first and only issue in 4 months there will perhaps not be a quick feedback on the fix :) |
|
Thanks. I recreated the problem using packet filtering to simulate a dropped network connection and confirmed that the fix resolved things. The downside to the new behaviour is that if name resolution is not possible when kplex reconnects a persistent interface, it will wait until DNS does become available before reconnecting but this does at least now work as intended rather than segmentation faulting. I'll leave this open for a couple more days then close it. If you see the problem again do please re-open or open another one. |
An issue last night where kplex stopped because of what seems a DNS lookup issue. Had to do a restart of the service. It has been running fine for at least 3-4 months so this came sort of out of the blue.
Running version 1.4-17-g6e1f698 (which was an experimental branch with kplex doing new lookups if disconnected if i remember correctly)
Dec 01 01:33:18 host1 kplex[443]: Lookup failed for host ais.domainname.com/service 6344: Temporary failure in name resolution
Dec 01 01:33:18 host1 systemd[1]: kplex.service: Main process exited, code=killed, status=11/SEGV
Dec 01 01:33:18 host1 systemd[1]: kplex.service: Failed with result 'signal'.
The service stopping could perhaps be solved with a Restart=on-failure or Restart=always in the service file, but failing in the first place is perhaps something that should not occur just because of a failed DNS lookup.
The text was updated successfully, but these errors were encountered: