when cpost is running as a daemon, and is idle for a long time, it shows up as hung to sr3.

MetPX / sarrac

C implementation of (a subset of) Sarracenia (large scale file transfer utility)

GNU General Public License v2.0

4 stars 1 forks source link

when cpost is running as a daemon, and is idle for a long time, it shows up as hung to sr3. #157

Closed petersilva closed 1 month ago

petersilva commented 2 months ago

sr3 status lists it is hung
sr3 sanity "fixes" it by restarting... every ten minutes...

petersilva commented 2 months ago

sanity log sample:

[2024-09-23 17:21:03] sanity:
[2024-09-23 17:21:03] found hung cpost/bulletins_from_sundew/1 pid: 65176
[2024-09-23 17:21:09] .killing hung processes... (no point in SIGTERM if it is hung)
[2024-09-23 17:21:09] missing: [['cpost', 'bulletins_from_sundew', 1]]
[2024-09-23 17:21:09] starting them up...
[2024-09-23 17:21:09] no stray processes found

petersilva commented 2 months ago

have patch that prints:

2024-09-23 15:46:57,606 [INFO] no directory changes in last 63 seconds

only when it's idle... It's very coarse, the actual limit is 60 seconds, but it can happen upto 10 seconds later...