Bug #22531
open
Rudder agent check doesn't clean everything when inventory is stalled
Added by François ARMAND about 1 year ago.
Updated 8 days ago.
Description
In #11102 we corrected some part of the problem, but there's still some leaf processes remaining when inventory is killed, cf: https://issues.rudder.io/issues/11102#note-49
- replace /sbin/fdisk by a shell script "sleep 10000"
- run rudder agent inventory, it get blocked on the false fdisk
- after some minutes (5 I think), inventory and children processes get killed
- but the sleep 10000 remains
- Related to Bug #11102: Parent fix does not work: Fusioninventory is not tracked by check-rudder-health added
sleep uses a syscall, that is not interrupted, this is expected.
After the end of sleep it will terminate.
Replace sleep with 100 sleep of 1mn and it should work.
- Target version changed from 7.2.6 to 7.2.7
- Target version changed from 7.2.7 to 7.2.8
- Target version changed from 7.2.8 to 7.2.9
- Target version changed from 7.2.9 to 7.2.10
- Target version changed from 7.2.10 to 7.2.11
- Target version changed from 7.2.11 to 1046
- Target version changed from 1046 to 7.3.8
- Target version changed from 7.3.8 to 7.3.9
- Target version changed from 7.3.9 to 7.3.10
- Target version changed from 7.3.10 to 7.3.11
- Target version changed from 7.3.11 to 7.3.12
- Target version changed from 7.3.12 to 7.3.13
- Target version changed from 7.3.13 to 7.3.14
- Target version changed from 7.3.14 to 7.3.15
- Subject changed from Rudder agent health doesn't clean everything when inventory is stalled to Rudder agent check doesn't clean everything when inventory is stalled
Also available in: Atom
PDF
Updated by 
Updated by 
Updated by 
Updated by