After upgrading to NetVault 12.2 under CentOS and RHEL 7.5 and 7.6, the nvsched process (Schedule Manager) crashes several times on startup.
- The segfault from nvsched generates a core dump.
- /var/log/messages files show:
++++++++++++++++
Mar 13 17:35:52 localhost kernel: nvsched[1646]: segfault at 7f5d00000000 ip 0000003de0047e2c sp 00007fffb7172220 error 4 in libc-2.12.so[3de0000000+18a000]
Mar 13 17:35:52 localhost abrtd: Directory 'ccpp-2019-03-13-17:35:52-1646' creation detected
Mar 13 17:35:52 localhost abrt[1702]: Saved core dump of pid 1646 (/usr/netvault/bin/nvsched) to /var/spool/abrt/ccpp-2019-03-13-17:35:52-1646 (18874368 bytes)
Mar 13 17:35:53 localhost abrtd: Executable '/usr/netvault/bin/nvsched' doesn't belong to any package and ProcessUnpackaged is set to 'no'
Mar 13 17:35:53 localhost abrtd: 'post-create' on '/var/spool/abrt/ccpp-2019-03-13-17:35:52-1646' exited with 1
Mar 13 17:35:53 localhost abrtd: Deleting problem directory '/var/spool/abrt/ccpp-2019-03-13-17:35:52-1646'
Mar 13 17:35:53 localhost kernel: nvsched[1728]: segfault at 7f8400000000 ip 0000003de0047e2c sp 00007fff40734210 error 4 in libc-2.12.so[3de0000000+18a000]
Mar 13 17:35:53 localhost NetVault[1604]: NetVault: Client: 'localhost' Class: 'System' Job: 'N/A' Warnlevel: 'Severe Error' Msg: 'Process 'Schedule Manager' (PID '5') has crashed!'
++++++++++++++++
- NetVault logs show that the nvsched process tries to load scheduled phases and crashes:
---
Loading scheduled phases, if any
Process 'Schedule Manager' (PID '5') has crashed!
---
nvsched also leaves a NetVault dump which shows a problem with the db query results:
---
6 DBCONNCT:11079 25 0 123533.245067 (1,0/0){SELECT * FROM scheduledphaseinfo ord
...
4 DBWRKFLW:11079 4 0 123533.302911 pgSchedDbSetPhaseInfoFromPGresult: Query returned null value for column jobmanagerid
---
- nvsched will crash several times but eventually recovers and NetVault starts up normally.