-
Change Request
-
Resolution: Done
-
Trivial
-
None
-
4.0.41
-
* CentOS Linux release 7.9.2009 (Core
* HyperV hypervisor
* SSD Disk
* Postgresql 9
After a standard OS security update the housekeeper process suddenly get's SIGSEGVed, see zabbix-housekeeper-sigsegv.log
Postgresql itself only sees a vanishing connection: zabbix-postgresql.log
It seems to happen right after the select itemid,min(clock) from history_uint group by itemid query finishes.
I attached via strace. But the log is not very helpfull: zabbix-housekeeper-strace.log
So far, a package downgrade, OS restart etc. has not helped. Befor the OS level updates postgresql was presenting an error to the housekeeper that did not seem to create any SIGSEGV:
2022-06-28 07:12:10 CEST STATEMENT: delete from history_uint where itemid=74542 and clock<1655788142
2022-06-28 07:12:10 CEST ERROR: could not access status of transaction 138739714
2022-06-28 07:12:10 CEST DETAIL: Could not open file "pg_subtrans/0845": No such file or directory.
This postgresql error is no longer present after OS level updates and a reboot. But now housekeeper no longer works.