21.7.3 Processes are killed due to swap space outage

[astrandb]

Upgraded  some 10 routers last night without reboot. After a few hours, four of the units failed with one or more services stopped.
Found following in the system/general log

Code Select Expand


2021-09-22T23:46:58   configctl[19087]   event @ 1632347211.39 msg:   
.
approx 7000 identical lines within 1 second
.
2021-09-22T23:46:58   configctl[19087]   event @ 1632347211.39 msg:   
2021-09-22T23:46:58   kernel   pid 34492 (unbound), jid 0, uid 59, was killed: out of swap space

Unbound was stopped in three units, unspecified python process in one and syslog-ng in one.
Just restarted the failed processes and everlything looks good after 1 hour.

[meyergru]

On my box, a configctl job was at 100% the whole time until reboot. Seems that a reboot is needed for this update.

[Benst]

I updated to 21.7.3 yesterday. Same here, suricata and unbound were killed with lots of out-of-swapspace messsages. I could manually restart them.

Will try the reboot now.

Ben