PAB webserver process was killed due to OOM, event though PC has 32Gb memory

[code-brewer]

Hi ,
I run plutus demo on my AWS server with 8core, 32Gb.

After start process with follow command:
cabal exec -- plutus-starter-pab --config plutus-pab.yaml --passphrase "Secure Passphrase" webserver
then it was observed that this process use 100% CPU and 21% percent mempry.

It run serval hours while the CPU and memory usage have little fluctuation.
But after I leave and get back some hour later, this process was killed.

After investigate into it, from the dmesg command I get following message:

[Thu May 12 06:51:23 2022] oom-kill:constraint=CONSTRAINT_NONE,nodemask=(null),cpuset=/,mems_allowed=0,global_oom,task_memcg=/user.slice,task=plutus-starter-,pid=7209,uid=1000 [Thu May 12 06:51:23 2022] Out of memory: Killed process 7209 (plutus-starter-) total-vm:1074491512kB, anon-rss:27228080kB, file-rss:0kB, shmem-rss:0kB, UID:1000 pgtables:53528kB oom_score_adj:0 [Thu May 12 06:51:24 2022] oom_reaper: reaped process 7209 (plutus-starter-), now anon-rss:0kB, file-rss:0kB, shmem-rss:0kB

It seem that PAB webserver was killed due to OOM, but I not watch the on-the-spot memory.

Just wonder is there minimum memory requirement? Or there is memory leak for PAB webserver ?

Thanks for any feedbacks.

Thanks for the issue!

Yes, there is a problem with PAB at the moment IntersectMBO/plutus-apps#420.

You may try to change the rollback history parameter which should help a bit before a proper fix.