[Sat Dec 17 16:23:13.476229 2022] [slotmem_shm:error] [pid 1634859:tid 140714406766464] (17)File exists: AH02611: create: apr_shm_create(/var/log/apache/s1-eu.ariba.com/shm/slotmem-shm-p15744fb4_s4taskcxml_c12_0.shm) failed [Sat Dec 17 16:23:13.476283 2022] [:emerg] [pid 1634859:tid 140714406766464] AH00020: Configuration Failed, exiting hosts running with RH8 OS are not restarting with the above issue.
I think it's a duplicate of https://bz.apache.org/bugzilla/show_bug.cgi?id=65180 I added some info in that ticket as it contains more detailed debugging info.
tried various code fixes, but seemed unable to completely solve the issue. There is some speculation that the underlying kernel code that Apache relies on to clear shared memory may “fail silently”. So, we are possibly hitting a known bug in the Apache code or kernel bug, that only happens when there are a large number of shared memory segments that are deleted and re-created very quickly. This is true even in the more recent Apache release we are using: 2.4.54