And:gabrieles wrote: ↑Thu Aug 31, 2023 7:49 am Ok, if talking about it is somewhat permitted I would like to share this detail:
When we received the "delete these 3 files" workaround, we first checked if these file existed.
We found them existing on every production environment, but NOT on the test environments: on the test envs only the first existed (it's on the base package).
We tried to call the first of the files, the only one reachable under public, and we found that the request generates the other two files, the ones under jetty/work.
This explicit request leaves a trace on the nginx.access.log, easily greppable with
So we supposed that the existence of the second and third files is subordinate to a request of the first one.halfgaar wrote: ↑Wed Aug 30, 2023 9:30 pmCode: Select all
grep --fixed-strings --ignore-case hostedlogin /opt/zimbra/log/access_log*
But on the production environments, all the three files have been found, and no evidence on the logs of that request.
We store over a year of logs, and checked all of them.
Are we looking in the wrong place? Can someone point us in the right direction?
I just helped someone over DM and found that two of the three files had been regenerated for me too, in 'public_'. Timestamped 'aug 30 19:04', the same time I accessed the URLs (of non-existing files) with curl.ghen wrote: ↑Thu Aug 31, 2023 8:36 am That's strange, because all 3 files belong to the zimbra-mbox-webclient-war package... So you should have all 3 of them?
(rpm -ql zimbra-mbox-webclient-war | grep hostedlogin)
If you suspect two were generated at runtime, what is the age (mtime) of those files, according to ls -l ?