Rainer Gerhards

@ctramnitz as far as I could track, you are the original contributor of rsyslog's omhttp. Did you have a special reason to threat 4xx http errors as unrecoverable? Maybe you...

`rscript_get_property-vg.sh` is probably just an artifact of a change that I am currently implemeting and that does not go well at the moment ;)

@kortemik sry, I overlooked the comment. Will look into it, but it will take a short while.

@StrongestNumber9 I have amended the commit. IMHO removing the PID file at the point of failure looks like the best option. I would appreciate if you could give the patch...

doc in https://github.com/rsyslog/rsyslog-doc/pull/987

Without special config, the behavior is intended: we consider it preferable to fail one action instead of all processing. This is the original spirit of (r)syslogd for over 30 years....

@kortemik I have created https://github.com/rsyslog/rsyslog/pull/4904 CI is currently running, but manual tests look good. Doc is missing, but I will add either tomorrow or begin of week. I think you...

After careful consideration and as we had error message, I do not think this was a bug. So I removed the "bug" label und treat it as "enhancement". Not that...

As you know, we do only change defaults in very, very important cases. Even more so, I think most users will not expect rsyslog abort on a queue startup failure....

The most important reason for this probably is not setting a working directory and using disk or disk-assisted queue. In that case, we try to write to file system root...