https://bugzilla.novell.com/show_bug.cgi?id=728565
https://bugzilla.novell.com/show_bug.cgi?id=728565#c49
--- Comment #49 from Marius Tomaschewski
I don't think that will work. The admin may provide /other files/ in /etc/rsyslog.d that require networky bits.
The admin that prefers other files should be able to set: SYSLOG_REQUIRES_NETWORK=yes
Why does rsyslog-early.conf even exist?
Artifact from attic versions that were not smart enough to catch some failures.
Won't rsyslog retry forever if the network is down?
Yes, it will retry from time to time (every 30sec?).
What's the problem with *always* using rsyslog.conf ?
There is _no_ problem. The rsyslog.early.conf _is_ obsolete in the packages above (comment 43): http://download.opensuse.org/repositories/home:/mtomaschewski:/branches:/ope... The packages _are_ using rsyslog.conf only (+ its includes). The early config is still shipped, but only to catch the systemd generator problem. A syslog restart after update would fail when I would drop it completely now [-> 12.2]. When on-disk queues are enabled, rsyslog will not loose any messages. When on-disk queues are not enabled, it will log into local files and start to log remotely as soon as it reaches the log host. But using any kind of remote or db logging without on-disk queues is simply a admin configuration bug.
I think the default should be: assume network needed - it's likely to be the far more common case that machines will have networking than not - why _default_ to the far-less-likely case? Perhaps if somebody really doesn't want rsyslog to require networking, then /etc/sysconfig/syslog is an appropriate place to set that -- but I'd suggest that the default should be that it requires networking (at least to start) and not the other way around.
No, I do not agree. I'd say >50% of the users are desktop/notebook users that do not log remotely. It makes sense to start the syslog daemon as soon as possible to write the messages down before they get dropped because the buffers aren't endless. Note, that remote logging can't be even a default, because it _has_ to be configured: you have to provide the name / IP of the log host. (In reply to comment #48)
Default configuration do not specify remote logging so it will work as expected. I think a sysadmin will only enable remote logging if he knows he has network access!
It will work as expected also _without_ any network dependency. With on-disk queues, you don't need network to start rsyslog. It will write the messages into the on-disk queues until it is able to send them to the log host.
I would suggest that the only patch needed is a comment in remote.conf with a recommendation that the spool file should also be enabled, when remote logging is setup.
The recommendation was always there: .. # Remote Logging (we use TCP for reliable delivery) # An on-disk queue is created for this action. If the remote host is # down, messages are spooled to disk and sent when it is up again. #$WorkDirectory RSYSLOG_SPOOL_DIR # where to place spool files #$ActionQueueFileName uniqName # unique name prefix for spool files #$ActionQueueMaxDiskSpace 1g # 1gb space limit (use as much as possible) #$ActionQueueSaveOnShutdown on # save messages to disk on shutdown #$ActionQueueType LinkedList # run asynchronously #$ActionResumeRetryCount -1 # infinite retries if host is down # remote host is: name/ip:port, e.g. 192.168.0.1:514, port optional #*.* @@remote-host .. The current packages already make it even clearer: ## ## === When you're using remote logging, enable on-disk queues === ## === in rsyslog.d/remote.conf. When neccesary also set the === ## === SYSLOG_REQUIRES_NETWORK=yes in /etc/sysconfig/syslog. === ## .. # ######### Enable On-Disk queues for remote logging ########## # # An on-disk queue is created for this action. If the remote host is # down, messages are spooled to disk and sent when it is up again. # Note again: SYSLOG_REQUIRES_NETWORK=yes in /etc/sysconfig/syslog is not required, except in special cases, e.g. you want to bind to listen on a specific IP (see comment 40), that is when you are on the _log_ _host_ and you configure it like this: # UDP Syslog Server: $ModLoad imudp.so # provides UDP syslog reception $UDPServerAddress 172.16.20.1 #^^^^^^^^^^^^^^^^^^^^^^^^^^^^ $UDPServerRun 514 # start a UDP syslog server at standard port 514 -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.