On Wed, 2008-11-26 at 13:03 -0500, Chuck Anderson wrote:
> On Wed, Nov 26, 2008 at 12:00:59PM -0600, Mike Chambers wrote:
> > When using NetworkManager, none of my log files (such as from epylog,
> > logwatch, crons, etc..) get emailed to my server (local network here at
> > home), all using static network. When I go to the *old* network, they
> > are sent just fine.
>
> Probably related to NetworkManager not changing the system hostname
> and leaving it at the default localhost.localdomain.
NM will change the hostname just like dhclient-script does, if you have
not set a persistent hostname in /etc/sysconfig/network. If you have
set a persistent hostname in /etc/sysconfig/network, NM will *always*
use that hostname.
It may be the issue where sendmail is happily unaware of network changes
including hostname changes. If, after connecting, you do '/sbin/service
sendmail restart' does that start things working? The real fix here is
to hit stupid services that don't handle network changes in the face
with a nail-studded board, and repeat if they get back up, until they
stop getting up and just lie there. But failing that, forcing them to
restart via a dispatcher script when network changes occur may be the
quickest alternative.
Dan
--
fedora-devel-list mailing list
fedora-devel-list@redhat.com
https://www.redhat.com/mailman/listinfo/fedora-devel-list
11-26-2008, 05:40 PM
Mike Chambers
NetworkManager doesn't email log files
On Wed, 2008-11-26 at 13:31 -0500, Dan Williams wrote:
> O
> NM will change the hostname just like dhclient-script does, if you have
> not set a persistent hostname in /etc/sysconfig/network. If you have
> set a persistent hostname in /etc/sysconfig/network, NM will *always*
> use that hostname.
The hostname was there and listed the whole time, from install (I used
askmethod during install so I could get NM to use static, as GUI didn't
allow me to).
>
> It may be the issue where sendmail is happily unaware of network changes
> including hostname changes. If, after connecting, you do '/sbin/service
> sendmail restart' does that start things working? The real fix here is
> to hit stupid services that don't handle network changes in the face
> with a nail-studded board, and repeat if they get back up, until they
> stop getting up and just lie there. But failing that, forcing them to
> restart via a dispatcher script when network changes occur may be the
> quickest alternative.
No, sendmail wouldn't still send it. I did go to network service and
relooked at files (ifcfg-eth0 mainly), saw any differences with
NetworkManger and with network. Went back to NM just now and ran epylog
cron manually and it got sent. Sooo, it *might* be working now. So
maybe something in ifcfg helped and allows it to go now?