Hi,

 

Just after upgrading from 4.7 to 5.0RC10 I noticed that I was getting these crond errors.  Any idea what they relate too?  They are every hour which is when crond is set to run.  47mins after the hour.

 

Jan 17 21:47:01 Tower kernel: crond[20340]: segfault at b77b051c ip b77b051c sp bfea4df4 error 15 in ld-2.11.1.so[b77b0000+1000] (Errors)

 

I have Plex installed as the only plug in.  I've uninstalled simple features & drive mount (removed the plg files & the folder created by them), and run unmenu, and SAB, SB and UMS off my apps drive as I always did under 4.7.  And I have rebooted, and it still remains.

 

Any ideas?

 

Syslog attached.

syslog-2013-01-17.txt

What happens when you run the command that is failing yourself:

/usr/bin/run-parts /etc/cron.daily

or

/usr/bin/run-parts /etc/cron.hourly

Thanks Joe for the help.

 

root@Tower:~# /usr/bin/run-parts /etc/cron.daily

/etc/cron.daily/certwatch:

 

 

/etc/cron.daily/logrotate:

 

 

and when I ran hourly the drives spun up.

root@Tower:~# /usr/bin/run-parts /etc/cron.hourly

/etc/cron.hourly/smtp_status.sh:

 

 

When I ran these in Putty no errors where generated in the Syslog, it only showed me logging in.

also I rebooted again today, and they are still appearing.

Any ideas if smtp_status.sh is causing the issue?  It's the only think under cron hourly.  I do have smtp set up in unmenu, or is this left over from simple features?

Looks like I'm on my own with this one, I'll just document my progress to so that if anyone can help, they can see what I've already done.

 

I've disabled plex and it still occurred, so that's makes the install exactly the same as my previous 4.7 install.  Possibly something is different between 4.7 and RC10.

 

So I went back to stock RC10, I removed everything from my go script apart from the default lines and the error didn't appear at 47mins past, so assuming it's not an unraid issue.

 

Next I reactivated my apps drivein the go script, no crond error

 

Then I reactivated unmenu in the go script, no crond error

 

Then I reactivated UMS.sh (Universal Media Server) and the crond error appeared!

 

Now I just have to nail down what is causing the crond issue.  Might be time to try something else due to the number of packages UMS requires on installation.  I've found plex to be a little unreliable with my WDTV live, but it's fantastic with my LG TV.  I might see if the WDTV firmware upgrade makes it any better.

 

Just realised that UMS.sh fails to start on a reboot, so it must be one of the 53 packages that it requires that cause the crond fault!   Likelyhood of me being able to find out what one it is will be rather slim.

 

Linux support isn't up there with with windoz for PMS/UMS, so I'll try a few other DLNA servers I think if plex doesn't work well.

 

At least the crond fault has been stopped, and I can now get my daily reports from the server.

All of the segfaults relate to ld, so while that does not necessarily mean that it's ld's fault, it does make it a good starting point for investigation.

 

Most of the segfaults are related to a cron call that's hourly (47th minute), as you noticed.  But they also occur with a daily cron call at 4:40am.  Perhaps that could help narrow it down?

Thanks RobJ,

 

Well I had Universal Media Server (an offshoot of PS3MS) running nicely on 4.7, and I upgraded to RC10 really so that I could run Plex with my new LG TV.

 

It's a blessing in disguise really, because I've now disabled UMS, and I'll just run Plex (which also does DLNA).  I had previously had issues with my WDTV but I upgraded the firmware, and watched a show on it while someone else was watching on the LG in the other room, and it handled it well.

 

UMS was a bit messy to install, lots of dependencies to get it to run right, including running Java, and 53 packages.  It could possibly be that some packages included ran fine in 4.7 and not in 5RC10.

 

Thanks everyone for the help, I'm happy to say this issue is solved for now.