Bug 218552

Please install the corresponding cups-debuginfo package and then, after
restarting cups (so that it is expected to exhibit this bug), find out the
process ID of cupsd and attach gdb to it like this:

gdb /usr/sbin/cupsd 4295

(if 4295 is the process ID of cupsd).

Then, at the (gdb) prompt, enter:

  continue

The next (gdb) prompt should be from the crash -- at that point, enter 'bt' and
copy the backtrace you get into this bug report.  If possible, don't exit the
session and I will try to get back to you in order to ask for any further
information I need.  Thanks.

I am looking at it. I installed the cups-debuginfo and attached gdb to it, and
the problem stopped happening, but I am not seeing the "Remote Destination" line
appearing anymore either. I am trying to see if I can casue the "Remote
destination" and see if it triggers cups from dieing. Will let you know.

Sorry for the long delay. It would happen when I didnt have gdb running.. and
when I had gdb running it wouldnt happen for a week or more. I finally got it
today with gdb

Long time tracking it down. The only settings away from the default cups conf
are adding two lines:

BrowsePoll print.unm.edu:631
BrowsePoll print2.unm.edu:631

cups-1.2.7-1.8.fc6
cups-debuginfo-1.2.7-1.8.fc6
cups-libs-1.2.7-1.8.fc6
cups-pdf-2.4.4-1.fc6


Program received signal SIGSEGV, Segmentation fault.
[Switching to Thread -1208961328 (LWP 19605)]
delete_rules (rules=0x81006e6f) at mime.c:452
452         if (rules->child != NULL)
(gdb) where
#0  delete_rules (rules=0x81006e6f) at mime.c:452
#1  0x80048624 in delete_rules (rules=0x81bac768) at mime.c:453
#2  0x800493a6 in mimeDeleteType (mime=0x81b327c0, mt=0x81bebe48) at mime.c:169
#3  0x80042cdf in cupsdDeletePrinter (p=0x81bb3ba8, update=1) at printers.c:743
#4  0x8001dbac in cupsdSendBrowseList () at dirsvc.c:839
#5  0x80020931 in main (argc=1, argv=0xbfae2a34) at main.c:935
(gdb) t
[Current thread is 1 (Thread -1208961328 (LWP 19605))]

Let me know what other commands you want me to try.

(gdb) frame
#0  delete_rules (rules=0x81006e6f) at mime.c:452
452         if (rules->child != NULL)
(gdb) up
#1  0x80048624 in delete_rules (rules=0x81bac768) at mime.c:453
453           delete_rules(rules->child);
(gdb) up 
#2  0x800493a6 in mimeDeleteType (mime=0x81b327c0, mt=0x81bebe48) at mime.c:169
169       delete_rules(mt->rules);
(gdb) up
#3  0x80042cdf in cupsdDeletePrinter (p=0x81bb3ba8, update=1) at printers.c:743
743       mimeDeleteType(MimeDatabase, p->filetype);
(gdb) up
#4  0x8001dbac in cupsdSendBrowseList () at dirsvc.c:839
839             cupsdDeletePrinter(p, 1);
(gdb) up
#5  0x80020931 in main (argc=1, argv=0xbfae2a34) at main.c:935
935           cupsdSendBrowseList();
(gdb) up
Initial frame selected; you cannot go up.

There is nothing in the logs that I can tell occuring on the client. I am
looking at the upstream cups printers to see if they are triggering something.
Any ideas would be welcome. I will keep the gdb open for the time being in case
there are stack items you want me to give you.

Try this:

select-frame 3
print p->name
print p->uri

(gdb) select-frame 3
(gdb) print p->name
$1 = 0x81bebdd0 "vaad5"
(gdb) print p->uri
$2 = 0x81b50e90 "ipp://print2.unm.edu:631/printers/vaad5"


I will look on the second print server to see what was going on with this one
last night (if anything.)

Ok vaad5 is the last printer on the list. We have 78 printers defined on print
and print2.. I do not see any errors listed for vaad5 on the print server which
is currently running ESP-Pro's version.

vaad5 was being removed from the printer list because too much time had elapsed
since it was last advertised by the server (this is the BrowseTimeout).  Somehow
while we were cleaning up the MIME database to remove printer/vaad5, we followed
a stale pointer.

Not sure at this point what the root cause is.

1. Could you attach your /etc/cups/mime.types file please?

2. It would also be helpful if I could see the debugging output when the printer
is first seen.  Here's how to get that (as root):

Stop CUPS:

/sbin/service cups stop

Clear out the error_log:

>/var/log/cups/error_log

Edit /etc/cups/cupsd.conf and change the 'LogLevel' line to read:

LogLevel debug2

Remove the remote printers cache:

rm -f /var/cache/cups/remote.cache

Start CUPS again:

/sbin/service cups start

Now wait for a few minutes and see if 'vaad5' is available to print to.  Once it
is, please attach the /var/log/cups/error_log file to this bug report.

Thanks!

Created attachment 149977 [details]
Default mime-types no change from rpm -Vf /etc/cups/mime.types

This is the requested mime.types. Looks to be default file.

Created attachment 149980 [details]
Requested error log

This is the error log from debug2. One interesting thing is that when I put the
system into debug2.. I think the issue may be with one of the polled servers
timing out. We have two polls (print.unm.edu and print2.unm.edu). Should I look
at the cups-polld next time also?

(Fixing bug status)

Did you need more info? By the way... the poll patch seems to have shown up in
the RHEL-3.9 cups and is causing problems on a lot of boxes with /var/log/cups
filling up.

Only today spotted this had the information added; haven't checked through it yet.

What poll patch?  Does that bug have its own bugzilla report?  If not, would you
mind filing one?  Thanks.

It has a RH Trouble ticket.. but I am guessing we are one of the few places
using browsepolling :).

Bugzilla Bug 246545: Cups fills up logfiles if queue is turned on

As far as I can tell.. this is fixed except for the RHEL-3 bug. Should you or I
combine the two now

I'll close this one and try to get the RHEL-3 one addressed.