Hi JC,
Unfortunately, this didn't seem to fix the crashes. Today I got the
e-mail at 1:05 AM, though the core file has a timestamp of 1:04 AM.
This time frame still matches up with it being the dailyreport task
that is triggering the crashes (since there are no crashes any other
time of the day).
[root@<xymon server> ~]# xymoncmd --version
Xymon version 4.3.30-0.5.el7.terabithia
[root@<xymon server> ~]# cat /etc/redhat-release
Red Hat Enterprise Linux Server release 7.7 (Maipo)
The latest core backtrace looks to be the same as previously (same
client and service too), but I'm including it here [1] just for
completeness.
Let me know if there's anything else I can provide to debug this.
Thanks!
[1]
[root@<xymon server> ~]# gdb -q /usr/libexec/xymon/xymongen core.1312
Reading symbols from /usr/libexec/xymon/xymongen...Reading symbols
from /usr/lib/debug/usr/libexec/xymon/xymongen.debug...done.
done.
[New LWP 1312]
[Thread debugging using libthread_db enabled]
Using host libthread_db library "/lib64/libthread_db.so.1".
Core was generated by `/usr/libexec/xymon/xymongen
--reportopts=1566446400:1566532799:0:nongr --recent'.
Program terminated with signal 6, Aborted.
#0 ?0x00007fc746f7e377 in __GI_raise (sig=sig at entry=6) at
../nptl/sysdeps/unix/sysv/linux/raise.c:55
55 ?return INLINE_SYSCALL (tgkill, 3, pid, selftid, sig);
(gdb) bt
#0 ?0x00007fc746f7e377 in __GI_raise (sig=sig at entry=6) at
../nptl/sysdeps/unix/sysv/linux/raise.c:55
#1 ?0x00007fc746f7fa68 in __GI_abort () at abort.c:90
#2 ?0x00005559fc1ce4f5 in sigsegv_handler (signum=<optimized out>) at
sig.c:57
#3 ?<signal handler called>
#4 ?strchrnul () at ../sysdeps/x86_64/strchrnul.S:33
#5 ?0x00007fc746f90681 in __find_specmb (format=0xfce <Address 0xfce
out of bounds>) at printf-parse.h:109
#6 ?_IO_vfprintf_internal (s=s at entry=0x7ffeb3e6d340,
? ? format=format at entry=0xfce <Address 0xfce out of bounds>,
ap=ap at entry=0x7ffeb3e6d478) at vfprintf.c:1308
#7 ?0x00007fc74705dc78 in ___vsprintf_chk (s=0x7ffeb3e6d6c2 "",
flags=1, slen=18446744073709551615,
? ? format=0xfce <Address 0xfce out of bounds>,
args=args at entry=0x7ffeb3e6d478) at vsprintf_chk.c:83
#8 ?0x00007fc74705dbcd in ___sprintf_chk (s=<optimized out>,
flags=flags at entry=1,
? ? slen=slen at entry=18446744073709551615, format=<optimized out>) at
sprintf_chk.c:32
#9 ?0x00005559fc1bf96a in sprintf (__fmt=<optimized out>,
__s=<optimized out>)
? ? at /usr/include/bits/stdio2.h:33
#10 parse_histlogfile (starttime=1566446400,
? ? timespec=0x5559fc431f50 <timespec.7157> "Wed_Sep_2_19:34:55_2015",
servicename=0x5559fd61b2d0 "procs",
? ? hostname=0x5559fdc94520 "<client hostname>") at availability.c:174
#11 parse_historyfile (fd=fd at entry=0x5559fdc9be00, repinfo=<optimized
out>,
? ? hostname=0x5559fdc94520 "<client hostname>",
servicename=0x5559fd61b2d0 "procs",
? ? fromtime=<optimized out>, totime=1566532799,
for_history=for_history at entry=0, warnlevel=97,
? ? greenlevel=99.995000000000005, warnstops=-1, reporttime=0x0) at
availability.c:475
#12 0x00005559fc1b496c in init_state (filename=<optimized out>,
? ? filename at entry=0x7ffeb3e7f950 "<client hostname>.procs",
log=log at entry=0x7ffeb3e7f860)
? ? at loaddata.c:275
#13 0x00005559fc1b568e in load_state
(sumhead=sumhead at entry=0x5559fc3fad48 <dispsums>) at loaddata.c:626
#14 0x00005559fc1af794 in main (argc=5, argv=0x7ffeb3e84b58) at
xymongen.c:599
--
Matt Vander Werf
On Thu, Aug 22, 2019 at 5:29 PM Matt Vander Werf <user-dfc3cf2ca434@xymon.invalid
<mailto:user-dfc3cf2ca434@xymon.invalid>> wrote:
Hi JC,
Ah ha! That is one place I did not look and the timing certainly
matches up!
I have installed that new version on my Xymon server (running
actual RHEL 7) and we'll see how it fares tomorrow morning....
Thanks!
--
Matt Vander Werf
On Thu, Aug 22, 2019 at 5:12 PM Japheth Cleaver
<user-87556346d4af@xymon.invalid <mailto:user-87556346d4af@xymon.invalid>> wrote:
Hi,
I think this might be xymongen in report mode from the
"dailyreport" file in /tasks.d/; the timing would check out.?
I believe the problem here is one of the Terabithia patches
now doing the wrong thing after some of the string-handling
changes in 4.3.29 -- causing core dumps in certain situations.
If you're running actual RHEL7 on this (not CentOS, which
hasn't released 7.7 yet), would you mind checking the
xymon-4.3.30-0.5 package in the EL7 Terabithia testing repo
and see if this helps?
https://repo.terabithia.org/rpms/xymon/testing/el7/x86_64/
Regards,
-jc
On 8/22/2019 11:34 AM, Matt Vander Werf wrote: Hi Torsten,
No, there wasn't anything running from cron or anything else
around that time, let alone anything that restarts the
network or Xymon.
Thanks.
--
Matt Vander Werf
On Wed, Aug 21, 2019 at 5:43 AM Torsten Richter
<user-c862b499d9fa@xymon.invalid <mailto:user-c862b499d9fa@xymon.invalid>> wrote:
Hi Matt,
dumb question: is there any cron job running at this time
that is restarting XYmon fiddling with the network, like
restarting the network for some reason?
Regards,
Torsten
Matt Vander Werf <user-dfc3cf2ca434@xymon.invalid
<mailto:user-dfc3cf2ca434@xymon.invalid>> hat am 20. August 2019 um
17:10 geschrieben:
Hi all,
Every day since we updated our Xymon server to 4.3.29
(from 4.3.28), I've gotten an e-mail alert due to xymond
turning red that reads:
red xymongen program crashed
Fatal signal caught!
The strange thing is that this has happened at 1:04 AM
every day...like clockwork. I have xymongen set to run
every 1 minute and it has no problems running any other
time of the day. We are using the Terabithia RPMs and
the Xymon server is running RHEL 7.
I've scoured the system to find anything that is set to
run at/around that time via cron, etc. and haven't found
anything. The system logs don't show anything is
happening around that time either.
I turned on debug logging for xymond and xymongen and
haven't been able to find anything unusual in either
logs around that time. But it is dumping core files for
xymongen every time it crashes.
I used gdb to get the backtrace on all of the core files
(so far) and I've found that they all show the same
thing. It shows the same host in the backtrace too
(although I'm farily confident it isn't specific or
isolated to that host but just the first one it runs
into that it has issues with when processing).
I've included an example gdb output below (the most
recent one) [1].
Is anyone else running into this by chance? Or any idea
what might be the cause?
Thanks!
[1]
# gdb -q /usr/libexec/xymon/xymongen core.16327
Reading symbols from
/usr/libexec/xymon/xymongen...Reading symbols from
/usr/lib/debug/usr/libexec/xymon/xymongen.debug...done.
done.
[New LWP 16327]
[Thread debugging using libthread_db enabled]
Using host libthread_db library "/lib64/libthread_db.so.1".
Core was generated by `/usr/libexec/xymon/xymongen
--reportopts=1566187200:1566273599:0:nongr --recent'.
Program terminated with signal 6, Aborted.
#0 ?0x00007f4657c49377 in __GI_raise (sig=sig at entry=6)
at ../nptl/sysdeps/unix/sysv/linux/raise.c:55
55 ?return INLINE_SYSCALL (tgkill, 3, pid, selftid, sig);
(gdb) bt
#0 ?0x00007f4657c49377 in __GI_raise (sig=sig at entry=6)
at ../nptl/sysdeps/unix/sysv/linux/raise.c:55
#1 ?0x00007f4657c4aa68 in __GI_abort () at abort.c:90
#2 ?0x00005589375dd455 in sigsegv_handler
(signum=<optimized out>) at sig.c:57
#3 ?<signal handler called>
#4 ?strchrnul () at ../sysdeps/x86_64/strchrnul.S:33
#5 ?0x00007f4657c5b681 in __find_specmb (format=0xfce
<Address 0xfce out of bounds>) at printf-parse.h:109
#6 ?_IO_vfprintf_internal (s=s at entry=0x7ffd5dabcc00,
? ? format=format at entry=0xfce <Address 0xfce out of
bounds>, ap=ap at entry=0x7ffd5dabcd38) at vfprintf.c:1308
#7 ?0x00007f4657d28c78 in ___vsprintf_chk
(s=0x7ffd5dabcf82 "", flags=1, slen=18446744073709551615,
? ? format=0xfce <Address 0xfce out of bounds>,
args=args at entry=0x7ffd5dabcd38) at vsprintf_chk.c:83
#8 ?0x00007f4657d28bcd in ___sprintf_chk (s=<optimized
out>, flags=flags at entry=1,
slen=slen at entry=18446744073709551615, format=<optimized
out>) at sprintf_chk.c:32
#9 ?0x00005589375ce8ca in sprintf (__fmt=<optimized
out>, __s=<optimized out>)
? ? at /usr/include/bits/stdio2.h:33
#10 parse_histlogfile (starttime=1566187200,
? ? timespec=0x558937840f50 <timespec.7157>
"Wed_Sep_2_19:34:55_2015", servicename=0x5589383b6d70
"procs",
? ? hostname=0x558938a335d0 "<client hostname>") at
availability.c:174
#11 parse_historyfile (fd=fd at entry=0x558938a3aea0,
repinfo=<optimized out>,
? ? hostname=0x558938a335d0 "<client hostname>",
servicename=0x5589383b6d70 "procs",
? ? fromtime=<optimized out>, totime=1566273599,
for_history=for_history at entry=0, warnlevel=97,
? ? greenlevel=99.995000000000005, warnstops=-1,
reporttime=0x0) at availability.c:475
#12 0x00005589375c38cc in init_state
(filename=<optimized out>,
? ? filename at entry=0x7ffd5dacf210 "<client
hostname>.procs", log=log at entry=0x7ffd5dacf120)
? ? at loaddata.c:275
#13 0x00005589375c45ee in load_state
(sumhead=sumhead at entry=0x558937809d48 <dispsums>) at
loaddata.c:626
#14 0x00005589375be6f4 in main (argc=5,
argv=0x7ffd5dad4418) at xymongen.c:599
--
Matt Vander Werf