When polling results are null, segmentation errors may occur

[a-vasilishin]

Today I compiled newest spine https://www.cacti.net/downloads/spine/cacti-spine-1.2.10.tar.gz
tar zxvf cacti-spine-1.2.10.tar.gz
cd cacti-spine-1.2.10
./bootstrap
./configure
make
make install
cp /usr/local/spine/bin/spine /usr/sbin/spine
Now I have many entries in log:
2020/03/26 20:40:31 - SPINE: Poller[Main Poller] PID[13424] FATAL: Spine Encountered a Segmentation Fault [4, Interrupted system call] (Spine thread)

I am using
mysql Ver 15.1 Distrib 10.1.44-MariaDB, for debian-linux-gnu (x86_64) using readline 5.2
Debian 3.16.7-ckt25-2 (2016-04-08) x86_64 GNU/Linux

[TheWitness]

Grab the 1.2.x branch and try again. It'll be released soon.

[TheWitness]

My guess is issue this is a duplicate of this one #149

[a-vasilishin]

I tried it. :(
-rwsr-sr-x 1 root root 429128 Mar 27 01:14 /usr/sbin/spine

2020/03/27 01:20:32 - SPINE: Poller[Main Poller] PID[9999] FATAL: Spine Encountered a Segmentation Fault [4, Interrupted system call] (Spine thread)

2020/03/27 01:25:32 - SPINE: Poller[Main Poller] PID[15310] FATAL: Spine Encountered a Segmentation Fault [4, Interrupted system call] (Spine thread)

2020/03/27 01:30:32 - SPINE: Poller[Main Poller] PID[19006] FATAL: Spine Encountered a Segmentation Fault [4, Interrupted system call] (Spine thread)

Every 5 minute, second per second

[TheWitness]

Okay, please run from the command line and do a backtrace

gdb ./spine
run -V 3 -S -R
bt

Post the results of the backtrace after the segfault.

[a-vasilishin]

POLLER: Active Threads is 2, Pending is 2
[Thread 0x7f2df8ff9700 (LWP 8235) exited]
Device[89] HT[1] DQ[1] RECACHE OID: .1.3.6.1.2.1.2.2.1.2.1, (assert: lo = output: U)
Device[89] HT[1] DQ[1] RECACHE OID: .1.3.6.1.2.1.2.2.1.2.2, (assert: eth0 = output: (null))

Thread 4 "spine" received signal SIGSEGV, Segmentation fault.
[Switching to Thread 0x7f2dfbfff700 (LWP 7901)]
poll_host (host_id=host_id@entry=89, host_thread=host_thread@entry=1, last_host_thread=last_host_thread@entry=1, host_data_ids=host_data_ids@entry=0,
host_time=host_time@entry=0x7f2dfbffede0 "2020-03-27 11:59:39", host_errors=host_errors@entry=0x7f2dfbffedd8, host_time_double=host_time_double@entry=1585303179.0394671) at poller.c:836
836 if ((IS_UNDEFINED(poll_result)) || (STRIMATCH(poll_result, "No Such Instance"))) {
(gdb) bt
#0 poll_host (host_id=host_id@entry=89, host_thread=host_thread@entry=1, last_host_thread=last_host_thread@entry=1, host_data_ids=host_data_ids@entry=0,
host_time=host_time@entry=0x7f2dfbffede0 "2020-03-27 11:59:39", host_errors=host_errors@entry=0x7f2dfbffedd8, host_time_double=host_time_double@entry=1585303179.0394671) at poller.c:836
#1 0x00007f2e030c0065 in child (arg=) at poller.c:77
#2 0x00007f2e018c54a4 in start_thread (arg=0x7f2dfbfff700) at pthread_create.c:456
#3 0x00007f2e0139ed0f in clone () at ../sysdeps/unix/sysv/linux/x86_64/clone.S:97
(gdb)

[TheWitness]

Okay, update poller.c from the 1.2.x branch, recompile and try again. This is a corner case, but a case just the same.

[a-vasilishin]

Thanks! Now almost fine in log

[TheWitness]

There is a recent fix for unresolvable hostname. You might want to look at #155. Closing this one.