Ad Widget

**jsaintro** · 01-03-2012, 19:22

I'm having the exact same problem. Here is the output from the zabbix_server.log with DebugLevel=4

Code:

  2618:20120301:120125.138 In process_ping() hosts_count:1
  2618:20120301:120125.138 /tmp/zabbix_server_pgsql_2618.pinger
  2618:20120301:120125.138 192.168.96.201
  2618:20120301:120125.138 /usr/sbin/fping -q -C3 2>&1 </tmp/zabbix_server_pgsql_2618.pinger;/usr/sbin/fping -q -C3 2>&1 </tmp/zabbix_server_pgsql_2618.pinger;
  2618:20120301:120125.143 End of process_ping()
  2618:20120301:120125.143 End of do_ping():SUCCEED
  2618:20120301:120125.143 In process_values()
  2618:20120301:120125.143 Host [192.168.96.201] cnt=0 rcv=0 min/max/avg=0.000000/0.000000/0.000000
  2618:20120301:120125.143 In process_value()

And here is a test as the zabbix user from the command line:

Code:

-bash-4.1$ whoami
zabbix
-bash-4.1$ echo "192.168.96.201" > /tmp/hostname
-bash-4.1$  /usr/sbin/fping -q -C3 2>&1 </tmp/hostname;/usr/sbin/fping -q -C3 2>&1 </tmp/hostname;
192.168.96.201 : 0.79 0.16 0.16
192.168.96.201 : 0.18 0.16 0.11

I'm guessing the problem is in the parsing of the fping output as I'm getting results when I do it manually but the debug output shows zero values. This is the same when I use icmppingsec.

I'm using the following environment:

OS: Centos 6.2 (Final) 32 Bit
fping: Version 2.4b2_to $Date: 2002/01/16 00:33:42 $
Zabbix: Version 1.8.10
Postgres: Version 8.4.9

**greenlight** · 01-03-2012, 19:29

the versions for my set up are
centos 6.2
zabbix 1.8.10
fping 2.4b2
mysql-5.1.61

basically the latest released centos versions. But on one system its working ant the other not.

I've not tries debug level 4 yet

**greenlight** · 02-03-2012, 12:25

this is definitely something to do with the parsing of the fping output into zabbix.

on the working system the output of fping is the same as the non working system

not working:

[root@monitor /]# echo 212.20.226.130 > /tmp/host
[root@monitor /]# /usr/sbin/fping -q -C3 2>&1 </tmp/host
212.20.226.130 : 13.66 13.21 12.94

[root@monitor richardr]# echo 212.20.226.130 > /tmp/host
[root@monitor richardr]# /usr/sbin/fping -q -C3 2>&1 </tmp/host
212.20.226.130 : 10.27 10.44 10.30

This rules out fping as the culprit

**jsaintro** · 02-03-2012, 21:09

Confirmed the correct behavior using the Zabbix Virtual Appliance. Exact same config but values are correctly parsed. So it's something specific to CentOS and the Zabbix install would be my guess.

I also checked the version of zabbix between Appliance and CentOS:

Code:

[me@centos ~]$ /usr/sbin/zabbix_server_pgsql -V
Zabbix Server v1.8.10 (revision 24303) (28 December 2011)
Compilation time: Dec 29 2011 12:14:26

Code:

[root@appliance ~]$ /usr/sbin/zabbix_server -V
Zabbix Server v1.8.10 (revision 24303) (28 December 2011)
Compilation time: Dec 28 2011 11:48:13

**greenlight** · 03-03-2012, 19:12

I have two almosty identical centos installs one works and one doesn't. Can't see why one would work and one not. its defo to do with parsing the data

**MrKen** · 04-03-2012, 12:34

Check this post (in Russian).
Same problem. Issue resolved.

http://translate.google.com/translat...26prmd%3Dimvns

MrKen

**greenlight** · 05-03-2012, 13:17

all permissions for fping are correct.

[root@monitor sbin]# su zabbix
sh-4.1$ fping 212.20.226.130
212.20.226.130 is alive
sh-4.1$ fping -q -C3 212.20.226.130
212.20.226.130 : 13.33 13.07 12.87
sh-4.1$ ls -la fping*
-rws--s---. 1 root zabbix 27824 May 19 2010 fping
-rws--s---. 1 root zabbix 27856 May 19 2010 fping6
sh-4.1$

logs show no data is being passed to zabbix from the fping

1755:20120305:110854.112 In get_pinger_hosts()
1755:20120305:110854.112 In DCconfig_get_poller_items() poller_type:3
1755:20120305:110854.112 End of DCconfig_get_poller_items():1
1755:20120305:110854.112 In substitute_simple_macros() data:'icmppingsec'
1755:20120305:110854.112 In add_icmpping_item() addr:'monitor.rikrobson.co.uk' count:3 interval:0 size:0 timeout:0
1755:20120305:110854.112 End of add_icmpping_item()
1755:20120305:110854.112 End of get_pinger_hosts():1
1755:20120305:110854.112 In process_pinger_hosts()
1755:20120305:110854.112 In add_pinger_host() addr:'monitor.rikrobson.co.uk'
1755:20120305:110854.112 End of add_pinger_host()
1755:20120305:110854.112 In do_ping() hosts_count:1
1755:20120305:110854.112 In process_ping() hosts_count:1
1755:20120305:110854.113 /tmp/zabbix_server_mysql_1755.pinger
1755:20120305:110854.113 monitor.rikrobson.co.uk
1755:20120305:110854.113 /usr/sbin/fping -q -C3 2>&1 </tmp/zabbix_server_mysql_1755.pinger;
1755:20120305:110854.118 End of process_ping()
1755:20120305:110854.118 End of do_ping():SUCCEED
1755:20120305:110854.118 In process_values()
1755:20120305:110854.118 Host [monitor.rikrobson.co.uk] cnt=0 rcv=0 min/max/avg=0.000000/0.000000/0.000000
1755:20120305:110854.118 In process_value()

on an almost identical system that works:

[root@monitor zabbix]# grep 031010 zabbix_server.log
32025:20120302:031010.219 read line [212.20.226.130 : 10.74 12.96 11.50]
32025:20120302:031010.223 read line [fping6: Address family for hostname not supported]
32025:20120302:031010.227 End of process_ping()
32025:20120302:031010.227 End of do_ping():SUCCEED
32025:20120302:031010.227 In process_values()
32025:20120302:031010.227 Host [212.20.226.130] cnt=3 rcv=3 min/max/avg=0.010740/0.012960/0.011733
32025:20120302:031010.227 In process_value()
32025:20120302:031010.227 In calculate_item_nextcheck() itemid:23257 delay:30 flex_intervals:'' now:1330657807
32025:20120302:031010.227 End of calculate_item_nextcheck() nextcheck:1330657837 delay:30

[root@monitor zabbix]# ls -la /usr/sbin/fping
-rws--x--- 1 root zabbix 27824 May 19 2010 /usr/sbin/fping
[root@monitor zabbix]#

**greenlight** · 05-03-2012, 19:07

fresh install of centos 6 and all the latest versions

fping works as the zabbix user:

[root@demeter ~]# su zabbix -c '/usr/sbin/fping -q -C3 2>&1 </tmp/host'
212.20.226.130 : 13.11 13.00 13.11
46.105.118.46 : 0.46 0.58 0.46

the logs appear to show that the host file is being written and the fping completeing correctly. there is no data parsed back to the server. I am assuming that the host file is being written correctly and can is being read before deletion. not sure how i can fine this out as the file is not in existance for very long

these are the logs:

19132:20120305:164852.659 In get_pinger_hosts()
19132:20120305:164852.659 In DCconfig_get_poller_items() poller_type:3
19132:20120305:164852.660 End of DCconfig_get_poller_items():1
19132:20120305:164852.660 In substitute_simple_macros() data:'icmppingsec'
19132:20120305:164852.660 In add_icmpping_item() addr:'78.129.180.115' count:3 interval:0 size:0 timeout:0
19132:20120305:164852.660 End of add_icmpping_item()
19132:20120305:164852.660 End of get_pinger_hosts():1
19132:20120305:164852.660 In process_pinger_hosts()
19132:20120305:164852.660 In add_pinger_host() addr:'78.129.180.115'
19132:20120305:164852.660 End of add_pinger_host()
19132:20120305:164852.660 In do_ping() hosts_count:1
19132:20120305:164852.660 In process_ping() hosts_count:1
19132:20120305:164852.660 /tmp/zabbix_server_mysql_19132.pinger
19132:20120305:164852.660 78.129.180.115
19132:20120305:164852.660 /usr/sbin/fping -q -C3 2>&1 </tmp/zabbix_server_mysql_19132.pinger;/usr/sbin/fping6 -q -C3 2>&1 </tmp/zabbix_server_mysql_19132.pinger;
19161:20120305:164852.662 In collect_selfmon_stats()
19161:20120305:164852.662 End of collect_selfmon_stats()
19161:20120305:164852.662 sleeping for 1 seconds
19132:20120305:164852.666 End of process_ping()
19132:20120305:164852.666 End of do_ping():SUCCEED
19132:20120305:164852.666 In process_values()
19132:20120305:164852.666 Host [78.129.180.115] cnt=0 rcv=0 min/max/avg=0.000000/0.000000/0.000000
19132:20120305:164852.666 In process_value()
19132:20120305:164852.666 In calculate_item_nextcheck() itemid:22162 delay:30 flex_intervals:'' now:1330966132
19132:20120305:164852.666 End of calculate_item_nextcheck() nextcheck:1330966162 delay:30
19132:20120305:164852.666 End of process_value()
19132:20120305:164852.666 End of process_values()
19132:20120305:164852.666 End of process_pinger_hosts()
19132:20120305:164852.666 icmp pinger #1 spent 0.006978 seconds while processing 1 items

**zabbixhack** · 12-04-2012, 00:35

pinger files in /tmp never created

I have 3 proxies - all 1.8.11 and all RHEL 6.2 x86_64.
All same config files, except host info.

1 has icmpping working - the other 2 are not.

I've been through all the threads and I have not found anyone with the solution. Fping permissions, location, ipv6 off, etc all checked.
The item config is in a template on the master - used by all 3 proxies.
1 works.

The only thing I see besides tcpdump not showing even an icmp attempt is:

The /tmp/zabbix_proxy_mysql_xxxx.pinger files are never created. That's used as the source for fping, so that explains why there's no icmp traffic, but WHY are those files not being created?

**zabbixhack** · 12-04-2012, 00:37

greenlight ..

run this - "watch -d ls /tmp"

Ever see the pinger files?
Have you solved your problem.

Please someone reply. I have never received a reply from anyone here. Would be nice and reassuring.

**zabbixhack** · 12-04-2012, 02:00

SELinux bit me. Fixed

Yes. My problem made no sense. It should have been working.
When I saw that the tmp pinger files weren't being created - I thought.. how could something be prevented from writing to /tmp?

SELinux.

setenforce 0

edit /etc/sysconfig/selinux - set to disabled

I had disabled it at build time on 1 of my proxies, but not the other 2.

Working now.

**zabbixhack** · 12-04-2012, 02:06

SELinux Killing icmpping pinger files create

Just one more post... If I had looked in /var/log/audit/audit.log on my RHEL 6 proxy .. I would have seen tons of these..

type=AVC msg=audit(1334188168.457:60898): avc: denied { read } for pid=28049 comm="fping" path="/tmp/zabbix_proxy_mysql_22688.pinger" dev=sda6 ino=20 scontext=unconfined_u:system_r

ing_t:s0 tcontext=unconfined_u

bject_r:initrc_tmp_t:s0 tclass=file

Bummer huh?

**elvar** · 16-07-2012, 20:04

Originally posted by greenlight

I have set up several simple checks for icmpping but i am always getting a 0 result when fping gives results back

I have fping installed and this can be accesses by the zabbix user

-rws--x---. 1 root zabbix 27824 May 19 2010 /usr/sbin/fping

[root@monitor tmp]# su zabbix -c "fping -c 3 -p 100 -b 56 -t 150 monitor.rikrobson.co.uk"
monitor.rikrobson.co.uk : [0], 84 bytes, 7.19 ms (7.19 avg, 0% loss)
monitor.rikrobson.co.uk : [1], 84 bytes, 5.08 ms (6.13 avg, 0% loss)
monitor.rikrobson.co.uk : [2], 84 bytes, 5.13 ms (5.80 avg, 0% loss)

monitor.rikrobson.co.uk : xmt/rcv/%loss = 3/3/0%, min/avg/max = 5.08/5.80/7.19
[root@monitor tmp]#

as you can see this has the correct permissions.

tmp dir is also accessible by zabbix as you would expect.

I have a test server set up the same and this just works fine, but i cannot find the problem as they look the same

I have also tried icmppingsec and icmppingloss, Both return 0.

This indicates that the problem may be with zabbix communicating with fping

Does anyone have ant suggestions as to the cause of this problem?

Did you ever figure your problem with this? I'm having the exact same issue in Centos 6.3 and Zabbix 2.0.1.

**elvar** · 16-07-2012, 20:16

There definitely appears to be some sort of parsing issue. I ran this manually and the data is returned properly but never parsed / inserted correctly. I just end up with 0's in the 'latest data' section.

Code:

 [zabbix@hostname ~]$ /usr/sbin/fping -q -C10 -p1000 -t1000 2>&1 </tmp/zabbix_server_31256.pinger
96.120.26.249 : 10.79 10.21 10.85 10.21 10.66 9.34 10.28 9.96 14.46 10.55

Ad Widget

icmpping

icmpping

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment