zabbix server process 100% cpu every 5 hours?

[Tristan]

sinds i have update my zabbix server form 1.4.4 to 1.4.5 approx. every 5 hours my zabbix process eats 100% cpu load on my zabbix server. I've read about fixes in 1.4.5 but it gives me only troubles? now i need to restart my zabbix_server process every 5 hours. is there a solution for this?

[Alexei]

Please could you run:

strace -p <pid>

where <pid> is a process ID of the ZABBIX process eating 100% of CPU time.

Please double-check you're running 1.4.5!

[Tristan]

Please could you run:

strace -p <pid>

where <pid> is a process ID of the ZABBIX process eating 100% of CPU time.

Please double-check you're running 1.4.5!

Hi,

in the webinterface i read zabbix 1.4.5 so i think it's 1.4.5. I've upgrade my zabbix install 2 times to ensure it's good. this is my output:

it is not every 5 hours. my last restart was 2 hours ago:

this is my strace output:

gettimeofday({1207080123, 79585}, NULL) = 0
select(1, [0], NULL, NULL, {0, 0}) = 1 (in [0], left {0, 0})
recvfrom(0, 0x81438a8, 65536, 0, 0x811e568, 0xbfae51b4) = -1 ENOTSOCK (Socket operation on non-socket)
gettimeofday({1207080123, 77019}, NULL) = 0
select(1, [0], NULL, NULL, {0, 0}) = 1 (in [0], left {0, 0})
recvfrom(0, 0x81438a8, 65536, 0, 0x8114710, 0xbfae51b4) = -1 ENOTSOCK (Socket operation on non-socket)
gettimeofday({1207080123, 77226}, NULL) = 0
select(1, [0], NULL, NULL, {0, 0}) = 1 (in [0], left {0, 0})
recvfrom(0, 0x81438a8, 65536, 0, 0x811e568, 0xbfae51b4) = -1 ENOTSOCK (Socket operation on non-socket)
gettimeofday({1207080123, 77244}, NULL) = 0
select(1, [0], NULL, NULL, {0, 0}) = 1 (in [0], left {0, 0})
recvfrom(0, 0x81438a8, 65536, 0, 0x8114710, 0xbfae51b4) = -1 ENOTSOCK (Socket operation on non-socket)
gettimeofday({1207080123, 77453}, NULL) = 0
select(1, [0], NULL, NULL, {0, 0}) = 1 (in [0], left {0, 0})
recvfrom(0, 0x81438a8, 65536, 0, 0x811e568, 0xbfae51b4) = -1 ENOTSOCK (Socket operation on non-socket)
gettimeofday({1207080123, 77652}, NULL) = 0
select(1, [0], NULL, NULL, {0, 0}) = 1 (in [0], left {0, 0})
recvfrom(0, 0x81438a8, 65536, 0, 0x8114710, 0xbfae51b4) = -1 ENOTSOCK (Socket operation on non-socket)
gettimeofday({1207080123, 77941}, NULL) = 0
select(1, [0], NULL, NULL, {0, 0} <unfinished ...>

thnx!

[bbrendon]

Double check like this!

$ zabbix_server --version
ZABBIX Server (daemon) v1.4.5 (25 March 2008)
Compilation time: Mar 27 2008 09:35:37
$

[Alexei]

gettimeofday({1207080123, 79585}, NULL) = 0
select(1, [0], NULL, NULL, {0, 0}) = 1 (in [0], left {0, 0})
recvfrom(0, 0x81438a8, 65536, 0, 0x811e568, 0xbfae51b4) = -1 ENOTSOCK (Socket operation on non-socket)

It looks very much like a problem we had in 1.4.4. The problem was fixed in 1.4.5.

[Tristan]

It looks very much like a problem we had in 1.4.4. The problem was fixed in 1.4.5.

thnx for your reply:
this is my version output:
zabbix:~ # zabbix_server --version
ZABBIX Server (daemon) v1.4.5 (25 March 2008)
Compilation time: Apr 1 2008 14:57:28
zabbix:~ #

I hope you can help me

[Alexei]

I am still not convinced. I do not believe the problem still exists in 1.4.5.

Sorry for asking this, but please check again that this is the only zabbix_server binary on your system and the running ZABBIX server is indeed 1.4.5

[Tristan]

Hello,

yesterday i have taken a hp server (xeon 3.06 with 4gb ram) and i have installed a fresh installation of zabbix, and restore my database. This morning the zabbix_server process eats 100% again. It's still collecting data. when i look at the graphs i see that al of my monitored windows servers eats more cpu, but this isn't true.

strace output is the same. any idea?

[Alexei]

strace output is the same. any idea?

Please re-read my previous message. The trace file does not look like an output of ZABBIX 1.4.5!

[Tristan]

update

Please re-read my previous message. The trace file does not look like an output of ZABBIX 1.4.5!

Hi

i've read about a problem in a topic with a error in the trapper.c file.
when i look at my file it doesn't display the 2 lines

http://www.zabbix.com/forum/showthread.php?t=9274

so i have compiled everything again en it seems to work now.

maybe it happens becouse i grabbed the source form my old zabbix server. I have reinstalled it now!

thnx for your time and i keep you informed.

b.t.w strange that the trace output is not of 1.4.5 and my zabbix_server --version said that it is 1.4.5. I think that something has terrible gone wrong.
I know for sure that there is 1 zabbix_server binary on my system

[Alexei]

i've read about a problem in a topic with a error in the trapper.c file.
when i look at my file it doesn't display the 2 lines

The 2 lines is a fix for a crash-related problem. It has nothing to do with your issue.

Anyway I am glad you are doing good progress

[Tristan]

to bad.. it happen again

Hello,

the problem wasn't solved. to bad. the same evening it happens again around 10 o'clock. But i have disabled the housekeeping and the problem is gone for 2 days. do jou have an idea?

monday i will enable the house keeping to see what heappens

[Alexei]

It seems that you saw the housekeeper process eating good percentage of CPU time.

[Tristan]

It seems that you saw the housekeeper process eating good percentage of CPU time.

I understand that. But if i enable the housekeeper again the problem is back. And it is running at 100% until i kill the process by my self.

is this a problem within my database?

i need to resolve this, otherwise i can't use my housekeeping.

[stefanw]

Have the same problem, zabbix_server using 100% of the CPU. I upgraded to 1.4.5 yesterday, but still the same issue.


recvfrom(0, 0x67d960, 65536, 0, 0x633b60, 0x7fff30b53e3c) = -1 ENOTSOCK (Socket operation on non-socket)
select(1, [0], NULL, NULL, {0, 0}) = 1 (in [0], left {0, 0})
recvfrom(0, 0x67d960, 65536, 0, 0x631420, 0x7fff30b53e3c) = -1 ENOTSOCK (Socket operation on non-socket)
select(1, [0], NULL, NULL, {0, 0}) = 1 (in [0], left {0, 0})
recvfrom(0, 0x67d960, 65536, 0, 0x62fdc0, 0x7fff30b53e3c) = -1 ENOTSOCK (Socket operation on non-socket)
select(1, [0], NULL, NULL, {0, 0}) = 1 (in [0], left {0, 0})
recvfrom(0, 0x67d960, 65536, 0, 0x61eef0, 0x7fff30b53e3c) = -1 ENOTSOCK (Socket operation on non-socket)
select(1, [0], NULL, NULL, {0, 0}) = 1 (in [0], left {0, 0})
recvfrom(0, 0x67d960, 65536, 0, 0x633b60, 0x7fff30b53e3c) = -1 ENOTSOCK (Socket operation on non-socket)
select(1, [0], NULL, NULL, {0, 0}) = 1 (in [0], left {0, 0})
recvfrom(0, 0x67d960, 65536, 0, 0x631420, 0x7fff30b53e3c) = -1 ENOTSOCK (Socket operation on non-socket)
select(1, [0], NULL, NULL, {0, 0}) = 1 (in [0], left {0, 0})
recvfrom(0, 0x67d960, 65536, 0, 0x62fdc0, 0x7fff30b53e3c) = -1 ENOTSOCK (Socket operation on non-socket)
select(1, [0], NULL, NULL, {0, 0}) = 1 (in [0], left {0, 0})
recvfrom(0, 0x67d960, 65536, 0, 0x61eef0, 0x7fff30b53e3c) = -1 ENOTSOCK (Socket operation on non-socket)
select(1, [0], NULL, NULL, {0, 0}) = 1 (in [0], left {0, 0})
recvfrom(0, 0x67d960, 65536, 0, 0x633b60, 0x7fff30b53e3c) = -1 ENOTSOCK (Socket operation on non-socket)
select(1, [0], NULL, NULL, {0, 0}) = 1 (in [0], left {0, 0})
recvfrom(0, 0x67d960, 65536, 0, 0x631420, 0x7fff30b53e3c) = -1 ENOTSOCK (Socket operation on non-socket)
select(1, [0], NULL, NULL, {0, 0}) = 1 (in [0], left {0, 0})
recvfrom(0, 0x67d960, 65536, 0, 0x62fdc0, 0x7fff30b53e3c) = -1 ENOTSOCK (Socket operation on non-socket)
select(1, [0], NULL, NULL, {0, 0}) = 1 (in [0], left {0, 0})
recvfrom(0, 0x67d960, 65536, 0, 0x61eef0, 0x7fff30b53e3c) = -1 ENOTSOCK (Socket operation on non-socket)
select(1, [0], NULL, NULL, {0, 0}) = 1 (in [0], left {0, 0})
recvfrom(0, 0x67d960, 65536, 0, 0x633b60, 0x7fff30b53e3c) = -1 ENOTSOCK (Socket operation on non-socket)
select(1, [0], NULL, NULL, {0, 0}) = 1 (in [0], left {0, 0})
recvfrom(0, 0x67d960, 65536, 0, 0x631420, 0x7fff30b53e3c) = -1 ENOTSOCK (Socket operation on non-socket)

# /usr/local/sbin/zabbix_server -V
ZABBIX Server (daemon) v1.4.5 (25 March 2008)
Compilation time: Apr 9 2008 10:15:10

# uname -a
Linux xxxxxx 2.6.16.53-0.16-smp #1 SMP Tue Oct 2 16:57:49 UTC 2007 x86_64 x86_64 x86_64 GNU/Linux