Hello, everybody!
I configured Zabbix to recieve SNMP-traps using PERL-script. PERL-script collected traps into the file /tmp/zabbix_traps.tmp. Zabbix put traps conforming regexp {$REGEXP} from the file into the configured items «SNMP trap». These items have a «LOG» data type and a snmptrap[{$REGEXP}] key. Everything**works correctly. The size of file zabbix_traps.tmp is much less than 2Gb.
But!
Periodically, spontaneously Zabbix repeatedly parses the file with traps (zabbix_traps.tmp) and processes the already processed traps again. When it happens, in all traps which were written into the Zabbix database, Zabbix sets the same false timestamp and I got alarm messages onto the interface because opening-traps and closing-traps have the same timestamp and change places.
Has anybody encountered a similar problem?
Could anybody help me or propose a suggestion?
As illustration, in the latest data I saw this:
Timestamp Value
2016-12-11 04:27:39 15:19:32 2016/12/01 PDU INFO: ... CLOSING TRAP ...
2016-12-11 04:27:39 15:19:25 2016/12/01 PDU INFO: ... OPENING TRAP ...
2016-12-11 04:27:39 15:17:57 2016/12/01 PDU INFO: ... CLOSING TRAP ...
2016-12-11 04:27:39 15:17:52 2016/12/01 PDU INFO: ... OPENING TRAP ...
2016-12-11 04:27:39 15:07:30 2016/12/01 PDU INFO: ... CLOSING TRAP ...
2016-12-11 04:27:39 15:07:27 2016/12/01 PDU INFO: ... OPENING TRAP ...
I configured Zabbix to recieve SNMP-traps using PERL-script. PERL-script collected traps into the file /tmp/zabbix_traps.tmp. Zabbix put traps conforming regexp {$REGEXP} from the file into the configured items «SNMP trap». These items have a «LOG» data type and a snmptrap[{$REGEXP}] key. Everything**works correctly. The size of file zabbix_traps.tmp is much less than 2Gb.
But!
Periodically, spontaneously Zabbix repeatedly parses the file with traps (zabbix_traps.tmp) and processes the already processed traps again. When it happens, in all traps which were written into the Zabbix database, Zabbix sets the same false timestamp and I got alarm messages onto the interface because opening-traps and closing-traps have the same timestamp and change places.
Has anybody encountered a similar problem?
Could anybody help me or propose a suggestion?
As illustration, in the latest data I saw this:
Timestamp Value
2016-12-11 04:27:39 15:19:32 2016/12/01 PDU INFO: ... CLOSING TRAP ...
2016-12-11 04:27:39 15:19:25 2016/12/01 PDU INFO: ... OPENING TRAP ...
2016-12-11 04:27:39 15:17:57 2016/12/01 PDU INFO: ... CLOSING TRAP ...
2016-12-11 04:27:39 15:17:52 2016/12/01 PDU INFO: ... OPENING TRAP ...
2016-12-11 04:27:39 15:07:30 2016/12/01 PDU INFO: ... CLOSING TRAP ...
2016-12-11 04:27:39 15:07:27 2016/12/01 PDU INFO: ... OPENING TRAP ...