-
Problem report
-
Resolution: Fixed
-
Trivial
-
4.0.19, 4.4.7, 5.0.0beta1, 5.0.0beta2
-
None
-
Sprint 63 (Apr 2020), Sprint 64 (May 2020)
-
0.25
Retry was added under ZBX-8538, however after retry was made and check is moved to unreachable poller there is no reason to retry such checks as it results in 2 * Timeout per check which makes unreachable poller busier.
Timeout=7
60522:20200415:155808.673 In zbx_snmp_get_values() num:1 level:0 60522:20200415:155822.686 zbx_snmp_get_values() snmp_synch_response() status:2 s_snmp_errno:-24 errstat:-1 mapping_num:1
example fix but better to pass retry count as parameter:
diff --git a/src/zabbix_server/poller/checks_snmp.c b/src/zabbix_server/poller/checks_snmp.c index 58c2d760e5..183e83782c 100644 --- a/src/zabbix_server/poller/checks_snmp.c +++ b/src/zabbix_server/poller/checks_snmp.c @@ -28,6 +28,7 @@ #include "comms.h" #include "zbxalgo.h" #include "zbxjson.h" +extern unsigned char process_type; /* * SNMP Dynamic Index Cache @@ -1283,7 +1284,7 @@ static int zbx_snmp_get_values(struct snmp_session *ss, const DC_ITEM *items, ch goto out; } - ss->retries = (1 == mapping_num && 0 == level ? 1 : 0); + ss->retries = (1 == mapping_num && 0 == level && ZBX_POLLER_TYPE_UNREACHABLE != process_type ? 1 : 0); retry: status = snmp_synch_response(ss, pdu, &response);
- caused by
-
-
- Closed
-
