starlingx/snmp-armada-app
Code Issues Proposed changes

Retry get alarms/events for SNMP if it fails

Browse Source 
Root cause:
SNMP sub-agent uses fm-common library to connect database
to obtain alarm/event list. But after swact, it does not
detect disconnection and as the result, the functions to
obtain alarm/event returns false and does not return proper
list.

Analysis:
fm-common library to connect to the database. It checks
connection by PQstatus() and it will reconnect if it find
connection error. But PQstatus responses CONNECTION_OK
even after swact. The other option to check connection is
using PQPing(). But this method also responses PQPING_OK
even after awact, too.
Other option is sending real SQL and it is same as to
retry. So this solution is to retry to obtain alarm/
event list if it fails.

Solution:
When SNMP sub-agent calls to obtain alarm/event list
and it returns false, retry it. Internally, it reconnects
automatically and it returns the proper list.

Test Plan:
Test the following steps for snmpget, snmpgetnext, snmpbulkget
and snmpwalk in IPv4 and IPv6:
1. Execute "system host-swact controller-n"
2. Send snmp command and record response
3. Send snmp command and record response
4. Compare 2. and 3. response and confirm they are matched
=> all cases PASS

Closes-bug: 1949083

Signed-off-by: Takamasa Takenaka <takamasa.takenaka@windriver.com>
Change-Id: Ib58f9f657bb5ac8c3b505eee5443a24bedb2463e
This commit is contained in:
Takamasa Takenaka 2021-10-28 16:41:20 -03:00
parent 1904f3ad6d
commit 5642841ebc
2 changed files with 30 additions and 4 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

17
stx-snmp-helm/centos/docker/stx-fm-subagent/src/snmpActiveAlarms.c
View File

@ -1,5 +1,5 @@
/*
* Copyright (c) 2020 Wind River Systems, Inc.
* Copyright (c) 2020-2021 Wind River Systems, Inc.
*
* SPDX-License-Identifier: Apache-2.0
*
@ -26,6 +26,7 @@ static long Active_Alarm_Count = 0;
static struct activealarm *alarm_list;
static struct activealarm *alarmaddr, savealarm, *savealarmaddr;
static int saveIndex = 0;
static int retryCount = 5;
static char saveUuid[36];
static long LastLoad = 0; /* ET in secs at last table load */
extern long long_return;
@ -64,7 +65,19 @@ Alarm_Scan_Init()
/*
* query active alarm list from DB
*/
if (fm_snmp_util_get_all_alarms(getAlarmSession(), &aquery) != true){
bool isAlarmObtained = false;
for (i = 0; i < retryCount; i++){
if (fm_snmp_util_get_all_alarms(getAlarmSession(), &aquery)){
DEBUGMSG(("cgtsAgentPlugin", "get_all_alarms done\n"));
isAlarmObtained = true;
break;
} else {
DEBUGMSG(("cgtsAgentPlugin",
"get_all_alarms returns false (%zu/%d). Try again\n",
(i+1), retryCount));
}
}
if (!isAlarmObtained) {
DEBUGMSG(("cgtsAgentPlugin", "get_all_alarms from db failed\n"));
return;
}

17
stx-snmp-helm/centos/docker/stx-fm-subagent/src/snmpEventLogs.c
View File

@ -1,5 +1,5 @@
/*
* Copyright (c) 2020 Wind River Systems, Inc.
* Copyright (c) 2020-2021 Wind River Systems, Inc.
*
* SPDX-License-Identifier: Apache-2.0
*
@ -26,6 +26,7 @@ static long Event_Log_Count = 0;
static struct activealarm *alarm_list;
static struct activealarm *alarmaddr, savealarm, *savealarmaddr;
static int saveIndex = 0;
static int retryCount = 5;
static char saveUuid[36];
static long LastLoad = 0; /* ET in secs at last table load */
extern long long_return;
@ -64,7 +65,19 @@ Event_Log_Scan_Init()
/*
* query event log list from DB
*/
if (fm_snmp_util_get_all_event_logs(getAlarmSession(), &aquery) != true){
bool isEventObtained = false;
for (i = 0; i < retryCount; i++){
if (fm_snmp_util_get_all_event_logs(getAlarmSession(), &aquery)){
DEBUGMSG(("cgtsAgentPlugin", "get_all_event_logs done\n"));
isEventObtained = true;
break;
} else {
DEBUGMSG(("cgtsAgentPlugin",
"get_all_event_logs returns false (%zu/%d). Try again\n",
(i+1), retryCount));
}
}
if (!isEventObtained) {
DEBUGMSG(("cgtsAgentPlugin", "get_all_event_logs from db failed\n"));
return;
}