11-14-2012 12:56 PM
We recently upgraded our controllers to 18.104.22.168 and have been fairly happy with it. However, since upgrading I have random AP's that will be reported as Down when they are really not down at all. I run scripts against our controllers to poll for information, such as AP up/down status.
These AP's primarily live on one of our Local controllers. (Current setup is: A Master and 2 Locals). They are set to failover to the Master if there is any issue with the local.
What I "think" is happening is the AP is associating with the Master for some short time, then going back to the Local. The problem is the Master is seeing the AP as down, once it goes back to the local.
I run several snmp scripts against the controllers. I first grab a list of all of the AP's:
snmpwalk -v2c Aruba-master -c some_snmp_password .22.214.171.124.4.1.148126.96.36.199.188.8.131.52.1.3
This returns a complete list of AP's:
I then poll each one using the AP string from above:
snmpwalk -v2c Aruba-master -c some_snmp_password .184.108.40.206.4.1.148220.127.116.11.18.104.22.168.1.19."AP string"
snmpwalk -v2c Aruba-master -c some_snmp_password .22.214.171.124.4.1.148126.96.36.199.188.8.131.52.184.108.40.206.200.
This returns either a 1 for up, or a 2 for down:
In this case, 2 means it's down.
If I run the same snmp script against the Local, it will come back with a INTEGER: 1 or UP.
Logging on to both controllers gives me similar results. Doing a: show AP database | include Down will show the AP as Down on the Master and "Up" on the Local. The AP is also working and passing data.
The Master controller will eventually update it' status to show the correct AP state, but only after I've gotten a bunch of email messages that an AP is down. I've tried several things to get the Master to update itself, including doing a Write Mem command. Which does not update the AP's to their correct state. If I show all of the Controller's status - show switches, they claim to be ins sync - Update Successful. (A reboot does get things back on track but for obvious reasons, I'd like to not do this.)
This brings up several questions....
1. What is the best way to poll four up/down status? (Since the above method no longer seems to be reliable).
2. Why are the controllers reporting different UP/Down status?
3. How do I modify the master's behavior to work as it should (i.e. show the correct up/down state of the AP)?
Anyone have any insight into this?
11-14-2012 03:07 PM
This isn't particularly helpful I know, but I think you'd best log that with TAC. If there's delays between the master and local in terms of AP up/down status, there might be a timer in 6.1.3,.5 for data sync that wasn't there before that I'm not aware of (a guess).
Debugging this will be a nightmare I'd suspect, so TAC will likely be your way forward.