Commit Failed (HA active-passive)

[paloalto.netfos]

Hi,

Commit on customer PA500 Cluster running in Active-Passive mode on PANOS 3.1.6

Details      device: Client device registered in the middle of a commit. Aborting current commit.
Commit failed

system log:

Receive Time    Type      Severity      Object  Event    Description
10/26 13:54:20   general  high                        general  Commit job failed for user admin
10/26 13:54:10   general  informational            general  localhost.localdomain disconnected

Could you please help us to diagnose and resolve this issue  ?

Thanks

[bradenmcg]

It would be helpful to have some idea of the configuration of the unit...  based on the error message alone, I'm wondering if you are using SSL-VPN or GlobalProtect?  Those are the only "clients" I can think of that would "register"...

[gswcowboy]

Get more details from the the job id that failed via >show job id <value here> to garner more detailed information. You could also scour the ms.log and/or devsrvr.logs for additional information as we have little to go by. Are you able to confirm whether or not either the mgmtsrvr and/or the devsrvr processes are running high memory utilization? >show system software resources | match dev or show system software resources | match mgmt.

[paloalto.netfos · Accepted solution]

Manual reboot PA, device auto-commit job is fail,

after commit job fail (3 time),device auto reboot and into maintenance mode
manual reboot again,still commit fail and into maintenance mode again
so i reset device to factory default (PanOS3.1.6)

when factory reset finish, situation still happen

system log:

Type      Severity         Object    Description
general   high               general   Commit job failed for user admin
general   informational  general   localhost.localdomain disconnected

finally ,upgrade FW to 3.1.10,problem no longer occurs

[gswcowboy]

autocom job failing can be rooted to the PAN device's inability to commit a policy to the config, for some reason or another. glad you were able to resolve the issue nonetheless.