I wanted to share an interesting incident appeared when upgrading a CUCM Cluster ( Mega size one ) from 8.6.2(a) to 8.6.2aSU3 edition. After Fully successful upgrade with full services up and running and with full successful DB Replication , we have noticed that ALL the IP Phones ( Hard phones and softphones SCCP and SIP based but NOT GWs ,Trunks , or Media Resources ) get resested with "Reset-Restart" cause ( shown as cause 522 on IP Phone console logs ) every 30 minutes pricesly . All trials to find out why the mark of 30 minutes and why every 30 mintues All phone get instructed to reset and go back to their main subscriber . The reset it self is instructed , so the phones when it comes back up it register to its orginal subscriber. There was no apparent reason to have this behaviour. When engaging Cisco TAC support and even Cisco BU Engineers , they didn't able to locate the reason behind this specific 30 minutes reset instructions from All subscribers to their registered IP Phones. Note that :
- when shutting down the publisher node , this behaviour stops till the publisher comes up. once the publisher node is up , Immediatly Subscribers initiate this reset-restart.
- TFTP service is not function probarly and not responding to phone request for some reason but this should not affect phone reset behaviour.
- No Network issues has been noted at all and Once rolled back , This behaviour stopped and everything back to its normal operational behaviour.
I wonder if some of you faced a similar interesting problem ? What do you think the service or feautre that cause this exact 30 mintue restart ? where is this time-out timer that expires and cause the subscribers to instruct ALL phones to restart ?