Cisco Bug: CSCuj89557 - DME failed after upgrading to build El-cap 323
Feb 12, 2018
- Cisco Unified Computing System
Known Affected Releases
Symptom: After upgrading from an older(delmar) to newer(elcap) release or after a power outage, when the DME comes up, we have seen that the read from sqlite db present in the nvram or flash fails sometimes. When this read fails, DME asserts and restarts. After restarting for 4 times, DME would give up and in this case UCSM wont come up. In case of an HA setup which is how customer always deploys UCSM, we have generally seen this failure always occurs on the secondary FI. Usually this wont be observed by the customer immediately when the db corruption occurs, since the secondary FI's DME is not involved at all and hence there is no impact at all.The customer will only notice this when they are doing a failover or upgrade/downgrade. In which case the downgrade/upgrade or failover will fail. If the customer has traffic going through the secondary FI, that would continue unaffected until the customer decides to downgrade/upgrade or failover . If the vnics are configured for failover, then during the recovery of the corrupted db, all vnic will failover to primary and hence there is no impact. The FI on which DB is corrupted, should have the following logs in /var/sysmgr/sam_logs/svc_sam_dme.log [INFO][0xb35a7080][Jun 26 16:25:19.785][exception_handling:rep] FATAL[3|0] nextRecord: Failed to Read record from SelectAll stmt 11 Conditions: Generally this occurs on the secondary FI and is only detected when we do downgrade/upgrade or failover. It is also suspected that sometimes power outages cause this issue. Exactly how the db corruption occurs is still unknown.
Bug details contain sensitive information and therefore require a Cisco.com account to be viewed.
Bug Details Include
- Full Description (including symptoms, conditions and workarounds)
- Known Fixed Releases
- Related Community Discussions
- Number of Related Support Cases