[PCA 2.x] - After a Power Outage a PCA Operating on only One Leaf Switch Can Not Locate its Peer vPC
(Doc ID 2987504.1)
Last updated on DECEMBER 14, 2023
Applies to:
Private Cloud Appliance X8-2 - Version All Versions and laterInformation in this document applies to any platform.
Symptoms
In this case, the customer had only one leaf switch up and functional and suffered a power outage.
As a test, we simulated the issue by unplugging one of the leaf switches and then shutting down the PCA via the PDUs.
When the rack is powered on, the functioning leaf switch fails to come up fully because it can't locate its peer vPC, which results in "None" as its operational state and the network is in an incomplete state.
Symptoms may include:
- Unable to ping the spines from the leaf switch.
- Unable to ping the .40 net, which is the ZFS storage.
- Unable to fully bring the Management Nodes online (ovcamn05r1/ovcamn06r1) because they can't contact the storage.
As a test with the PCA rack up and unplugging one of of the switches, if ovcasw16r1 goes down, ovcasw15r1 remains the "primary".
If ovcasw15r1 goes down, ovcasw16r1 becomes "secondary, operational primary".
When the rack is running and either one of those goes down, the rack continues to functions properly.
Even taking down both leaf switches and restoring their functionality, in one test, ovcasw16r1 ended up as "secondary, operational primary".
Everything still functioned properly.
TEST SIMULATING THE CUSTOMER'S SCENARIO
Powered down the whole rack. Unplugged ovcasw15r1 and powered the rack back on.
ZFS came online. ovcasw16r1, Spine switches and Management switch come online.
ovcasw16r1 status after power outage with ovcasw15r1 unplugged:
Changes
Power outage while peer leaf switch down (ovcasw16r1). Auto-recovery is disabled as part of the original design.
Cause
To view full details, sign in with your My Oracle Support account. |
|
Don't have a My Oracle Support account? Click to get started! |
In this Document
Symptoms |
Changes |
Cause |
Solution |
References |