Cisco Bug: CSCur76872 - Stack manager goes down when the system runs out of SDP buffer.
Oct 27, 2016
- Cisco Catalyst 3850 Series Switches
Known Affected Releases
Symptom: The SDP (stack discovery protocol) buffer pool gets filled completely and hence the stack manager process goes down. This may manifest itself as any of a number of symptoms, many of which would appear at first glance to be due to stack cable instability (e.g. switches dropping from the stack and rejoining, stack merges, keepalive timeouts, etc). The fingerprint of this issue is one of more sets of messages similar to the following in the "Trace buffer: stack-mgr-events" in any system-report or switch-report files which are generated during an event: [01/26/15 23:50:49.498 UTC 283424 5943] No buffers available in chunk ssm 0x55c40ae0, buf (nil)Try increasing the chunk size [01/26/15 23:50:49.498 UTC 283425 5943] smn-error: ran out of buffers Conditions: The issue is seen after SSO In mixed stack and the issue also gets triggered randomly whenever there is a PDS buffer overflow (a high rate of L2 & L3 protocol packets) in any stack. Issue has been seen: - After an SSO in a mixed stack - When any SDP buffer overflow occurs in any stack, which can be triggered by a high rate of L2 / L3 protocol packets.
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