Percona XtraDB Cluster moved to https://jira.percona.com/projects/PXC

  1. Bug #1108035
  2. Comment #10

Comment 10 for bug 1108035

Revision history for this message
Jay Janssen (jay-janssen) wrote : Re: PXC innodb_buffer_pool_restore_at_startup and node joining cluster io resources race condition

Ovais,
  I agree with you, but it depends on what symptoms are actually being seen. If the "cluster stalls" are really just flow control, then I agree this isn't technically a bug. If they are something worse, then there may be a larger problem.

 Ilias, can you confirm the "stalls" you see? Are they just flow control because the node doing the LRU restore is busy? I would expect to see that node increasing wsrep_flow_control_sent during this window. Or, is it some other symptoms on the cluster at this time?