Not the answer you need?
Register and ask your own question!

Cluster always sync with the same Donor

mstenzmstenz EntrantInactive User Role Beginner
Hi, I don't know why this happens. I have a cluster with 6 nodes + 1 arbitrator and everytime it get out of sync or there is a problem with the cluster (which is currently about every 1 hours in average (the Cluster Software as very unstable in my environment) the nodes are syncing with the same donor. Problem of this is that this is the node with the highest latency. But for me it would make more sense to sync with the node with lowest latency. My nodes are not in the same datacenter but all over the world. Biggest latency is about 500ms.

WSREP: Member 2.0 (garb) requested state transfer from '*any*'. Selected 0.0 (xxxxxxxx)(SYNCED) as donor.

Comments

  • rollsrolls Contributor Inactive User Role Beginner
    Change the order on wsrep_cluster_address=gcomm:// because the donor is always the first address of the pool
  • mstenzmstenz Entrant Inactive User Role Beginner
    thank you. thats a very good information.
  • aaronlaaronl Entrant Inactive User Role Beginner
    Have you tweaked the timeout settings?
    wsrep_provider_options  = "gmcast.segment=0; gcache.size=1G; evs.suspect_timeout = PT90S; evs.inactive_timeout = PT5M; evs.install_timeout = PT5M; evs.join_retrans_period=PT1.0S"
    
    I had to increase the timeouts significantly to handle a cluster spread across data centers on the east and west coasts of the US.
Sign In or Register to comment.

MySQL, InnoDB, MariaDB and MongoDB are trademarks of their respective owners.
Copyright ©2005 - 2020 Percona LLC. All rights reserved.