Bug #17787 Restart NDBD before NF is complete

• Submitted: 28 Feb 2006 15:52
• Modified: 28 Feb 2006 17:17
• Reporter: Johan Andersson
• Status: Duplicate
• Category: MySQL Cluster: Cluster (NDB) storage engine
• Severity: S3 (Non-critical)
• Version: 5.x
• OS: Linux (Linux)
• CPU Architecture: Any

[28 Feb 2006 15:52] Johan Andersson

Description:
32-bit system. 
2 node cluster , 2 replicas, nodes are "started"

2 restart (from ndb_mgm) makes the ndbd node restart _before_  node 3 has completed NF handling of node 2.

A wait would be nice or a that node 2 actually asks node 3 if it has finished NF, before it tries to rejoin the cluster

How to repeat:
Have a large dataset so that NF handling takes some time so that the restarting node tries to join cluster before the started node is complete with NF

Suggested fix:
See above.

[28 Feb 2006 17:17] Jonas Oreland

duplicate of #16772