Bug #45371 StartPartitionedTimeout and StartPartialTimeout def values not applied
Submitted: 8 Jun 2009 12:28 Modified: 10 Aug 2009 9:04
Reporter: Joffrey MICHAIE Email Updates:
Status: Closed Impact on me:
None 
Category:MySQL Cluster: Cluster (NDB) storage engine Severity:S3 (Non-critical)
Version:mysql-5.1-telco-7.0 OS:Any
Assigned to: CPU Architecture:Any
Tags: Found in 5.1.32-ndb-7.0.5, StartPartialTimeout, StartPartitionedTimeout

[8 Jun 2009 12:28] Joffrey MICHAIE 
Description:
Hi,

Documentation mentions that default value for :
- StartPartialTimeout is 30000 ms
- StartPartitionedTimeout 60000 ms

But, when not explicitely set :

2009-06-05 20:09:04 [MgmSrvr] INFO     -- Node 10: Initial start, waiting for 0000000000000800 to connect,  nodes [ all: 0000000000000c00 connected: 0000000000000400 no-wait: 0000000000000000 ]
2009-06-05 20:12:28 [MgmSrvr] INFO     -- Node 10: Initial start, waiting for 0000000000000800 to connect,  nodes [ all: 0000000000000c00 connected: 0000000000000400 no-wait: 0000000000000000 ]

(...)

So the NDBD node will never start.

----

And, when explicitly set (see change with timeout in log output) :

2009-06-05 21:29:45 [MgmSrvr] INFO     -- Node 10: Waiting 28 sec for nodes 0000000000000800 to connect, nodes [ all: 0000000000000c00 connected: 0000000000000400 no-wait: 0000000000000000 ]
2009-06-05 21:30:13 [MgmSrvr] INFO     -- Node 10: Waiting 1 sec for nodes 0000000000000800 to connect, nodes [ all: 0000000000000c00 connected: 0000000000000400 no-wait: 0000000000000000 ]
2009-06-05 21:30:15 [MgmSrvr] INFO     -- Node 10: Waiting 58 sec for non partitioned start, nodes [ all: 0000000000000c00 connected: 0000000000000400 missing: 0000000000000800 no-wait: 0000000000000000 ]
2009-06-05 21:31:12 [MgmSrvr] INFO     -- Node 10: Waiting 1 sec for non partitioned start, nodes [ all: 0000000000000c00 connected: 0000000000000400 missing: 0000000000000800 no-wait: 0000000000000000 ]
2009-06-05 21:31:22 [MgmSrvr] INFO     -- Node 10: Started (mysql-5.1.32 ndb-7.0.5)

----

Verified with and without --intial

Workaround : start ndbd --nowait-nodes=[Node_ID]

----

ndb_config output (without start*timeout set):

StartPartialTimeout (Non-negative Integer)
Time to wait before trying to start wo/ all nodes. 0=Wait forever
Default: 30000 (Min: 0, Max: 4294967039)

StartPartitionedTimeout (Non-negative Integer)
Time to wait before trying to start partitioned. 0=Wait forever
Default: 60000 (Min: 0, Max: 4294967039)

How to repeat:
Run ndb_mgmd with default (no Start*Timeout in config)

Start node 1 (ndbd --inital)

Look at the logs

Ndbd never starts

Add to config :

StartPartialTimeout=30000
StartPartitionedTimeout=60000

Restart mgmd

Restart ndbd

It will start after both Timeouts (1m30s)

Regards,
Joffrey