| Bug #48380 | ndbd not showing connected when in start-phase 0 and 1 | ||
|---|---|---|---|
| Submitted: | 28 Oct 2009 14:45 | Modified: | 30 Oct 2009 11:30 |
| Reporter: | Geert Vanderkelen | Email Updates: | |
| Status: | Duplicate | Impact on me: | |
| Category: | MySQL Cluster: Cluster (NDB) storage engine | Severity: | S3 (Non-critical) |
| Version: | 5.1.35-ndb-6.3.26 | OS: | Any |
| Assigned to: | CPU Architecture: | Any | |
| Tags: | cluster, NDB_MGM, start | ||
[28 Oct 2009 16:41]
Magnus Blåudd
After fixing bug#48301, it should be possible to return CONNECTED(which is actually a new status for API and MGM) when asking status for the ndbd. I.e in the case when node is not connected to ClusterMgr with transporter yet but has allocated node id.
[30 Oct 2009 11:30]
Geert Vanderkelen
Duplicate of bug#47818.
Description: The output of SHOW in ndb_mgm doesn't show the data node as connected, although it is starting and allocating a (huge) amount of memory. shell> ndbd Cluster log shows: 2009-10-28 08:41:23 [MgmSrvr] INFO -- Mgmt server state: nodeid 3 reserved for ip 127.0.0.1, m_reserved_nodes 000000000000000000000000000000000000000000000000000000000000000a. 2009-10-28 08:41:23 [MgmSrvr] INFO -- Node 1: Node 3 Connected 2009-10-28 08:41:25 [MgmSrvr] INFO -- Mgmt server state: nodeid 3 freed, m_reserved_nodes 0000000000000000000000000000000000000000000000000000000000000002. ndb_mgm> SHOW [ndbd(NDB)] 1 node(s) id=3 (not connected, accepting connect from localhost) Then it continues: 2009-10-28 08:42:02 [MgmSrvr] INFO -- Node 3: Start phase 2 completed (system restart) 2009-10-28 08:42:03 [MgmSrvr] INFO -- Node 3: Start phase 3 completed (system restart) 2009-10-28 08:42:30 [MgmSrvr] INFO -- Node 3: DICT: index 329 activated done ndb_mgm> SHOW [ndbd(NDB)] 1 node(s) id=3 @127.0.0.1 (mysql-5.1.35 ndb-6.3.26, starting, Nodegroup: 0, Master) How to repeat: Start a data node which need to allocate a huge amount of memory. I used a 1 data node setup with 50Gb DataMemory. Used MySQL Cluster 6.3.26 Suggested fix: Discussed already: the ndbd already connected to ndb_mgmd and fetched configuration. Also, it allocated a nodeid, so there shouldn't be a reason why not showing the status in the SHOW-output of ndb_mgm.