Bug #7124 ndb_mgmd is aborted on startup when using SHM connection.
Submitted: 9 Dec 2004 10:05 Modified: 13 Dec 2004 11:57
Reporter: Taro Kono Email Updates:
Status: Closed Impact on me:
Category:MySQL Cluster: Cluster (NDB) storage engine Severity:S1 (Critical)
Version:4.1.8 BK OS:Linux (Fedora Core 2)
Assigned to: Jonas Oreland CPU Architecture:Any

[9 Dec 2004 10:05] Taro Kono
I have tried to connect DB node and API node with SHM. 
But ndb_mgmd was aborted. I used config.ini file shown below 
with MySQL 4.1.8 built with --with-ndb-shm option 
on Fedora Core 2. 


ExecuteOnComputer: 1 

NoOfReplicas: 1 

ExecuteOnComputer: 1 
FileSystemPath: /home/taro/ndb/db/ndb_data 

ExecuteOnComputer: 1 

NodeId1: 1 
NodeId2: 2 

NodeId1: 2 
NodeId2: 3 
ShmKey: 1

How to repeat:
Run ndb_mgmd.

Suggested fix:
I found that ndb_mgmd was aborted at this line of fixPortNumber() in ConfigInfo.cpp 
when processing [SHM] block of config.ini file. 

Line 2970: 
require(ctx.m_currentSection->get("HostName1", &hostName1));
[13 Dec 2004 11:57] Jonas Oreland

I fixed this, but building 4.1.8 is already "in the build team"

So, the fix will show up in next 4.1.9 (which I don't know when it will be released)
and in 5.0.4 (which I don't know when it will be released).