Bug #41965 ndb_mgmd does not disconnect client after failed "transporter connect"

• Submitted: 8 Jan 2009 14:30
• Modified: 11 Feb 2009 16:40
• Reporter: Magnus Blåudd
• Status: Closed
• Category: MySQL Cluster: Cluster (NDB) storage engine
• Severity: S2 (Serious)
• Version: 6.2
• OS: Any
• Assigned to: Magnus Blåudd
• CPU Architecture: Any

[8 Jan 2009 14:30] Magnus Blåudd

Description:
While starting a transporter between ndbd and ndb_mgmd, the ndbd will normally first connect to the ndb_mgmd's port and issue the command "connect transporter", if the ndb_mgmd does not suceed in converting the connection to a transporter, it should disconnect the client. But that doe snot currently happen. Since this means we have different behaviour when a ndbd connects to a ndbd, we should make ndb_mgmd disconnect after failure to handle "connect transporter". If not disconnecting, the client will be hanging for "reead timeout" seconds, shich normally is set to 60.

How to repeat:
$ ndb_mgmd --nodaemon -f ../config.ini
NDB Cluster Management Server. mysql-5.1.30 ndb-6.2.17-GA
Id: 2, Command port: *:1186

$ telnet localhost 1186
transporter connect

23 99

<hang for 60 seconds>

Suggested fix:
Disconnect the client immediately after failing to convert socket to transporter.

[8 Jan 2009 14:36] Bugs System

A patch for this bug has been committed. After review, it may
be pushed to the relevant source trees for release in the next
version. You can access the patch from:

  http://lists.mysql.com/commits/62699

2786 Magnus Svensson	2009-01-08
      Bug #41965  ndb_mgmd does not disconnect client after failed "transporter connect"

[9 Jan 2009 12:25] Bugs System

Pushed into 5.1.30-ndb-6.2.17 (revid:[e-mail]-20090108143549-wca0uz8cohwsxz82) (version source revid:[e-mail]-20090108143549-wca0uz8cohwsxz82) (merge vers: 5.1.30-ndb-6.2.17) (pib:6)

[9 Jan 2009 12:26] Bugs System

Pushed into 5.1.30-ndb-6.3.21 (revid:[e-mail]-20090108152514-y0rxk5kqbpjk1tg9) (version source revid:[e-mail]-20090108152514-y0rxk5kqbpjk1tg9) (merge vers: 5.1.30-ndb-6.3.21) (pib:6)

[9 Jan 2009 12:30] Bugs System

Pushed into 5.1.30-ndb-6.4.1 (revid:[e-mail]-20090109121748-t0gy78i6su7ntseu) (version source revid:[e-mail]-20090109121748-t0gy78i6su7ntseu) (merge vers: 5.1.30-ndb-6.4.1) (pib:6)

[10 Feb 2009 20:13] Bugs System

Pushed into 6.0.10-alpha (revid:[e-mail]-20090210194937-s7xshv5l3m1v7wi9) (version source revid:[e-mail]-20090123110300-k3vjhnm1iop8c0h7) (merge vers: 6.0.10-alpha) (pib:6)

[11 Feb 2009 16:40] Jon Stephens

Documented in the NDB-6.2.17, 6.3.21, and 6.4.1 changelogs as follows:

        Formerly, the management server failed to create a transporter
        for a data node connection, read_timeout seconds elapsed before
        the data node was actually allowed to disconnect. Now in such
        cases the disconnection occurs immediately.