Bug #20418 NdbConnection::connect() and DNS timeouts
Submitted: 13 Jun 2006 8:30 Modified: 3 May 2012 6:50
Reporter: Hartmut Holzgraefe Email Updates:
Status: Can't repeat Impact on me:
None 
Category:MySQL Cluster: Cluster (NDB) storage engine Severity:S3 (Non-critical)
Version:mysql-5.0 OS:Solaris (Solaris)
Assigned to: CPU Architecture:Any

[13 Jun 2006 8:30] Hartmut Holzgraefe
Description:
When trying to connect an API node to the cluster while the management nodes DNS hostname is not resolveable on the API client side due to DNS problems connect() can get stuck for several minutes until finaly receiving a DNS lookup timeout error. The connect() parameters for "number of retries" and "time between retries" implicitly suggest that one has full control over the timeing behavior of the connect() call and that one can rely on connect() returning after e.g. max. 15s when specifying 3 retries and 5s delays between retries.

How to repeat:
see description

Suggested fix:
Use async DNS lookups or move DNS lookups to a seperate application thread and terminate this after the retry delay time has expired
[25 Mar 2007 7:44] Stewart Smith
With mgmapi timeouts (specifically, with connect_timeout and now mgmapi timeouts) I don't think this should be a problem. If can be repeated with latest 5.1, reopen.

I don't have the intention of backporting mgmapi timeouts to 5.0 though.
[25 Apr 2007 23:00] Bugs System
No feedback was provided for this bug for over a month, so it is
being suspended automatically. If you are able to provide the
information that was originally requested, please do so and change
the status of the bug back to "Open".
[18 May 2009 22:22] Jared S
Server bug here in progress, http://bugs.mysql.com/38247