[ejabberd] safe recovery from inconsistent_database/running_partitioned_network error

Igor Goryachev igor at goryachev.org
Thu Nov 8 16:54:08 MSK 2007

"Zbyszek Żółkiewski" <zbyszek at toliman.pl> writes:

>>      I am thinking on implementing a safe recovery routine against
>>      inconsistent_database/running_partitioned_network errors which
>>      occur
>>      sometimes in our network environment. Any suggestions?
>    can you explain more? we're running cluster also so we can exchange
>    experiences
This behaviour occurs while network failures, even the short ones. The
nodes begin to live as they are standalone, not connected in cluster.
The only one solution (for now) is to reboot all nodes except one in
order they could sync mnesia once again.

    Igor Goryachev              E-Mail/Jabber: igor at goryachev.org

More information about the ejabberd mailing list