This behaviour is due to a bug (#3657) which has been solved and the fix is on its way out.
Rgrds Mikael
2004-05-19 kl. 13.35 skrev Maciek Dobrzanski:
Hi,
I have configured MySQL Cluster on two machines with 2 DB nodes
(NoOfReplicas = 2) and 2 MySQL API nodes, one of each node type on both
systems. The config is almost the same as the one of 2-node demo. The
cluster is working fine as long as all DB nodes are operational, but if one
of them is gone (i.e. I shut it down), all queries that are sent to the
MySQL API nodes seem to hang for about 6.60sec before they are actually
executed. As soon as the dead DB node becomes available again, everything
starts to work as it supposed to.
If one of DB nodes is dead:
mysql> SELECT * FROM t; +----+-------+ | id | name | +----+-------+ | 2 | test2 | | 1 | test1 | +----+-------+ 2 rows in set (6.60 sec)
with all DB nodes working:
mysql> SELECT * FROM t; +----+-------+ | id | name | +----+-------+ | 2 | test2 | | 1 | test1 | +----+-------+ 2 rows in set (0.00 sec)
It looks like MySQL is waiting for the dead node to respond, gets timed out
after about 6 seconds and then requests the answer from the other node. I
did not find anything in the Administrator Guide that would say about such
behaviour, which makes the cluster rather useless in case of a node crash.
Any ideas how to fix it?
Regards, Maciek
-- MySQL General Mailing List For list archives: http://lists.mysql.com/mysql To unsubscribe: http://lists.mysql.com/[EMAIL PROTECTED]
Mikael Ronström, Senior Software Architect MySQL AB, www.mysql.com
Clustering: http://www.infoworld.com/article/04/04/14/HNmysqlcluster_1.html
http://www.eweek.com/article2/0,1759,1567546,00.asp
-- MySQL General Mailing List For list archives: http://lists.mysql.com/mysql To unsubscribe: http://lists.mysql.com/[EMAIL PROTECTED]
