Suddenly your application starts throwing "error 157" and
performance degrades or is non-existing. It is easy to panic then
and try all sorts of actions to get past the problem. We have
seen several users doing:
- rolling restart
- stop cluster / start cluster
because they also see this in the error logs:
120828 13:15:11 [Warning] NDB: Could not acquire global schema lock (4009)Cluster Failure
That is not a really a nice error message. To begin with, it is a
WARNING when something is obviously wrong. IMHO, it should be
CRITICAL. Secondly, the message ‘Cluster Failure’ is misleading.
The cluster may not really have failed, so there is no
point trying to restart it before we know more.
So what does error 157 mean and what can we do about it?
By using perror we can get a hint what it means:
$ …