Got error 127 and 22 when replicating

[Ville Mattila]

Hello,

I have set up a replication between three servers: one master and two slaves that are directly connected to the master server.

For some reason, when the slave begins to get binary information from the master, there appears a few "Got error 127 from table handler" (or 22 sometimes). This requires to stop the slave, fix the tables with myisamchk and then restart again. This must be repeated two or three times per starting a slave until the slave is running properly.

Once slave is started and errors resolved, everything works fine.

I would be happy to get any information why this happens and how this could be avoided?

Thanks,
Ville