MySQL clsuter v7.1.15a
I am doing performancing testing with flexAsynch
It got a lots of below error.
NDB error code 1218: Send Buffers overloaded in NDB kernel: Temporary error: Temporary Resource error
It still got the error with below situations:
1. TotalSendBufferMemory=1024M without setting SendBufferMemory
2. TotalSendBufferMemory=1024M, SendBufferMemory=8M. Seems better result
3. Stop all the SQL nodes
Ny netowkr card has a speed of 1000Mbits/s and the max utilization is just about 220Mbits/s.
someone said:
The "Send Buffers overloaded" error was added specifically for
overloaded binlogging to mysqld. The mysqld receives events
(data changes) from all DB nodes. Either the connections from DB nodes,
or the mysqld, or an attached replication channel cannot keep up.
But it has no effect when I do below things:
1. In my.cnf set ndb-log-bin = 0 and ndb-log-binlog-index = 0
2. Stop all the SQL nodes
The only bottleneck is that I found a cpu core is very busy, 95% utilization while ohers are all idle.
I want to full use the all CPU cores by running MTD, but all instances were forced to down once it got a 1218 error. Is it a bug of MTD mode? is it fixed by 7.1.18 or not?
I am doing performancing testing with flexAsynch
It got a lots of below error.
NDB error code 1218: Send Buffers overloaded in NDB kernel: Temporary error: Temporary Resource error
It still got the error with below situations:
1. TotalSendBufferMemory=1024M without setting SendBufferMemory
2. TotalSendBufferMemory=1024M, SendBufferMemory=8M. Seems better result
3. Stop all the SQL nodes
Ny netowkr card has a speed of 1000Mbits/s and the max utilization is just about 220Mbits/s.
someone said:
The "Send Buffers overloaded" error was added specifically for
overloaded binlogging to mysqld. The mysqld receives events
(data changes) from all DB nodes. Either the connections from DB nodes,
or the mysqld, or an attached replication channel cannot keep up.
But it has no effect when I do below things:
1. In my.cnf set ndb-log-bin = 0 and ndb-log-binlog-index = 0
2. Stop all the SQL nodes
The only bottleneck is that I found a cpu core is very busy, 95% utilization while ohers are all idle.
I want to full use the all CPU cores by running MTD, but all instances were forced to down once it got a 1218 error. Is it a bug of MTD mode? is it fixed by 7.1.18 or not?