How can I check if my data nodes are using the correct config.ini?
I just tried to restart my cluster with changed parameters, but the parameter change didn't seem to have any effect.
This is what I'm trying to do:
I have a SQL file with 100K inserts, each adding between 100 and 1000 rows to a table in my 4 datanode cluster (7.2.4). After 22K inserts (around 700MB) I got the 410 error: 'REDO log files overloaded'.
So I thought: if the buffer is full after 22K inserts, multiplying the buffer size by a factor 5 should be enough for my 100K inserts.
I changed the RedoBuffer parameter in the config.ini from 64M to 320M and I set the NoOfFragmentLogFiles from 50 to 100, just to be sure.
I stopped all data nodes, stopped the ndb_mgmd and restarted it like this:
ndb_mgmd --configdir=C:\my_cluster\conf -f C:\my_cluster\conf\config.ini
Then I restarted my data nodes and tried to perform my 100K inserts again.
I got the same error again at around 22K inserts, so my guess is, that the parameters weren't changed. I don't know what's going wrong. Did I tune the wrong parameters, didn't the management node load the changed config file when restarting or didn't the data nodes get the new config from the management node?
I just tried to restart my cluster with changed parameters, but the parameter change didn't seem to have any effect.
This is what I'm trying to do:
I have a SQL file with 100K inserts, each adding between 100 and 1000 rows to a table in my 4 datanode cluster (7.2.4). After 22K inserts (around 700MB) I got the 410 error: 'REDO log files overloaded'.
So I thought: if the buffer is full after 22K inserts, multiplying the buffer size by a factor 5 should be enough for my 100K inserts.
I changed the RedoBuffer parameter in the config.ini from 64M to 320M and I set the NoOfFragmentLogFiles from 50 to 100, just to be sure.
I stopped all data nodes, stopped the ndb_mgmd and restarted it like this:
ndb_mgmd --configdir=C:\my_cluster\conf -f C:\my_cluster\conf\config.ini
Then I restarted my data nodes and tried to perform my 100K inserts again.
I got the same error again at around 22K inserts, so my guess is, that the parameters weren't changed. I don't know what's going wrong. Did I tune the wrong parameters, didn't the management node load the changed config file when restarting or didn't the data nodes get the new config from the management node?