Dear all,

I found that slave replication is slow or in other words slave is back from master.

in my case master have total records
+----------+
| count(*) |
+----------+
| 52433939 |
+----------+

where as slave is on

+----------+
| count(*) |
+----------+
| 50397367 |
+----------+

the total difference is 2036572.

my question is how do i force the slave or master to replicate data.

i have dedicated lan 1000Mbps for replication between master and slave, and master updates after a min.

regards.

I don't think that this has much to do with bandwidth as long as you don't do much more queries on your slave than on your master.

First thing to check is that the slave process is not hanging due to some error by

show slave status;

on your slave machine.

Other reasons might be that you are doing queries on your slave that lock the affected tables for a long time so that the slave is not able to process the updates in time.

- check and optimize your tables, for some reason or another there might be corrupt index files.

- check your buffer and io settings, if these are different from your master server the slave might perform slower in general (not that they have to be the same, but might not be optimized)

Andreas

thanks Andreas,

I will optimize master dB server, your suggestions will help me to do more best in this regard.

my single table count increases rapidly, you can see my last post counts were

+----------+
| count(*) |
+----------+
| 52433939 |
+----------+
and current count on master DB are
+----------+
| count(*) |
+----------+
| 54210071 |
+----------+

I am using MySQL enterprise 5.0. can you help me, how do i get performance on this DB. should i use "Archive storage engine" on slave and "MOVE" all records till previous month to "archive_table" and leave the current month records.

or

i change the DB to MySQL 5.1 to use partitioning , only on slave?

or
DB should be changed on both Master and slave and other than MySQL e.g Oracle.................etc.

thanks

Hi Ammad,

trying to run different tabletypes on master and slave
servers will very likely run you into trouble as replication
is ment to give exact copy of the master server besides the
delay. So different features or restrictions of the table type
could cause failure on the replication server then replaying the
sql-log.

As I don't no very much about your table structure and row size I'm
afraid I cannot give much more hints besides that you could also think
about putting the table into a dedicated database and mount the
database directory on a dedicated harddisk if IO performance is
the bottleneck.

Andreas