I am running two MySQL servers (5.0.92-community-log). One (host name gw8) is in a SingleHop data center, the other (host name db1) is an Amazon EC2 instance. They are configured for circular replication - gw8 is master to db1, db1 is master to gw8. Both hosts use MyISAM.
They have been running flawlessly for about 3 months. This last Friday we started noticing weird errors in our application, and investigation showed that we were having replication issues. When I insert a row on db1, it immediately shows up on gw8. When I insert a row on gw8 it takes anywhere from 5 minutes to 8 hours to show up on db1.
When I "show full processlist" on both hosts I don't see any queries running for inordinately long times. The slow-query log shows just the normal assortment of occasional non-indexed SELECTs.
Neither machine is showing unusual CPU loading, neither is showing signs of network issues. I can "mysql -h gw8" from db1 with no noticeable delay, and can "mysql -h db1" from gw8 with no noticeable delay.
Suggestions on where I should be looking?
They have been running flawlessly for about 3 months. This last Friday we started noticing weird errors in our application, and investigation showed that we were having replication issues. When I insert a row on db1, it immediately shows up on gw8. When I insert a row on gw8 it takes anywhere from 5 minutes to 8 hours to show up on db1.
When I "show full processlist" on both hosts I don't see any queries running for inordinately long times. The slow-query log shows just the normal assortment of occasional non-indexed SELECTs.
Neither machine is showing unusual CPU loading, neither is showing signs of network issues. I can "mysql -h gw8" from db1 with no noticeable delay, and can "mysql -h db1" from gw8 with no noticeable delay.
Suggestions on where I should be looking?