Hello,
I have a master/slave replication set up. There is only a small amount of traffic on both the server and the slave (tens of transactions per day). The slave is basically set up so that it is read only, to make reads faster for users on that part of the network.
The network in this part is a little unreliable. It is a remote area on a large mine site, and so it is connected via a wireless network. Every now and then the network loses a few packets or loses its connection completely, but it sorts itself out again within 30 seconds.
The problem is that whenever this happens (and it happens more than 10 times a day) the slave replication sometimes stops working. Any updates on the master are not reflected on the slave. I would be OK with this if some kind of error message came up on the slave, but it just sits there saying "Waiting for master event", no matter how many updates on the master occur. The only way to fix the problem is to stop the slave and then start it again. Then it quickly catches up to the right position in the log file and everything is good again.
Does anyone have a workaround for this? Or is it some kind of setting I am missing? I know this network is a little more unreliable than most, but you can never guarantee that a packet is going to reach its destination on any network. In my mind this shouldn't be breaking replication at all.
I have a master/slave replication set up. There is only a small amount of traffic on both the server and the slave (tens of transactions per day). The slave is basically set up so that it is read only, to make reads faster for users on that part of the network.
The network in this part is a little unreliable. It is a remote area on a large mine site, and so it is connected via a wireless network. Every now and then the network loses a few packets or loses its connection completely, but it sorts itself out again within 30 seconds.
The problem is that whenever this happens (and it happens more than 10 times a day) the slave replication sometimes stops working. Any updates on the master are not reflected on the slave. I would be OK with this if some kind of error message came up on the slave, but it just sits there saying "Waiting for master event", no matter how many updates on the master occur. The only way to fix the problem is to stop the slave and then start it again. Then it quickly catches up to the right position in the log file and everything is good again.
Does anyone have a workaround for this? Or is it some kind of setting I am missing? I know this network is a little more unreliable than most, but you can never guarantee that a packet is going to reach its destination on any network. In my mind this shouldn't be breaking replication at all.