Re: [Maria-discuss] WL#4925 and multi-source replication

2013-03-15 Thread Greg
2013/3/15 Kristian Nielsen > How many transactions do you have running in parallel? Binlog group commit > only helps if you have more than 2 transactions writing concurrently (and > the > more, the better). > I didn't add write queries to this master, so I only have the 4 sources (I removed 2 so

Re: [Maria-discuss] WL#4925 and multi-source replication

2013-03-15 Thread Kristian Nielsen
Greg writes: > This morning I've removed DRBD to verify that it is not the bottleneck, and > It's much faster with DRBD ! Yes, because with DRBD, an fdatasync() only needs to do a network trip to the other node. On local disk, it needs a physical disk write, which can take 10 milliseconds (

Re: [Maria-discuss] WL#4925 and multi-source replication

2013-03-15 Thread Henrik Ingo
On Wed, Mar 13, 2013 at 4:34 PM, Greg wrote: > Hi, > > I'm testing multi-source replication with 6 sources, it works great ! > Now I'm at the "HA" state, testing a master multi-source with DRBD. To do a > safe switch between masters, I have to enable sync_binlog=1 and then the > master delay grow.

Re: [Maria-discuss] WL#4925 and multi-source replication

2013-03-15 Thread Greg
2013/3/15 Kristian Nielsen > Isn't it? I do not have practical experience with MariaDB/MySQL in a HA > setup, > but I have always thought that DRBD was one of the best choices, being a > simple and robust solution. > Yoshinori Matsunobu described available solutions with pros and cons here : htt

Re: [Maria-discuss] WL#4925 and multi-source replication

2013-03-15 Thread Kristian Nielsen
Greg writes: > I think you're right. I know that DRBD is not the best choice to have > high-availibility on a MariaDB master, but I have my reasons to make this Isn't it? I do not have practical experience with MariaDB/MySQL in a HA setup, but I have always thought that DRBD was one of the best

Re: [Maria-discuss] WL#4925 and multi-source replication

2013-03-15 Thread Greg
2013/3/15 Kristian Nielsen > > Right. So this is mostly just my own academic interest, in practice it is > of > course real crashes/powerfailures we want to handle, not SIGKILL. > > If you are interested, this is my thinking: the server always does a > write(2) > system call on the binlog at (gro

Re: [Maria-discuss] WL#4925 and multi-source replication

2013-03-15 Thread Kristian Nielsen
Greg writes: > With "kill -9 mysqld", and sync_binlog=0, I'm not really surprised since > mysql will not fdatasync after each commit, right ? Right. So this is mostly just my own academic interest, in practice it is of course real crashes/powerfailures we want to handle, not SIGKILL. If you are