[robvas]

'peterlaursen' wrote:

It looks to me like your are using the same local port for SSH tunnel to two different servers. This is not possible. You will eventually try to 'sync a server with itself'.

You guys are right, I was trying to sync the database with itself. I suspected this was the problem last week, but when I did 'show databases' on the test server, it showed 19, and then connect to the mirror server through the ssh tunnel, it showed 41. I have it working now. The problem was either in the creation of SSH tunnel or not connecting through it properly. This is the line I was using to connect:

ssh -fN mysql-prod-mirror -L 3309:127.0.0.1:3306

This is the new line I connect with:

ssh -fN -L 3309:127.0.0.1:3306 deploy@mysql-prod-mirror

I then connect to mysql using this to test:

mysql -h 127.0.0.1 -P 3309 -u deploy -p

The problem was, before I was not using the -h option to specify the host. I didn't think I needed to, I thought it would connect by default.

sja appears to be running as it should now:

Sync started at Tue Jul 19 11:21:35 2011

Table SrcRows TgtRows Inserted Updated Deleted

========================= ======= ======= ======== ======= =======

`activities` 0 0 0 0 0

`addresses` 8632 8334 298 0 0

`award_eligibilities` 0 0 0 0 0

`awards` 0 0 0 0 0

`competitions` 28 28 0 0 0

`contacts` 8640 8342 298 2 0

`custom_prompts` 38 36 3 0 1

`destinations` 0 0 0 0 0

`edge_points` 925 925 0 0 0

`eligibilities` 66805 66805 0 277 0

`entries` 401825 389598 12227 1523 0

`evaluations` 1228 1226 2 0 0

Thanks for the help, dumb mistake on my part. I ended up changing the XML file from 'localhost' to '127.0.0.1'

[robvas]

XML file I am using:

localhost

deploy

secret

3309

0

0

cc

mysql-prod-test

deploy

secret

3306

0

0

cc

(we use ssh tunnel from target machine to source machine, sja is executed on target machine)

[robvas]

Source DB:

Before lock state: http://pastebin.us/1602

During lock state: http://pastebin.us/1601

Target DB:

Before lock state: http://pastebin.us/1604

During lock state: http://pastebin.us/1603

[robvas]

Thanks for the fast reply.

[robvas]

'peterlaursen' wrote:

I believe this is not a a problem wiht hardware. It looks like there is some other application that has LOCKED a lot of tables because of open transactions. Maybe there is a backup job or similar running? Or some app (or user!) that just START TRANSACTION and does not commit for very long time. Or execute SELECT * FOR UPDATE? Try to get details from SHOW FULL PROCESSLIST and SHOW ENGINE INNODB STATUS about what is happening.

The server I am using for a source is replicating from a server on the internet (as a slave). I stopped replication and still get the error.

Here is the output you requested from the target server. The source didn't show anything interesting.

http://pastebin.us/643

'peterlaursen' wrote:

Maybe (?) I could also suggest to try restart the server (just in case).

I restarted both and had no effect.

'peterlaursen' wrote:

Also tell: Are you using SJA for the first time. Or are you an old user that have been using it for long and only now this happens?

This is the first time i have used SJA

[Author]

Posts