[Bucardo-general] Config Problem (WHY CTL Could not add to q sync ?)

[Greg Sabino Mullane]

> Sep 16 13:36:03 garrafa Bucardo[6732]: CTL Timed out - force a sync for
> "pais"
> Sep 16 13:36:03 garrafa Bucardo[6732]: CTL Could not add to q
> sync=pais,source=Master,target=Slave,count=1. Sending manual notification
> 
> WHATS WRONG?
> There is a more detailed log?

This simply means that there is already an entry in the q table for this
sync. In this case, there is already an entry created at 
2008-09-15 18:45:51.623033-03, so the controller sees no need to add
another one. The kid processes are responsible for seeing these entries in
the q table (those with a null 'started' column), doing the actual
replication, and updating the row with first an entry in started, and then
in 'ended' (or 'aborted' if there was a problem).

None of this indicates a problem, unless that row has been sitting around
for a long time, which means the kids processes are not processing it. If
this is still a problem, please send more of the logs, so we can see what
the KID processes are doing.

To recap, the process works like this: when a row changes in a table
Bucardo is replicating, a trigger signals the MCP process. The MCP process
tells the CTL for that particular sync that something has changed. The
controller checks for a non-started row in the q table. If such a row is
not there it adds one. Either way, it sends a signal to the KID to check
the q table. The KID then checks the q table and does the actual copying
of data from one database to the other. It updates the q row to set
'ended' or 'aborted', then waits for another signal from the CTL.


-- 
Greg Sabino Mullane greg at endpoint.com
End Point Corporation
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 163 bytes
Desc: not available
Url : https://mail.endcrypt.com/pipermail/bucardo-general/attachments/20080916/4521b6b2/attachment.bin