[Bucardo-general] Long, predictable delay in synchronization

[Paul Theodoropoulos]

On 9/27/13 11:43 AM, Paul Theodoropoulos wrote:
> On 9/26/13 8:16 PM, Paul Theodoropoulos wrote:
>>
>> On 9/26/2013 20:13, Greg Sabino Mullane wrote:
>>>> I just ran an strace on the Bucardo VAC process, and it appears that
>>>> it's ignoring the vac configurations. Is it expected that updates
>>>> will propogate only after vac runs?
>>> No, they should be independent of anything else. What made you think 
>>> it is
>>> ignoring the config?
>> The logs only show sync results occurring after the vac runs. I set 
>> my vac run to 5, and vac sleep to 15, but the logs still showed the 
>> consistent 30 second intervals between any updates to syncs.
>>
> "bucardo reload config" is helpful to do before reporting that the 
> application is ignoring one's configuration changes. :)
>
Throwing another datapoint out there, I not infrequently get this sort 
of behavior:

postgres at MIA-BUCARDO-tru-main-dev: ~ $ bucardo reload config
Forcing Bucardo to reload the bucardo_config table
Waited, but Bucardo never confirmed the configuration reload!
postgres at MIA-BUCARDO-tru-main-dev: ~ $ bucardo set vac_sleep=15
Set "vac_sleep" to "15"
postgres at MIA-BUCARDO-tru-main-dev: ~ $ bucardo reload config
Forcing Bucardo to reload the bucardo_config table
Looks like Bucardo is not running, so there is no need to reload*
*
But bucardo is indeed happily running, /var/run/bucardo/bucardo.mcp.pid 
has the correct pid in it, and there's no fullstopbucardo anywhere. Not 
sure why this happens.

-- 
Paul Theodoropoulos
www.anastrophe.com

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mail.endcrypt.com/pipermail/bucardo-general/attachments/20130927/d4cd578f/attachment.html>