On 10-11-15 12:29 PM, Aleksey Tsalolikhin wrote:
> On Mon, Nov 15, 2010 at 3:19 AM, Aleksey Tsalolikhin
> <atsaloli.tech at gmail.com>  wrote:
>> On Mon, Nov 15, 2010 at 1:27 AM, Aleksey Tsalolikhin
>> <atsaloli.tech at gmail.com>  wrote:
>>>
>>> we are going to take an emergency maintenance window to upgrade from
>>> slony-I 1.2.20 to 1.2.21 and resync, see how it looks then.
>
> Lag is over an hour now.  Here is a quick snapshot of the logs:
>
>
> Slony master:
>
>
> 2010-11-15 09:24:50 PSTDEBUG2 localListenThread: Received event 1,12426 SYNC
> 2010-11-15 09:24:52 PSTDEBUG2 syncThread: new sl_action_seq 22626110 -
> SYNC 12427
> 2010-11-15 09:24:52 PSTDEBUG2 localListenThread: Received event 1,12427 SYNC
> 2010-11-15 09:24:53 PSTDEBUG2 remoteListenThread_2: LISTEN
> 2010-11-15 09:24:54 PSTDEBUG2 syncThread: new sl_action_seq 22626117 -
> SYNC 12428
> 2010-11-15 09:24:54 PSTDEBUG2 localListenThread: Received event 1,12428 SYNC
>
> Slony slave:
>
>
> 2010-11-15 09:24:53 PSTDEBUG2 syncThread: new sl_action_seq 1 - SYNC 2353
> 2010-11-15 09:24:54 PSTDEBUG2 remoteListenThread_1: queue event 1,12428 SYNC
> 2010-11-15 09:24:55 PSTDEBUG4 remoteWorkerThread_1: returning lines to pool
> 2010-11-15 09:24:55 PSTDEBUG4 remoteHelperThread_1_1: fetch from cursor
> 2010-11-15 09:24:55 PSTDEBUG4 remoteHelperThread_1_1: fetched 100 log rows
> 2010-11-15 09:24:55 PSTDEBUG4 remoteHelperThread_1_1: deliver 10 lines to worker
> 2010-11-15 09:24:55 PSTDEBUG4 remoteHelperThread_1_1: allocate line buffers
> 2010-11-15 09:24:56 PSTDEBUG2 remoteListenThread_1: queue event 1,12429 SYNC
>
>
> What does "SYNC 2353" mean?  The number is off from the other numbers
> (in the 12 thousand  4 hundred range).
>
A sync is identified by two numbers, the origin on which the sync was 
generated and a sequence number for the sync.

1,1  is the first sync generated on node 1
1,2  is the second sync generated on node 1

2,1 is the first sync generated on node 2
2,2 is the second sync generated on node 2.

The slave slon  has a sync thread that generates sync events on the 
slave (your node 2).   The sync thread above is refering to event 
"2,2353" - we aren't printing the 2 but probably should be.

If the slon configuration is the same for both your master and slave 
(your aren't adjusting the sync timing intervals) and you created both 
nodes at about the same time and started the slons at about the same 
time then it sounds like your slave is generating events at a slower 
rate than the master, which isn't a problem per say but might be a 
consequence of your real problem.

It sounds to me like your transaction rate is too high for your setup. 
Either the network can't keep up or the hardware can't keep up.  Based 
on your other post it sounds like the slave is having problems keeping 
up.  Are your master + slave a similar hardware configuration?




> Best,
> -at
> _______________________________________________
> Slony1-general mailing list
> Slony1-general at lists.slony.info
> http://lists.slony.info/mailman/listinfo/slony1-general