On Tuesday 14 February 2006 09:16, Tim Goodaire wrote:
> On Tuesday 14 February 2006 03:14, Robert Treat wrote:
> > 3 node setup. I moved the origin of a set from svr1 to sv3, and now svr2
> > doesn't appear to be updating anymore, even after issuing a subscribe set
> > to specifically point svr2 to svr3.
> >
> > sl_status on all three nodes shows no signifigant lag_time, sl_subscribe
> > shows svr2 as a reciever for svr3 and svr1 with both active. even
> > sl_event seems to have timestamps relativly inline on all three machines.
> >  the thing is that the count(*) on the tables and the values of the
> > sequences are not updating on the svr2 machine, as if there was no
> > updating going on.
> >
> > is there something else i should be looking for?  suppose i could create
> > the setup on svr2 but theres a large table involved that i'd rather not
> > have to resync...  anyone have pointers on jump starting svr2 again?
>
> Have you tried killing your slons and starting them again? I've found some
> strangeness similar to what you describe in the past, and restarting slons
> straightened it out. (That was a long time ago though, in an older version
> of Slony-I) What version of Slony-I are you using?

Thanksfor the tip... you were right in that this was on a 1.0.5 set of slons.  
I killed the slon on svr2 and restarted it and now it is communicating ok; 
which is to say the sequences have synched up, but I see that the row counts 
are off... looks like the svr2 machine lost data during the outage... guess I 
am off to re-build that set :-(

-- 
Robert Treat
Build A Brighter Lamp :: Linux Apache {middleware} PostgreSQL