headers
Tory M Blue | 2 May 2008 17:46
Picon

Re: 1.2.14rc still does not appear to handle switchover cleanly

On Fri, May 2, 2008 at 5:58 AM, Charles Duffy
<charles.duffy@...> wrote:
> >  > I'd have expected such a patch would have been applied to the 1.2.14rc tarball? I guess I don''t
understand something about the release procedures etc.
>  >
>  >  I think so too.. the original patch which I sent was not applied correctly,
>  >  I hope the release 1.2.14 or 1.2.14rc2.
>  >
>
>  Oh, just noticed that it was committed:
>
>  http://lists.slony.info/pipermail/slony1-commit/2008-March/002214.html
>
>  Didn't make it into the tarball...

Okay going to patch the 1.2.14rc release and do some further testing
thanks
Tory
Permalink | Reply | 1 comment |
headers
Tory M Blue | 2 May 2008 19:02
Picon

Re: 1.2.14rc still does not appear to handle switchover cleanly

On Fri, May 2, 2008 at 8:46 AM, Tory M Blue <tmblue@...> wrote:
> On Fri, May 2, 2008 at 5:58 AM, Charles Duffy
<charles.duffy@...> wrote:
>  > >  > I'd have expected such a patch would have been applied to the 1.2.14rc tarball? I guess I don''t
understand something about the release procedures etc.
>  >  >
>  >  >  I think so too.. the original patch which I sent was not applied correctly,
>  >  >  I hope the release 1.2.14 or 1.2.14rc2.
>  >  >
>  >
>  >  Oh, just noticed that it was committed:
>  >
>  >  http://lists.slony.info/pipermail/slony1-commit/2008-March/002214.html
>  >
>  >  Didn't make it into the tarball...
>
>  Okay going to patch the 1.2.14rc release and do some further testing
>  thanks
>  Tory

So appears this does in fact fix the fail over issue for the most
part.Although I see this on one of the slave nodes after initial
switchover. I do not see a hanging slave node causing issues with
failover nor did I have to rebuild, but I did see this error, each
time I switchover. Have done this test 4 times now (switchover,
switchback, verifying replication between each switch).

any idea? It appears to be the last node each time..

Qslavehost2 (node 4)
(Continue reading)

Permalink | Reply | 0 comments |
headers
Christopher Browne | 2 May 2008 22:28

Re: 1.2.14rc still does not appear to handle switchover cleanly

Glyn Astill <glynastill@...> writes:
> Do we need to apply this patch? We have 3 nodes, 1 origin with 2 subscribers to it.
>
> I'd have expected such a patch would have been applied to the
> 1.2.14rc tarball? I guess I don''t understand something about the
> release procedures etc.

I think what happened is that I made a mistake when applying the
patch, which meant that the "rc" tarball was short a change.

I'll make sure this is in place; will pursue releasing 1.2.14 early
next week.
--

-- 
output = reverse("ofni.secnanifxunil" " <at> " "enworbbc")
http://www3.sympatico.ca/cbbrowne/sap.html
I always try to do things in chronological order.
Permalink | Reply | 0 comments |
headers
Christopher Browne | 2 May 2008 22:36

Re: configure:7134: error: Headers for libpqserver are not found in the includeserverdir

"Giuseppe Laccone" <giuseppe.laccone@...> writes:
> hi all,
> I am trying to compile slony1 with PG8.1.11 for win32 with mingwin32
> and msys as suggested in the
> documentation:
>
> ./configure --with-pgbindir=/usr/local/pgsql/bin
> --with-pgincludedir=/usr/local/pgsql/include
> --with-pgincludeserverdir=/usr/local/pgsql/include/server
> --with-pgconfigdir=/usr/local/pgsql/bin
>
> Unfortunately I  receive always the same error even is I specify
> pgincludeserverdir:

Well, in looking at the config log (thanks for including it!), the
following seems to be suggesting that:

a) Yes, you *do* have the server #includes installed (I can see it
   referencing PostgreSQL server includes, successfully).

   So the problem *isn't* relating to not having properly referenced
   pgincludeserverdir.

b) ERROR being redefined seems troublesome.

It seems as though you have some extra #includes being drawn in that
are causing a conflict.

In particular, you probably ought to see what is in
".../include/wingdi.h", as that is evidently in conflict with
(Continue reading)

Permalink | Reply | 0 comments |
headers
Giuseppe Laccone | 4 May 2008 22:09
Picon

Re: configure:7134: error: Headers for libpqserver are not found in the includeserverdir

Thanks a lot Christopher for your attention!!

  I will try to recompile it again, even if I have specifyed the
  location with --with-pgincludeserverdir=/usr/local/pgsql/include/server,
  it seems to me that here there aren't  some requested server headers.

  Do you know if there is somone who has compiled PG8.1.15 with slony1
  1.1.5 for Windows?

  I need this particular version on the master because the linux client
  is supporting just this version of slony and as I know from the
  documentation it's not possible to synchronize 2 diffenrent slony
  versions. Do you have any suggestion?

  Thanks a lot in advance!

  regards
  Giuseppe

  On Fri, May 2, 2008 at 10:36 PM, Christopher Browne
  <cbbrowne@...> wrote:
  > "Giuseppe Laccone" <giuseppe.laccone@...> writes:
  >  > hi all,
  >  > I am trying to compile slony1 with PG8.1.11 for win32 with mingwin32
  >  > and msys as suggested in the
  >  > documentation:
  >  >
  >  > ./configure --with-pgbindir=/usr/local/pgsql/bin
  >  > --with-pgincludedir=/usr/local/pgsql/include
  >  > --with-pgincludeserverdir=/usr/local/pgsql/include/server
(Continue reading)

Permalink | Reply | 0 comments |
headers
Jakub K | 5 May 2008 12:39
Picon

Beginner question about slony - push or pull replication?

Hello,

i am new in slony and i have question, about how it's works.

Q: slony is pull or push replication? or both?
I set replication using slonik, than start slon daemon on subsrciber (sub) and publisher (pub).
And now, subsrciber is asking for data or publisher is sending data? How often?
-S SYNC check interval means how often are data asking (sub) or sending (pub)

Thanks Jakub.






_______________________________________________
Slony1-general mailing list
Slony1-general@...
http://lists.slony.info/mailman/listinfo/slony1-general
Permalink | Reply | 3 comments |
headers
Christopher Browne | 5 May 2008 16:20

Re: Beginner question about slony - push or pull replication?

"Jakub K" <slonymaillist@...> writes:
> Hello,
> i am new in slony and i have question, about how it's works.
> Q: slony is pull or push replication? or both?

When it combines "publish" and "subscribe," that implies a bit of
both, though it mostly looks like "pull," as that's where the bulk of
the work takes place.

> I set replication using slonik, than start slon daemon on subsrciber
> (sub) and publisher (pub).  And now, subsrciber is asking for data
> or publisher is sending data? How often?  -S SYNC check interval
> means how often are data asking (sub) or sending (pub)

"How often" is essentially controlled by the origin; it is there that
the "check interval" is used to determine if a SYNC event should be
published.

  -> If, within the sync_interval period, replication work has arrived,
     the origin will generate a SYNC event.

  -> If the origin is quiet (e.g. - not receiving any replicable work),
     there will still be a SYNC event each sync_interval_timeout
     milliseconds.

The SYNC event is published via the LISTEN/NOTIFY system in
PostgreSQL, which will cause a quiescent subscriber to wake up and
know it has new work to do.  That's when the "pulling" takes place.
--

-- 
output = reverse("moc.enworbbc" " <at> " "enworbbc")
http://www3.sympatico.ca/cbbrowne/emacs.html
Horribly wedging my very own personal machine gives me a comfortable,
familiar, Lisp-machine feeling (like an old shoe), laced with that racy
hint of raw, cold-boot power.
Permalink | Reply | 2 comments |
headers
Stéphane A. Schildknecht | 5 May 2008 17:38

sl_status for cross replication

Hi,

I put up a replication composed of 2 sets, each one subscribed by different nodes.

(that is I have 5 nodes, 2 sets

First set is replicating that way :
1 -> 2 -> 3 -> 4
     \-> 5

Second set :
4 -> 3 -> 2 -> 1)

First, I wonder which set the information given by sl_status on different nodes
refers to, and how I could get relevant information about each replicated set.

Second, if I have to failover node 1 or node 4, how would I have to proceed?

Would it be enough to fail from 1 to 2 and drop node 1? (It seems indeed that
patched 1.2.14 version works that way, but I'd prefer be sure.)

Best regards,
--

-- 
Stéphane SCHILDKNECHT
Président de PostgreSQLFr
Tél. 09 53 69 97 12
http://www.postgresql.fr
Permalink | Reply | 0 comments |
headers
Glyn Astill | 8 May 2008 13:54
Picon
Favicon

subscribers failing with "child terminated status: 11"

Running slony 1.2.14rc, pg8.3.1 on debian. I've just dropped a set and am tyring to recreate it, Anyone got
any idea whats going off here?

2008-05-08_124040 BST DEBUG4 remoteWorkerThread_3: update provider configuration
2008-05-08_124040 BST DEBUG4 version for "dbname=SEE user=slony host=192.168.240.11" is 80301
2008-05-08_124040 BST DEBUG1 remoteListenThread_3: connected to 'dbname=SEE host=192.168.240.12 user=slony'
2008-05-08_124040 BST DEBUG1 remoteListenThread_1: connected to 'dbname=SEE host=192.168.240.10 user=slony'
2008-05-08_124040 BST DEBUG2 remoteWorkerThread_3: forward confirm 3,189839 received by 1
2008-05-08_124040 BST DEBUG2 remoteWorkerThread_3: forward confirm 1,316706 received by 3
2008-05-08_124040 BST DEBUG2 remoteWorkerThread_3: forward confirm 2,189839 received by 3
2008-05-08_124040 BST DEBUG2 remoteListenThread_1: queue event 1,316707 RESET_CONFIG
2008-05-08_124040 BST DEBUG2 remoteListenThread_1: queue event 1,316708 RESET_CONFIG
2008-05-08_124040 BST DEBUG2 remoteListenThread_1: queue event 1,316709 RESET_CONFIG
2008-05-08_124040 BST DEBUG2 remoteWorkerThread_3: forward confirm 2,189837 received by 1
2008-05-08_124040 BST DEBUG2 remoteWorkerThread_1: Received event 1,316707 RESET_CONFIG
2008-05-08_124040 BST DEBUG2 remoteListenThread_1: queue event 1,316710 DROP_SET
2008-05-08_124040 BST DEBUG2 remoteListenThread_1: queue event 1,316711 STORE_SET
2008-05-08_124040 BST DEBUG2 remoteListenThread_1: queue event 1,316712 SYNC
2008-05-08_124040 BST DEBUG2 remoteListenThread_1: queue event 1,316713 SYNC
2008-05-08_124040 BST DEBUG2 remoteListenThread_1: queue event 1,316714 SYNC
2008-05-08_124040 BST DEBUG2 remoteListenThread_1: queue event 1,316715 SYNC
2008-05-08_124040 BST DEBUG2 remoteListenThread_1: queue event 1,316716 SYNC
2008-05-08_124040 BST DEBUG2 remoteListenThread_1: queue event 1,316717 SYNC
2008-05-08_124040 BST DEBUG2 remoteListenThread_1: queue event 1,316718 SYNC
2008-05-08_124040 BST DEBUG2 remoteListenThread_1: queue event 1,316719 SYNC
2008-05-08_124040 BST DEBUG2 remoteListenThread_1: queue event 1,316720 SYNC
2008-05-08_124040 BST DEBUG2 remoteListenThread_1: queue event 1,316721 SYNC
2008-05-08_124040 BST DEBUG2 remoteListenThread_1: queue event 1,316722 SYNC
2008-05-08_124040 BST DEBUG2 remoteListenThread_1: queue event 1,316723 SYNC
2008-05-08_124040 BST DEBUG2 remoteListenThread_1: queue event 1,316724 SYNC
2008-05-08_124040 BST DEBUG2 remoteListenThread_1: queue event 1,316725 SYNC
2008-05-08_124040 BST DEBUG2 remoteListenThread_1: queue event 1,316726 SYNC
2008-05-08_124040 BST DEBUG2 remoteListenThread_1: queue event 1,316727 SYNC
2008-05-08_124040 BST DEBUG2 remoteListenThread_1: queue event 1,316728 SYNC
2008-05-08_124040 BST DEBUG2 remoteListenThread_1: queue event 1,316729 SYNC
2008-05-08_124040 BST DEBUG2 remoteListenThread_1: queue event 1,316730 SYNC
2008-05-08_124040 BST DEBUG2 remoteListenThread_1: queue event 1,316731 SYNC
2008-05-08_124040 BST DEBUG2 remoteListenThread_1: queue event 1,316732 SYNC
2008-05-08_124040 BST DEBUG2 slon: child terminated status: 11; pid: 23811, current worker pid: 23811
2008-05-08_124040 BST DEBUG1 slon: restart of worker in 10 seconds

      __________________________________________________________
Sent from Yahoo! Mail.
A Smarter Email http://uk.docs.yahoo.com/nowyoucan.html
Permalink | Reply | 0 comments |
headers
postgres | 9 May 2008 05:39

Dropping the master node

hi all.

Just a quick question...

I currently have a master-slave setup with postgres 8.1.10 and slony
1.2.12. I'm needing to drop slony usage on BOTH nodes for a while. I
have experience dropping slave nodes using DROP NODE... but I can't
find any info about its usage on master nodes.

Any advice?

Thanks
Permalink | Reply | 2 comments |

Gmane