postgres-xc-general

[Postgres-xc-general] replication problem

[Marcin K. <mr...@gm...>]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Hello everyone,

I have set up XC on two nodes according to detailed installation
instructions in installation manual and it seemed to be working at
first, but the changes to my 2nd node do not seem to be replicated. I'm
getting warnings in psql:

STATEMENT:  create database testd2;
WARNING:  Do not have a GTM snapshot available
CREATE DATABASE

I'm getting errors in GTM log:

157:1092475216:2011-01-04 19:58:35.646 CET -LOG:  Sending transaction id 680
LOCATION:  ProcessBeginTransactionGetGXIDCommand, gtm_txn.c:1088
158:1092475216:2011-01-04 19:58:35.668 CET -LOG:  unexpected EOF on
client connection
LOCATION:  ReadCommand, main.c:870
159:1092475216:2011-01-04 19:58:35.668 CET -LOG:  Cleaning up thread state
LOCATION:  GTM_ThreadCleanup, gtm_thread.c:265

I have come across this:

http://sourceforge.net/mailarchive/forum.php?thread_name=4CB71016.2020301%40enterprisedb.com&forum_name=postgres-xc-general

"I think this is related to what Koichi pointed out, that you have to
resync the other coordinators."

?? How do I do that, that is, resync the coordinators?

" You may just want to use the latest code
in the git repository so that you do not need to worry about it."

Ok, but is the code in git repo stable enough for production use?



- --

Regards,
mk

- --
Premature optimization is the root of all fun.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.9 (MingW32)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/

iQEcBAEBAgAGBQJNI3zRAAoJEFMgHzhQQ7hOYRwH/A249GOF7ywtL6Krg5Chjrfi
HQYh+tqa/or8qZJ6/KyrZSArROe3YYe+R6EM6vwpMhr/1HT1Uu0mEoNivGgenDIk
rINoqdxTpENsi3NQFo4sb61cqJdYyBTb3kThO0SOFRy1ov3/rV7opOER6Sr+q0JE
5AZxpCHZ/AB+g6Hq9kiMS5HOpBrnx2EVCsiIdImTVfxLUL294YP9k1BuxWiZDRxr
Y06uxIK/IFeF5RjPLTXPaTMK/wcsL7laQsidPvHN8H8LF9ANn3OMHbcFJJhRPmby
iCrgt8uJK4RFC4DFLM7N7wBeOZdcsV8E2DiQF/wy74ojfjbVD2n5npG93zEAHBw=
=UAyW
-----END PGP SIGNATURE-----

[Postgres-xc-general] replication problem

[Marcin K. <mr...@gm...>]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Hello everyone,

P.S. I have NOT set up GTM proxy. Do I have to do that? Is it related to
my problem?


- --

Regards,
mk

- --
Premature optimization is the root of all fun.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.9 (MingW32)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/

iQEcBAEBAgAGBQJNI32UAAoJEFMgHzhQQ7hOOloIAIEeNHDjM2cMzRK7R4yUPN77
VxvoxZPO8tiRKh7nkQtBwzLVXp2myYtfGc+oiKG2FdlGXZaF7V61RE8xWwGIEYDd
GDx0Xo4dYHAtIXLI+fR0sLGnrnPy1sQyybYTR/PDGklxPb+lrTV9emUjcgW/ArGH
lvqhwckW8GpGgec5PWS6WMKK6WCPYfs0aEc9mmxRHVgLkrZrpRX8QfA/nRohlmDO
ZzEcsptMjhKjsTF6zyleDWArxSziPvfuD13HVey4FdnP44o2AKDs3Hew1WC7S3vC
kVpWJtWKbT65hQ60/U0rAKxeqy208/aK+MaQaQAfaXNsXBgFBCyW+Hx4MB820ps=
=Ws44
-----END PGP SIGNATURE-----

Re: [Postgres-xc-general] replication problem

[Michael P. <mic...@gm...>]

For the answers to your questions... please see inline.

About CREATE DATABASE, there is since version 0.9.3 all he necessary
mechanisms to synchronize automatically Coordinators (well, DDL...)
through the whole cluster. So it is not necessary to resynchronize the
catalog files of Coordinators manually.

If you are getting snapshots warning when creating a Database, you may have
configuration problems.
You should double check your configuration files.

> I have NOT set up GTM proxy. Do I have to do that? Is it related to my
problem?
It is not necessary to set up a GTM Proxy to have your cluster working
correctly.

> Ok, but is the code in git repo stable enough for production use?
Code is not stable enough to be used in production, we are currently working
a lot on the code stabilization, SQL support and HA features.

-- 
Michael Paquier
http://michaelpq.users.sourceforge.net