postgres-xc-developers Mailing List for Postgres-XC

If any one of the tables is distributed, NodeA must have a subset of such
table raws and they were lost.    You need to restore them.  With there
could be a way to do this manually, although it is tricky and you need to
be very familiar with XC internals.   Unfortunately, XC does not provide
such built-in feature.

Thank you;

---
Koichi Suzuki
https://www.linkedin.com/in/koichidbms


2015-12-15 7:20 GMT+09:00 Timur Allamiyarov <tim...@gm...>:

> Hi guys,
>
> Say I have 3 nodes in my cluster: NodeA (which runs GTM, GTM_proxy, 1
> datanode and 1 coordinator), NodeB and Node C, each running GTM_Proxy, 1
> datanode and 1 coordinator, and
> GTM Standby.
>
> For some reason main NodeA goes down and becomes physically unavailable
> (someone shuts it down by mistake or something).
>
> Is there any way for nodes B and C to negotiate of who's going to become
> the main node - promote GTM standy and restart GTM_Proxy?
> Does Postgres-XC have any consensus algorithm implemented that could help
> solve this issue?
>
> Thank you.
> -Tim
>
>
>
>
> ------------------------------------------------------------------------------
>
> _______________________________________________
> Postgres-xc-developers mailing list
> Pos...@li...
> https://lists.sourceforge.net/lists/listinfo/postgres-xc-developers
>
>

Hi guys,

Say I have 3 nodes in my cluster: NodeA (which runs GTM, GTM_proxy, 1
datanode and 1 coordinator), NodeB and Node C, each running GTM_Proxy, 1
datanode and 1 coordinator, and
GTM Standby.

For some reason main NodeA goes down and becomes physically unavailable
(someone shuts it down by mistake or something).

Is there any way for nodes B and C to negotiate of who's going to become
the main node - promote GTM standy and restart GTM_Proxy?
Does Postgres-XC have any consensus algorithm implemented that could help
solve this issue?

Thank you.
-Tim

What symptom did you have?   Didn't DML work against tables?

Could you have table's grant list with \z command in psql?   If you find
that the table is not read only, could you try to create new table for the
test?    Another information will be found in $PGDATA/postgresql.conf at
each node.   If you have rebooted your machine several time, each node
should have done crash-recovery process and you may find what happened in
that log at each node, including both datanode and coordinator.

GTM and GTM_Proxy does not need any recovery actions so you may not find
useful information in these logs.

BTY, current head of Postgres-XC REL1_2_STABLE contains many more fixes for
stabilization.   I'm about to publish release 1.2.2.   All the code is
ready and now I'm preparing release note and merging document updates from
PostgreSQL minor releases.    It is at
https://github.com/postgres-x2/postgres-x2.

Thank you very much and good luck.
---
Koichi Suzuki



---
Koichi Suzuki
https://www.linkedin.com/in/koichidbms


2015-12-03 14:18 GMT-08:00 Timur Allamiyarov <tim...@gm...>:

> Good day,
>
> I have recently switched to postgresxc v1.2.1, and I find it great. In my
> test scenario, I have created GTM, GTM Proxy, Coordinator and Datanode all
> running on one server. I then created a new database:
>
> # create database test_db encoding='UFT8';
>
> +several simple tables. All works like a charm.
>
> However, what I discovered is that for some reason when I re-booted my
> server several times (trying to imitate some abnormal conditions) all my
> tables in that particular database (test_db) had become read-only.
>
> Initially I though - may be postgres is in the recovery mode, so I ran:
>
> postgres=# SELECT pg_is_in_recovery();
> pg_is_in_recovery
> -------------------
> f
> (1 row)
>
> I then executed 'SET transaction_read_only = off' and it didn't help
> either.
>
> Any idea what could cause that problem? What will be the solution to fix
> it should I face with that issue again?
> Thank you.
>
> -Timur
>
>
> ------------------------------------------------------------------------------
> Go from Idea to Many App Stores Faster with Intel(R) XDK
> Give your users amazing mobile app experiences with Intel(R) XDK.
> Use one codebase in this all-in-one HTML5 development environment.
> Design, debug & build mobile apps & 2D/3D high-impact games for multiple
> OSs.
> http://pubads.g.doubleclick.net/gampad/clk?id=254741911&iu=/4140
> _______________________________________________
> Postgres-xc-developers mailing list
> Pos...@li...
> https://lists.sourceforge.net/lists/listinfo/postgres-xc-developers
>
>

Good day,

I have recently switched to postgresxc v1.2.1, and I find it great. In my
test scenario, I have created GTM, GTM Proxy, Coordinator and Datanode all
running on one server. I then created a new database:

# create database test_db encoding='UFT8';

+several simple tables. All works like a charm.

However, what I discovered is that for some reason when I re-booted my
server several times (trying to imitate some abnormal conditions) all my
tables in that particular database (test_db) had become read-only.

Initially I though - may be postgres is in the recovery mode, so I ran:

postgres=# SELECT pg_is_in_recovery();
pg_is_in_recovery
-------------------
f
(1 row)

I then executed 'SET transaction_read_only = off' and it didn't help
either.

Any idea what could cause that problem? What will be the solution to fix it
should I face with that issue again?
Thank you.

-Timur