Menu
▾
▴
postgres-xc-general
|
From: Juned K. <jkh...@gm...> - 2014-04-04 10:15:21
|
Hi all, I am using pgxc which has two datanodes (node1 and node2). As of now my application is connected to node1, yesterday due to some problem ssh service just stopped on second server. when i was trying to run any query on node1 then it gave me "ERROR: Failed to get pooled connection" . I was just wondering node2 is went down and my application is connected to node1 then why this problem? how does it relates to that error. is it normal ?? Please suggest. -- Thanks, Juned Khan iNextrix Technologies Pvt Ltd. www.inextrix.com |
|
From: 鈴木 幸市 <ko...@in...> - 2014-04-04 10:50:51
|
Postgres-XC assumes all the nodes are running normally so that the planner can build a plan regarding clause push-down, table distribution and replication. The error indicates some of such nodes are down. Please restart the failed one and issue ‘select pgxc_pool_reload()’. --- Koichi Suzuki 2014/04/04 19:15、Juned Khan <jkh...@gm...<mailto:jkh...@gm...>> のメール： Hi all, I am using pgxc which has two datanodes (node1 and node2). As of now my application is connected to node1, yesterday due to some problem ssh service just stopped on second server. when i was trying to run any query on node1 then it gave me "ERROR: Failed to get pooled connection" . I was just wondering node2 is went down and my application is connected to node1 then why this problem? how does it relates to that error. is it normal ?? Please suggest. -- Thanks, Juned Khan iNextrix Technologies Pvt Ltd. www.inextrix.com<http://www.inextrix.com/> ------------------------------------------------------------------------------ _______________________________________________ Postgres-xc-general mailing list Pos...@li...<mailto:Pos...@li...> https://lists.sourceforge.net/lists/listinfo/postgres-xc-general |
|
From: Juned K. <jkh...@gm...> - 2014-04-04 11:12:37
|
Hi koichi, so it means if i have 10 clusters and if any of them goes down then whole system will be useless what's the solution of this problem ? How to failover ? On Fri, Apr 4, 2014 at 4:20 PM, 鈴木 幸市 <ko...@in...> wrote: > Postgres-XC assumes all the nodes are running normally so that the > planner can build a plan regarding clause push-down, table distribution and > replication. > > The error indicates some of such nodes are down. > > Please restart the failed one and issue ‘select pgxc_pool_reload()’. > --- > Koichi Suzuki > > 2014/04/04 19:15、Juned Khan <jkh...@gm...> のメール： > > Hi all, > > I am using pgxc which has two datanodes (node1 and node2). > > As of now my application is connected to node1, yesterday due > to some problem ssh service just stopped on second server. > > when i was trying to run any query on node1 then it gave me > "ERROR: Failed to get pooled connection" . > > I was just wondering node2 is went down and my application is connected > to node1 then why this problem? how does it relates to that error. > > is it normal ?? > > Please suggest. > > > -- > Thanks, > Juned Khan > iNextrix Technologies Pvt Ltd. > www.inextrix.com > > ------------------------------------------------------------------------------ > _______________________________________________ > Postgres-xc-general mailing list > Pos...@li... > https://lists.sourceforge.net/lists/listinfo/postgres-xc-general > > > -- Thanks, Juned Khan iNextrix Technologies Pvt Ltd. www.inextrix.com |
|
From: Koichi S. <koi...@gm...> - 2014-04-05 02:19:54
|
Each node can have its own slave. Example configuration for pgxc_ctl will be found at http://sourceforge.net/apps/mediawiki/postgres-xc/index.php?title=PGOpen2013_Postgres_Open_2013 You will also find example script of the demonstration I used at PG Open 2013. Regards; --- Koichi Suzuki 2014-04-04 20:12 GMT+09:00 Juned Khan <jkh...@gm...>: > Hi koichi, > > so it means if i have 10 clusters and if any of them goes down then whole > system will be useless > > what's the solution of this problem ? How to failover ? > > > On Fri, Apr 4, 2014 at 4:20 PM, 鈴木 幸市 <ko...@in...> wrote: >> >> Postgres-XC assumes all the nodes are running normally so that the planner >> can build a plan regarding clause push-down, table distribution and >> replication. >> >> The error indicates some of such nodes are down. >> >> Please restart the failed one and issue ‘select pgxc_pool_reload()’. >> --- >> Koichi Suzuki >> >> 2014/04/04 19:15、Juned Khan <jkh...@gm...> のメール： >> >> Hi all, >> >> I am using pgxc which has two datanodes (node1 and node2). >> >> As of now my application is connected to node1, yesterday due >> to some problem ssh service just stopped on second server. >> >> when i was trying to run any query on node1 then it gave me >> "ERROR: Failed to get pooled connection" . >> >> I was just wondering node2 is went down and my application is connected >> to node1 then why this problem? how does it relates to that error. >> >> is it normal ?? >> >> Please suggest. >> >> >> -- >> Thanks, >> Juned Khan >> iNextrix Technologies Pvt Ltd. >> www.inextrix.com >> >> ------------------------------------------------------------------------------ >> _______________________________________________ >> Postgres-xc-general mailing list >> Pos...@li... >> https://lists.sourceforge.net/lists/listinfo/postgres-xc-general >> >> > > > > -- > Thanks, > Juned Khan > iNextrix Technologies Pvt Ltd. > www.inextrix.com > > ------------------------------------------------------------------------------ > > _______________________________________________ > Postgres-xc-general mailing list > Pos...@li... > https://lists.sourceforge.net/lists/listinfo/postgres-xc-general > |
|
From: Juned K. <jkh...@gm...> - 2014-04-07 05:19:08
|
Hi koichi, Yeah we can have slave of each component but in case of failure i need to promote it manually right ? is there any way to make this process automatic ? On Sat, Apr 5, 2014 at 7:49 AM, Koichi Suzuki <koi...@gm...> wrote: > Each node can have its own slave. Example configuration for > pgxc_ctl will be found at > > http://sourceforge.net/apps/mediawiki/postgres-xc/index.php?title=PGOpen2013_Postgres_Open_2013 > > You will also find example script of the demonstration I used at PG Open > 2013. > > Regards; > --- > Koichi Suzuki > > > 2014-04-04 20:12 GMT+09:00 Juned Khan <jkh...@gm...>: > > Hi koichi, > > > > so it means if i have 10 clusters and if any of them goes down then whole > > system will be useless > > > > what's the solution of this problem ? How to failover ? > > > > > > On Fri, Apr 4, 2014 at 4:20 PM, 鈴木 幸市 <ko...@in...> wrote: > >> > >> Postgres-XC assumes all the nodes are running normally so that the > planner > >> can build a plan regarding clause push-down, table distribution and > >> replication. > >> > >> The error indicates some of such nodes are down. > >> > >> Please restart the failed one and issue ‘select pgxc_pool_reload()’. > >> --- > >> Koichi Suzuki > >> > >> 2014/04/04 19:15、Juned Khan <jkh...@gm...> のメール： > >> > >> Hi all, > >> > >> I am using pgxc which has two datanodes (node1 and node2). > >> > >> As of now my application is connected to node1, yesterday due > >> to some problem ssh service just stopped on second server. > >> > >> when i was trying to run any query on node1 then it gave me > >> "ERROR: Failed to get pooled connection" . > >> > >> I was just wondering node2 is went down and my application is connected > >> to node1 then why this problem? how does it relates to that error. > >> > >> is it normal ?? > >> > >> Please suggest. > >> > >> > >> -- > >> Thanks, > >> Juned Khan > >> iNextrix Technologies Pvt Ltd. > >> www.inextrix.com > >> > >> > ------------------------------------------------------------------------------ > >> _______________________________________________ > >> Postgres-xc-general mailing list > >> Pos...@li... > >> https://lists.sourceforge.net/lists/listinfo/postgres-xc-general > >> > >> > > > > > > > > -- > > Thanks, > > Juned Khan > > iNextrix Technologies Pvt Ltd. > > www.inextrix.com > > > > > ------------------------------------------------------------------------------ > > > > _______________________________________________ > > Postgres-xc-general mailing list > > Pos...@li... > > https://lists.sourceforge.net/lists/listinfo/postgres-xc-general > > > -- Thanks, Juned Khan iNextrix Technologies Pvt Ltd. www.inextrix.com |
|
From: 鈴木 幸市 <ko...@in...> - 2014-04-07 05:52:18
|
Because monitoring feature is not a part of Postgres-XC yet, you need to promote it manually. You will fold failover command in pgxc_ctl to help this operation. The background that automatic failover is not a part of XC is: We need to monitor many other failures such as server itself, storage, network among other software. It is more than Postgres-XC can do by itself. Of course, you can invoke pgxc_ctl command to promote a slave as a part of your resource agent of an automatic failover software such as Pacemaker. I hope Postgres-XC can have automatic failover system which takes care of XC components only and take other monitoring result from external middleware. Regards; --- Koichi Suzuki 2014/04/07 14:19、Juned Khan <jkh...@gm...<mailto:jkh...@gm...>> のメール： Hi koichi, Yeah we can have slave of each component but in case of failure i need to promote it manually right ? is there any way to make this process automatic ? On Sat, Apr 5, 2014 at 7:49 AM, Koichi Suzuki <koi...@gm...<mailto:koi...@gm...>> wrote: Each node can have its own slave. Example configuration for pgxc_ctl will be found at http://sourceforge.net/apps/mediawiki/postgres-xc/index.php?title=PGOpen2013_Postgres_Open_2013 You will also find example script of the demonstration I used at PG Open 2013. Regards; --- Koichi Suzuki 2014-04-04 20:12 GMT+09:00 Juned Khan <jkh...@gm...<mailto:jkh...@gm...>>: > Hi koichi, > > so it means if i have 10 clusters and if any of them goes down then whole > system will be useless > > what's the solution of this problem ? How to failover ? > > > On Fri, Apr 4, 2014 at 4:20 PM, 鈴木 幸市 <ko...@in...<mailto:ko...@in...>> wrote: >> >> Postgres-XC assumes all the nodes are running normally so that the planner >> can build a plan regarding clause push-down, table distribution and >> replication. >> >> The error indicates some of such nodes are down. >> >> Please restart the failed one and issue ‘select pgxc_pool_reload()’. >> --- >> Koichi Suzuki >> >> 2014/04/04 19:15、Juned Khan <jkh...@gm...<mailto:jkh...@gm...>> のメール： >> >> Hi all, >> >> I am using pgxc which has two datanodes (node1 and node2). >> >> As of now my application is connected to node1, yesterday due >> to some problem ssh service just stopped on second server. >> >> when i was trying to run any query on node1 then it gave me >> "ERROR: Failed to get pooled connection" . >> >> I was just wondering node2 is went down and my application is connected >> to node1 then why this problem? how does it relates to that error. >> >> is it normal ?? >> >> Please suggest. >> >> >> -- >> Thanks, >> Juned Khan >> iNextrix Technologies Pvt Ltd. >> www.inextrix.com<http://www.inextrix.com/> >> >> ------------------------------------------------------------------------------ >> _______________________________________________ >> Postgres-xc-general mailing list >> Pos...@li...<mailto:Pos...@li...> >> https://lists.sourceforge.net/lists/listinfo/postgres-xc-general >> >> > > > > -- > Thanks, > Juned Khan > iNextrix Technologies Pvt Ltd. > www.inextrix.com<http://www.inextrix.com/> > > ------------------------------------------------------------------------------ > > _______________________________________________ > Postgres-xc-general mailing list > Pos...@li...<mailto:Pos...@li...> > https://lists.sourceforge.net/lists/listinfo/postgres-xc-general > -- Thanks, Juned Khan iNextrix Technologies Pvt Ltd. www.inextrix.com<http://www.inextrix.com/> |
|
From: Mason S. <ms...@tr...> - 2014-04-07 15:08:00
|
On Mon, Apr 7, 2014 at 1:19 AM, Juned Khan <jkh...@gm...> wrote: > Hi koichi, > > Yeah we can have slave of each component but in case of failure i need to > promote it > manually right ? > > is there any way to make this process automatic ? > > > We use Corosync/Pacemaker. One should be careful though, it takes time; you may encounter false positives with premature failover, or one down component could impact another and also failover. So, I recommend a lot of testing. Also, one should be prepared to fall back and have a strategy for that in place. -- Mason Sharp TransLattice - http://www.translattice.com Distributed and Clustered Database Solutions |
Thanks for helping keep SourceForge clean.
X