Re: [Postgres-xc-general] Problem with taking database dump with pgxc-1.2.1

[Aaron J. <aja...@re...>]

Not a problem, that's why we test these things (and take backups).

Aaron
________________________________
From: 鈴木 幸市 [ko...@in...]
Sent: Tuesday, June 17, 2014 7:45 PM
To: Aaron Jackson
Cc: Postgres-XC mailing list
Subject: Re: [Postgres-xc-general] Problem with taking database dump with pgxc-1.2.1

Hmm,  I’m afraid GTM have falled back local XID somehow.   Could you let me test this.   May need a bit.

Thank you.
---
Koichi Suzuki

2014/06/18 3:46、Aaron Jackson <aja...@re...<mailto:aja...@re...>> のメール：

So, it would appear that after applying this patch, that pg_dump works the first time.  However, I prematurely terminated pg_dump (SIGINT on the terminal) and subsequently issued a count on one of my tables.  The query hung and eventually returned with “WARNING: Xid is invalid.” – the call to psql never actually finished.

I ended up restarting the gtm_proxy, data node and coordinator on host 1.  I think at this point, I’ll do a dump and rebuild the cluster since it seems unstable.

Aaron

From: 鈴木 幸市 [mailto:ko...@in...]
Sent: Monday, June 16, 2014 8:05 PM
To: Aaron Jackson
Cc: Postgres-XC mailing list
Subject: Re: [Postgres-xc-general] Problem with taking database dump with pgxc-1.2.1

Year, it’s XC bug.   Because of the sequence handling difference from PG, transaction should not be read-only.   Please try the attached patch.   It will go to 1.2 and master tree if it works.

Hope it helps.

Regards;
---
Koichi Suzuki

2014/06/17 4:25、Aaron Jackson <aja...@re...<mailto:aja...@re...>> のメール：


So, I tried the same thing and applied this patch.  I even verified that it was setting the transaction mode to "serializable, read only, deferrable" - yet the problem persists.  I believe my execution is similar to the OP.

pg_dump -d analysisdb --serializable-deferrable

pg_dump: [archiver (db)] query failed: ERROR:  cannot execute nextval() in a read-only transaction
pg_dump: [archiver (db)] query was: SELECT pg_catalog.nextval('customtype_customtypeid_seq');

Did I miss something else?

Aaron

________________________________
From: Juned Khan [jkh...@gm...<mailto:jkh...@gm...>]
Sent: Friday, April 11, 2014 2:05 AM
To: Pavan Deolasee
Cc: Postgres-XC mailing list
Subject: Re: [Postgres-xc-general] Problem with taking database dump with pgxc-1.2.1
Thanks to Pavan and Koichi for your inputs.
So i think as of now i do not need to worry about
this


On Fri, Apr 11, 2014 at 12:31 PM, Pavan Deolasee <pav...@gm...<mailto:pav...@gm...>> wrote:
On Fri, Apr 11, 2014 at 12:07 PM, Juned Khan <jkh...@gm...<mailto:jkh...@gm...>> wrote:
As of now this fix worked fine for me. is there any side effect of this
patch ?

IMV there won't be any significant side effect of the patch. pg_dump used to run in RW mode till we fixed it in 9.3.  If it has worked for so many years, it may work for some more time too :-) Marking pg_dump READ ONLY was exactly to catch bugs like these. But other than that, I think its OK.

Thanks,
Pavan

--
Pavan Deolasee
http://www.linkedin.com/in/pavandeolasee



--
Thanks,
Juned Khan
iNextrix Technologies Pvt Ltd.
www.inextrix.com<http://www.inextrix.com/>
------------------------------------------------------------------------------
HPCC Systems Open Source Big Data Platform from LexisNexis Risk Solutions
Find What Matters Most in Your Big Data with HPCC Systems
Open Source. Fast. Scalable. Simple. Ideal for Dirty Data.
Leverages Graph Analysis for Fast Processing & Easy Data Exploration
http://p.sf.net/sfu/hpccsystems_______________________________________________
Postgres-xc-general mailing list
Pos...@li...<mailto:Pos...@li...>
https://lists.sourceforge.net/lists/listinfo/postgres-xc-general