[Bucardo-general] Simple question regarding master-master

Athanasios Kostopoulos athanasios.kostopoulos at classmarkets.com
Fri Jun 14 15:50:54 UTC 2013


Can you elaborate a bit on the "bucardo set search_path ...."?
I am using version 4.99.5 that comes packaged with Debian wheezy, perhaps
this is the source of my problems?


On Fri, Jun 14, 2013 at 4:33 PM, Mitchell Perilstein <
Mitchell.Perilstein at trueposition.com> wrote:

>  Hm, works for me.  (despite one vs two dashes, oops).
>
> # bucardo start --log-destination=syslog -no-exit-on-nosync
> Checking for existing processes
> Removing /var/run/bucardo/fullstopbucardo
> Starting Bucardo
> # bucardo --version
> bucardo version 4.99.7
>
> Regarding your error below, did you do a "bucardo set search_path ...."?
>
>
> On 06/14/2013 10:08 AM, Athanasios Kostopoulos wrote:
>
> I have banged a quick shell script using Mitchell's advice and pseudocode
> (thanks again, this has been of tremendous help) but now I see that it
> "freezes" on activate sync. The following is from log.bucardo:
> (4138) [Fri Jun 14 14:49:56 2013] MCP Created VAC 5154
> (5154) [Fri Jun 14 14:49:56 2013] VAC New VAC daemon. PID=5154
> (5154) [Fri Jun 14 14:49:56 2013] VAC Warning! VAC was killed at line 4729:
> DBD::Pg::db do failed: ERROR:  invalid value for parameter "search_path":
> "bucardo,bucardo, public"
> DETAIL:  schema "bucardo" does not exist at /usr/share/perl5/Bucardo.pm
> line 4729.
> Can't call method "disconnect" without a package or object reference at
> /usr/share/perl5/Bucardo.pm line 6336.
> Issuing rollback() due to DESTROY without explicit disconnect() of
> DBD::Pg::db handle dbname=bucardo;port=5432;host=localhost at
> /usr/share/perl5/Bucardo.pm line 4729.
>
> anyone has an idea what might be the probable cause?
>
>
> On Fri, Jun 14, 2013 at 3:49 PM, Athanasios Kostopoulos <athanasios.kostopoulos at classmarkets.com> wrote:
>
>
>  Thanks for the code Mitchell. One quick question: which version of bucardo
> and under which OS you are using? I tried it on debian wheezy (and the
> packaged bucardo 4.99.5) and I get errors like below:
> Unknown option: log-destination
> Unknown option: no-exit-on-nosync
> Unknown option 'conflict_strategy'
>
> are you using the version from github?
>
>
> On Fri, Jun 14, 2013 at 3:31 PM, Mitchell Perilstein <Mitchell.Perilstein at trueposition.com> wrote:
>
>
>  Again, not saying it's right, but it's roughly something like this. Names
> changed, error checks not shown, repeated for a number of tables and
> databases, etc..
>
>     bucardo add database    source_xx host=tplocalvirt user=myuser db=xx
> dbport=5432
>     bucardo add database    --force target_xx host=mypeerhost user=myuser
> db=xx dbport=5432
>     bucardo add table       $ALLMYTABLES db=source_xx herd=xxherd
>     bucardo add dbgroup     xxgroup source_xx:source target_xx:target
>     bucardo start           --log-destination=syslog -no-exit-on-nosync
>     bucardo add sync        xx_sync herd=xxherd conflict_strategy=latest
> onetimecopy=1 dbs=xxgroup
>     bucardo set syslog_facility=local6
>     bucardo set mcp_dbproblem_sleep=0.5
>     bucardo set isolation_level='read uncommitted'
>     bucardo set statement_chunk_size=1000
>     bucardo activate sync   xx_sync
>
>
>
> On 06/14/2013 09:12 AM, Athanasios Kostopoulos wrote:
>
>
>  Thanks for the reply Mitchell. So you do something along the lines of:
> in master1:
> bucardo add dbgroup master1:source master2:target
> and fire off the relative sync
> and on master2:
> bucardo add dbgroup master2:source master1:target
> and fire off the relative sync?
>
>
> On Fri, Jun 14, 2013 at 3:02 PM, Mitchell Perilstein <
> Mitchell.Perilstein@**trueposition.com<Mitchell.Perilstein at trueposition.com> <Mitchell.Perilstein at trueposition.com>>
> wrote:
>
>    We're running differently from Rosser's one-bucardo suggestion--which
>
>  does make sense.  We're dual master but we need both boxes to be as
> identical as possible partly because the process has to be repeatable on
> many pairs and also because we bounce our apps on occasion but want the
> DBs
> to stay in sync.  So we have bucardo on each box, with each local
> bucardo
> writing to its remote peer.  So it's a criss-cross arrangement.
>
> For our purpose, at least, this also gets us a little bit of HA: even if
> our application layer and bucardo are down on one box, as long as its
> database is running, the fully-up peer box will continue writing its
> changes to the partly down box.
>
> I'm not saying our approach is correct, but it's working for us.
>
>
>
> On 06/14/2013 07:17 AM, Rosser Schwarz wrote:
>
> Bucardo only needs to be installed on one host, although obviously, that
> won't provide HA capability. Typically, in a master-master scenario,
> it's
> installed on one of the masters. In the event either of the masters then
> goes down, the Bucardo daemon wouldn't have anything to do anyway, so HA
> for it is pretty moot.
>
>
> On Fri, Jun 14, 2013 at 3:44 AM, Athanasios Kostopoulos <
> athanasios.kostopoulos@**classmarkets.com<athanasios.kostopoulos at classmarkets.com> <athanasios.kostopoulos at classmarkets.com>>
> wrote:
>
>
>   In continuation to my previous email?
> For example the following approach might be effective:
> - install bucardo on master1
> - setup replication as follows: master1: master - master2: slave
> - verify that replication works
> If the above are correct then:
> - install bucardo on master2
> - setup replication as follows: master2: master - master1: slave
> - verify replication works
> Will the previous be enough or I am missing a fundamental piece here
> (perhaps not needing to run bucardo on two different hosts and use one
> setup (say on master one) and two replication streams. How will this
> handle
> failover?
>
>
> On Thu, Jun 13, 2013 at 5:43 PM, Athanasios Kostopoulos <athanasios.kostopoulos@**classmarkets.com<athanasios.kostopoulos at classmarkets.com> <athanasios.kostopoulos at classmarkets.com>>
> wrote:
>
>
>   Dear all,
> first of all, thanks for the hard work that has been put into bucardo. I
> am a novice and I have the following perhaps trivial question:
> when wanting to create a master-master setup, do I need to install
> bucardo (e.g. bucardo install) on both of the masters? I am on Debian 7
> wheezy with bucardo 4.99.7, installed from the Debian repositories.
> Thanks
> in advance for any answers to my perhaps trivial question.
>
>
>
> ______________________________**_________________
> Bucardo-general mailing listBucardo-general at bucardo.**orghttps://mail.endcrypt.com/**mailman/listinfo/bucardo-**general<http://mail.endcrypt.com/mailman/listinfo/bucardo-general> <http://mail.endcrypt.com/mailman/listinfo/bucardo-general>
>
>
>
> ______________________________**_________________
> Bucardo-general mailing listBucardo-general at bucardo.**orghttps://mail.endcrypt.com/**mailman/listinfo/bucardo-**general<http://mail.endcrypt.com/mailman/listinfo/bucardo-general> <http://mail.endcrypt.com/mailman/listinfo/bucardo-general>
>
>
>
>
>
> Confidentiality Notice: This e-mail (including any attachments) is
> intended only for the recipients named above. It may contain
> confidential
> or privileged information and should not be read, copied or otherwise
> used
> by any other person. If you are not a named recipient, please notify the
> sender of that fact and delete the e-mail from your system.
>
>
>
>  Confidentiality Notice: This e-mail (including any attachments) is
> intended only for the recipients named above. It may contain confidential
> or privileged information and should not be read, copied or otherwise used
> by any other person. If you are not a named recipient, please notify the
> sender of that fact and delete the e-mail from your system.
>
>
>
>
>
>
> Confidentiality Notice: This e-mail (including any attachments) is
> intended only for the recipients named above. It may contain confidential
> or privileged information and should not be read, copied or otherwise used
> by any other person. If you are not a named recipient, please notify the
> sender of that fact and delete the e-mail from your system.
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mail.endcrypt.com/pipermail/bucardo-general/attachments/20130614/0222d5ad/attachment-0001.html>


More information about the Bucardo-general mailing list