[Bucardo-general] Possible bug?

[Adam Kozłowski]

Hi,

Try to remove table then add table once again and add sync.
#bucardo_ctl remove table ...
#bucardo_ctl add table …
#bucardo_ctl add sync ...

I can't check right now, but bucardo probable gather information about
tables while adding tables...


/Adam


2013/5/14 Olaf Rühenbeck <some.individuum at gmail.com>:
> Hey there,
>
> I try to use bucardo do replicate a few databases between 2 hosts to have a
> always up to date backup copy of our database in another host. In this case
> the bacula director catalog database.
>
> I try to add a sync:
>
> # bucardo add sync m_baculadir_v herd=maguro dbs=maguro onetimecopy=2
>
> and get:
>
> Failed to add sync: DBD::Pg::st execute failed: ERROR:  Table
> "public.jobhisto" must specify a primary key! at line 117. at line 30.
> CONTEXT:  PL/Perl function "validate_sync" at /usr/bin/bucardo line 4113.
>
> As its the first sync I used onetimecopy=2 to get the target database fully
> populated.
>
> I read in the archives of this list that bucardo always requires a primary
> key on every table it tries to sync, this is kinda annoying and I will have
> to write a script to take care of this after every major bacula database
> schema upgrade, but also if I try to set onetimecopy=1 to avoid the
> necessity of the primary key it comes up with the same message. So I added a
> primary key to the table in the source database and I still get the same
> error:
>
>
> baculadir=# alter table jobhisto add primary key (jobid);
> NOTICE:  ALTER TABLE / ADD PRIMARY KEY will create implicit index
> "jobhisto_pkey" for table "jobhisto"
> ALTER TABLE
> baculadir=# \d jobhisto
>                   Table "public.jobhisto"
>      Column      |            Type             | Modifiers
> -----------------+-----------------------------+-----------
>  jobid           | integer                     | not null
>  job             | text                        | not null
>  name            | text                        | not null
>  type            | character(1)                | not null
>  level           | character(1)                | not null
>  clientid        | integer                     |
>  jobstatus       | character(1)                | not null
>  schedtime       | timestamp without time zone |
>  starttime       | timestamp without time zone |
>  endtime         | timestamp without time zone |
>  realendtime     | timestamp without time zone |
>  jobtdate        | bigint                      |
>  volsessionid    | integer                     |
>  volsessiontime  | integer                     |
>  jobfiles        | integer                     |
>  jobbytes        | bigint                      |
>  readbytes       | bigint                      |
>  joberrors       | integer                     |
>  jobmissingfiles | integer                     |
>  poolid          | integer                     |
>  filesetid       | integer                     |
>  priorjobid      | integer                     |
>  purgedfiles     | smallint                    |
>  hasbase         | smallint                    |
>  hascache        | smallint                    |
>  reviewed        | smallint                    |
>  comment         | text                        |
> Indexes:
>     "jobhisto_pkey" PRIMARY KEY, btree (jobid)
>     "jobhisto_idx" btree (starttime)
>
> baculadir=# \q
>
> And still get the same error:
>
> # bucardo add sync m_baculadir_v herd=maguro dbs=maguro onetimecopy=1
> Failed to add sync: DBD::Pg::st execute failed: ERROR:  Table
> "public.jobhisto" must specify a primary key! at line 117. at line 30.
> CONTEXT:  PL/Perl function "validate_sync" at /usr/bin/bucardo line 4113.
>
> Any ideas on this? Would appreciate some help :)
>
> thanks,
>
> Olaf
>
> _______________________________________________
> Bucardo-general mailing list
> Bucardo-general at bucardo.org
> https://mail.endcrypt.com/mailman/listinfo/bucardo-general
>