Re: [Gnumed-bugs] Fixup Server failed

[Karsten Hilbert]

On Mon, Sep 02, 2019 at 09:28:00PM +0200, slappinjohn wrote:

> gm-fixup-server 22 failed after latest update.
>
> a pk seems to be in use twice - how can I get rid of this problem?

Indeed ...

> 2019-09-02 21:24:07  INFO      gm.bootstrapper  [140446077142464 MainThread]  
> (./bootstrap_gm_db_system.py::reindex_all() #1156): REINDEXing cloned target 
> database so upgrade does not fail in case of a broken index
> 2019-09-02 21:24:07  INFO      gm.bootstrapper  [140446077142464 MainThread]  
> (./bootstrap_gm_db_system.py::reindex_all() #1157): this may potentially take 
> "quite a long time" depending on how much data there is in the database
> 2019-09-02 21:24:07  INFO      gm.bootstrapper  [140446077142464 MainThread]  
> (./bootstrap_gm_db_system.py::reindex_all() #1158): you may want to monitor 
> the PostgreSQL log for signs of progress
> 2019-09-02 21:24:34  ERROR     gm.bootstrapper  [140446077142464 MainThread]  
> (./bootstrap_gm_db_system.py::reindex_all() #1168): >>>[REINDEX (VERBOSE) 
> DATABASE gnumed_v22]<<< failed
> Traceback (most recent call last):
>   File "./bootstrap_gm_db_system.py", line 1166, in reindex_all
>     curs_outer.execute(cmd)
>   File "/usr/lib/python2.7/dist-packages/psycopg2/extras.py", line 144, in 
> execute
>     return super(DictCursor, self).execute(query, vars)
> IntegrityError: could not create unique index "access_log_pkey"
> DETAIL:  Key (pk)=(67314) is duplicated.
>
> 2019-09-02 21:24:34  INFO      gm.bootstrapper  [140446077142464 MainThread]  
> (./bootstrap_gm_db_system.py::reindex_all() #1170): attempting REINDEXing 
> without VERBOSE
> 2019-09-02 21:25:02  ERROR     gm.bootstrapper  [140446077142464 MainThread]  
> (./bootstrap_gm_db_system.py::reindex_all() #1176): >>>[REINDEX DATABASE 
> gnumed_v22]<<< failed
> Traceback (most recent call last):
>   File "./bootstrap_gm_db_system.py", line 1174, in reindex_all
>     curs_inner.execute(cmd)
>   File "/usr/lib/python2.7/dist-packages/psycopg2/extras.py", line 144, in 
> execute
>     return super(DictCursor, self).execute(query, vars)
> IntegrityError: could not create unique index "access_log_pkey"
> DETAIL:  Key (pk)=(67314) is duplicated.
>
> 2019-09-02 21:25:02  ERROR     gm.bootstrapper  [140446077142464 MainThread]  
> (./bootstrap_gm_db_system.py::__bootstrap() #652): cannot REINDEX cloned 
> target database

Strange how this can have happened.

But this is exactly *why* we do a REINDEX at such time: to
catch such problems early.

Very interesting would be the output of

        SELECT * FROM gm.access_log WHERE pk = 67314;

which should show the duplicated rows.

There's a doable fix but it somewhat depends on what those
rows contain.

Best,
Karsten
--
GPG  40BE 5B0E C98E 1713 AFA6  5BC0 3BEA AC80 7D4F C89B