[Bucardo-general] FW: not reliable RESULTS with 'bucardo status' and 'bucardo kick'

[Markova, Nina]

Hi David,

I found that the sync is not active for other reasons: Mismatch schema in source and target database - columns with NULL/DEFAULT  value.
MCP marks it as inactive and the validation FAILED.  
So, it looks like it is not 'bucardo deactivate' -> 'bucardo activate' issue. 
But I still expect to see if validation of the sync is failing or not - that would safe me some digging through the logs. 

Have you or anyone else experienced that? I am running bicardo 4.99.8 github version from Oct 31 2013

*** Somehow bucardo commands doesn't catch that and doesn't report it  properly.  ***  Bug?

1) on startup bucardo doesn't report that failure on the command line, only in the log ->  [Fri Nov  8 22:17:56 2013]  MCP Validation of sync sta_sync_Mo3_So2 FAILED
      (12973) [Fri Nov  8 22:17:56 2013] MCP    Inspecting target table "public.sitechan" on database "sta_pgo2"
      (12973) [Fri Nov  8 22:17:56 2013] MCP   Inspecting source table "public.a0factor" on database "sta_pgo3"
      (12973) [Fri Nov  8 22:17:56 2013] MCP    Inspecting target table "public.a0factor" on database "sta_pgo2"
      (12973) [Fri Nov  8 22:17:56 2013] MCP Warning: Source database for sync "sta_sync_Mo3_So2" has column "initials" of table "public.a0factor" set as NULL, but target database  "sta_pgo2" has column set as NOT NULL
     (12973) [Fri Nov  8 22:17:56 2013] MCP Warning: Source database for sync "sta_sync_Mo3_So2" has column "initials" of table "public.a0factor" has no DEFAULT value, but target database "sta_pgo2" does
     (12973) [Fri Nov  8 22:17:56 2013] MCP Warning: Source database for sync "sta_sync_Mo3_So2" has column "lddate" of table "public.a0factor" set as NULL, but target database "sta_pgo2" has column set as NOT NULL
     (12973) [Fri Nov  8 22:17:56 2013] MCP Warning: Source database for sync "sta_sync_Mo3_So2" has column "normfactor" of table "public.a0factor" set as NULL, but target database "sta_pgo2" has column set as NOT NULL
     (12973) [Fri Nov  8 22:17:56 2013] MCP Validation of sync sta_sync_Mo3_So2 FAILED
     (12973) [Fri Nov  8 22:17:56 2013] MCP Activating sync "sta_sync_Mo3_Ss2"
     (12973) [Fri Nov  8 22:17:56 2013] MCP Running validate_sync on "sta_sync_Mo3_Ss2"
      ... etc ...

2) bucardo validate all
     Validating sync sta_sync ............... OK
      Validating sync sta_sync_Mo3_So2 ... OK               # validated OK, while on startup  bucardo reported it (in log.bucardo) as FAILED 

3) bucardo status
    PID of Bucardo MCP: 12973
    Name                   State              Last good               Time           Last I/D    Last bad                Time           
    ======================+==================+=======================+==============+===========+=======================+============
    sta_sync             | Good             | Nov 08, 2013 22:20:54 | 95h 36m 51s  | 0/1       | Oct 07, 2013 19:15:00 | 866h 42m 44s  
    sta_sync_Mo3_So2     | Good             | Oct 24, 2013 17:31:21 | 460h 26m 24s | 0/0       | none                  |                 

   # reported as Good

 4) bucardo activate   sta_sync_Mo3_So2
      Activating sync sta_sync_Mo3_So2

      bucardo kick sta_sync_Mo3_So2
      Kicked sync sta_sync_Mo3_So2

So, bucardo command reports the sync is kicked ... while bucardo log shows it is not :
      (9057) [Fri Nov  8 16:25:35 2013] MCP Cannot kick inactive sync "sta_sync_Mo3_So2"
      (9057) [Fri Nov  8 18:41:18 2013] MCP Cannot kick inactive sync "sta_sync_Mo3_So2"


Cheers,
Nina

-----Original Message-----
From: Markova, Nina 
Sent: November-08-13 17:05
To: 'David E. Wheeler'
Subject: RE: [Bucardo-general] deactivate a sync - bug/ not reliable RESULTS with 'bucardo status'

Thanks, that's encouraging.
I'll create a new sync and will do similar tests.
 
Maybe it will work if from scratch?!

Cheers,
Nina


-----Original Message-----
From: David E. Wheeler [mailto:david at justatheory.com] 
Sent: November-08-13 16:56
To: Markova, Nina
Subject: Re: [Bucardo-general] deactivate a sync - bug/ not reliable RESULTS with 'bucardo status'


On Nov 8, 2013, at 11:55 AM, Markova, Nina <Nina.Markova at NRCan-RNCan.gc.ca> wrote:

> For 4.99.8 I run 'bucardo upgrade' and at that time the issue (deactivated sync that appears as active) was there.
> 
> Just tried one more thing: I tried to deactivate it again and then to activate it:
> 
> $ bucardo deactivate sta_sync_Mo3_So2
>   Deactivating sync sta_sync_Mo3_So2
>   Log.bucardo:
>   (9057) [Fri Nov  8 19:50:49 2013] MCP Sync "sta_sync_Mo3_So2" is already deactivated
> 
> $ bucardo activate sta_sync_Mo3_So2
>  Activating sync sta_sync_Mo3_So2
> 
> $ bucardo kick sta_sync_Mo3_So2
>  (9057) [Fri Nov  8 19:51:36 2013] MCP Cannot kick inactive sync "sta_sync_Mo3_So2"

Hrm. I get:

    $ bucardo deactivate dwheeler_sync
    Deactivating sync dwheeler_sync

And the sync is not running anymore. The output of `bucardo status dwheeler_sync` shows it as inactive (that was the bug I fixed; previously it would say it was active, even though it wasn't).

then I activate it again:

    $ bucardo activate dwheeler_sync
    Activating sync dwheeler_sync

It's back in the process table, running the sync, and the output of `bucardo status dwheeler_sync` shows it as active. All good. If I try to kick it:

    $ bucardo kick dwheeler_sync
    Kicked sync dwheeler_sync

All good, exactly as it should be.

What is the output of `bucardo status sta_sync_Mo3_So2`?

Best,

David

PS: Mail lists are expected to be active; feel free to post there and Cc me. Others might find the conversation of use in the archives someday.