Re: pgAdmin4: Random failure of FTS test cases due to improper randomstring creation - Mailing list pgadmin-hackers

From Navnath Gadakh
Subject Re: pgAdmin4: Random failure of FTS test cases due to improper randomstring creation
Date
Msg-id CAOAJCYqB_dWHXAzNmQ0tZv5BKUOsdOz=uawRew4Zc2Q7xfNazQ@mail.gmail.com
Whole thread Raw
In response to Re: pgAdmin4: Random failure of FTS test cases due to improper randomstring creation  (Dave Page <dave.page@enterprisedb.com>)
Responses Re: pgAdmin4: Random failure of FTS test cases due to improper randomstring creation  (Dave Page <dave.page@enterprisedb.com>)
List pgadmin-hackers
Hi Dave,

      Please find the attached patch. The code added to tear down the FTS related objects. As this issue was random, I have tested this patch on all 12 servers (pg/ppas) with multiple time and got no errors. 

Thanks.


On Fri, Aug 18, 2017 at 1:35 PM, Dave Page <dave.page@enterprisedb.com> wrote:


On Fri, Aug 18, 2017 at 4:54 AM, Ashesh Vashi <ashesh.vashi@enterprisedb.com> wrote:
On Fri, Aug 11, 2017 at 1:24 PM, Navnath Gadakh <navnath.gadakh@enterprisedb.com> wrote:
Hi Dave,

        Please find the attached patch for UUID creation issues with test objects for FTS configurations, FTS dictionaries and FTS parsers. Previously(refer email with subject "Build failed in Jenkins: pgadmin4-master-python27 #279" and "Build failed in Jenkins: pgadmin4-master-python33 #207"), test cases were randomly failing due to repetitions of the test object names.

In the old code we used some part of the string for creating a UUID string, but it seems that at some point that created string gets repeated and due to which test cases were failing as this was a random behavior, now it is fixed.
Navnath,

We're still not removing the temporary objects, created by test-cases, in the tear-down function.
We should have removed them in the tear-down function to fix the issue in proper way.

Dave - thoughts?

Right - it only seems to be FTS and FDW related objects that suffer from this problem from what I can see, so I assume we're getting it right for everything else by removing objects in the tear-down.
 
--
Dave Page
VP, Chief Architect, Tools & Installers
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company

Blog: http://pgsnake.blogspot.com
Twitter: @pgsnake



--
Regards,
Navnath Gadakh

EnterpriseDB Corporation
The Enterprise PostgreSQL Company


Attachment

pgadmin-hackers by date:

Previous
From: Teng Zhang
Date:
Subject: Fwd: [pgadmin4][Patch] Greenplum specific DDL and Dashboard display
Next
From: Surinder Kumar
Date:
Subject: Re: Next release