Re: [pgAdmin4][Patch]: Allow user to cancel long running queries from dashboard

Hi Dave,

On Mon, Jul 24, 2017 at 4:44 PM, Dave Page <dpage(at)pgadmin(dot)org> wrote:

> Thanks, applied.
>
> How much effort do you think it would take to turn the table into a
> subnode control so we can show the SQL query and other missing fields from
> pg_stat_activity on that tab?
>
I think around 2-3 days would be needed.

>
>
On Mon, Jul 24, 2017 at 11:45 AM, Murtuza Zabuawala <murtuza.zabuawala@
> enterprisedb.com> wrote:
>
>> Hi Dave,
>>
>> Please find updated patch.
>>
>> On Tue, Jul 18, 2017 at 8:05 PM, Murtuza Zabuawala <
>> murtuza(dot)zabuawala(at)enterprisedb(dot)com> wrote:
>>
>>> Hi Shirley,
>>>
>>> On Tue, Jul 18, 2017 at 1:21 AM, Shirley Wang <swang(at)pivotal(dot)io> wrote:
>>>
>>>> Hi!
>>>>
>>>> I can't seem to get the patch to completely work on my computer, only
>>>> the close icon shows up plus the dialog and success/error messages, but
>>>> here are some comments:
>>>>
>>>> Because we are just cancelling the active running query, so if the
>>> start of the session is 'Active' when you cancel it will simply goto 'Idle'
>>> stat.
>>>
>>>> +1 to Dave's comment about refreshing after the cancel operation
>>>>
>>>> I'll fix this.
>>>
>>>> - We're working on a patch for updating alerts in the Dashboard tab
>>>> which updates the grays in the Database activities panel and changes the
>>>> border around the refresh button and search bar to 1px. This hasn't been
>>>> submitted yet but just a heads up as you work on the alignment.
>>>>
>>>> - Something to consider is how a super user will identify which session
>>>> should be closed. Is that information there?
>>>>
>>>> I think super user can cancel everything except main connection session
>>> & as Dave mentioned in previous email that background workers in PG10.
>>>
>>>> - Are there sessions that should never be closed? If so, do they also
>>>> need close buttons? (Probably not, because that will lead the user to an
>>>> error message, which is not fun)
>>>>
>>>> In Backgrid, we can not exclude specific column from certain rows if it
>>> renders in one row then it will render for every row in the grid, What we
>>> can do is, it will throw an error when user is not eligible to cancel the
>>> active running query.
>>>
>>>> - Perhaps this is a good feature to review with Chethana! :)
>>>>
>>>> On Mon, Jul 17, 2017 at 5:37 AM Murtuza Zabuawala <
>>>> murtuza(dot)zabuawala(at)enterprisedb(dot)com> wrote:
>>>>
>>>>> On Mon, Jul 17, 2017 at 3:01 PM, Dave Page <dpage(at)pgadmin(dot)org> wrote:
>>>>>
>>>>>> Hi
>>>>>>
>>>>>> On Thu, Jul 13, 2017 at 2:53 PM, Murtuza Zabuawala <
>>>>>> murtuza(dot)zabuawala(at)enterprisedb(dot)com> wrote:
>>>>>>
>>>>>>> hiHi,
>>>>>>>
>>>>>>> PFA patch to add functionality which will allow super user to cancel
>>>>>>> long running queries from dashboard.
>>>>>>> RM#1812
>>>>>>>
>>>>>>> *Steps used to test:*
>>>>>>> 1) Open psql session, Connect to 'test' database on respective server
>>>>>>> 2) Execute "select pg_sleep(1000);"
>>>>>>> 3) Open pgAdmin4
>>>>>>> 4) Connect to respective server
>>>>>>> 5) Click on Dashboard
>>>>>>> 6) Check "Sessions" tab under "Server activity" section then look
>>>>>>> for active sessions for test database.
>>>>>>> 7) Click on cancel button and cancel the active session
>>>>>>> 8) Check psql session now, you will see "ERROR: canceling statement
>>>>>>> due to user request"
>>>>>>>
>>>>>>> Some comments:
>>>>>>
>>>>>> - The action here is to cancel the active query in the backend, not
>>>>>> the session - so messages etc. should say things like "Cancel Active Query?"
>>>>>>
>>>>>>
>>>>> - The grid should refresh following the cancel operation.
>>>>>>
>>>>>> - Can you fix the vertical alignment while you're working on this?
>>>>>> The new button really makes the poor alignment stand out.
>>>>>>
>>>>>> - This should not be superuser only - regular users should be able to
>>>>>> cancel their own queries.
>>>>>>
>>>>>> - On PG10, background workers are also shown in the dashboard. Should
>>>>>> we prevent attempts to cancel their work (they'll fail anyway I believe).
>>>>>>
>>>>>> Thanks!
>>>>>>
>>>>>>
>>>>> Sure, I'll work on these comments & send updated patch.
>>>>>
>>>>>> --
>>>>>> Dave Page
>>>>>> Blog: http://pgsnake.blogspot.com
>>>>>> Twitter: @pgsnake
>>>>>>
>>>>>> EnterpriseDB UK: http://www.enterprisedb.com
>>>>>> The Enterprise PostgreSQL Company
>>>>>>
>>>>>
>>>
>>
>
>
> --
> Dave Page
> Blog: http://pgsnake.blogspot.com
> Twitter: @pgsnake
>
> EnterpriseDB UK: http://www.enterprisedb.com
> The Enterprise PostgreSQL Company
>