Re: UI not responsive until postgresql restart

Hi Michael,

Please provide pgAdmin logs. Please refer https://www.pgadmin.org/faq/#8 to
enable the logging.
Also, provide the network tab (Browser's Developer tool > Network tab)
logs.

Thanks,
Khushboo

On Thu, Jul 25, 2019 at 2:37 PM Michael Monerau <michael(dot)monerau(at)gmail(dot)com>
wrote:

> Hi,
>
> I'm not sure that's something you want to investigate, but to arouse your
> curiosity, here is another screenshot of the 100% reproducible behaviour on
> my Fedora 30 box:
>
> [image: image.png]
>
> It looks like there is definitely an issue with the way the initial
> requests are sent in this configuration. It is reproducible (each time I
> restart).
> The requests end up working in the end, but you can see it's hung in a
> whole bunch of useless stuff.
>
> Anyway, it looks weird to have the UI show negative values for counts. An
> indication that it may be worth investigating in my opinion.
>
> The 32000-ish +/- value may indicate it has something to do with a bug
> with 16bits integers and the first bit for sign. At least in the UI.
>
> Best,
> Michael
>
>
> Le mer. 17 juil. 2019 à 17:34, Michael Monerau <michael(dot)monerau(at)gmail(dot)com>
> a écrit :
>
>> Hello,
>>
>> I use pgadmin 4.10, postgresql 11.3 on Fedora 30.
>>
>> Whenever I log into pgadmin for the first time in a session, I can't move
>> around in the databases and it keeps "loading" with loading circles in the
>> tree layout.
>>
>> To unblock this, I found out that running "sudo service postgresql
>> restart" solves the issue immediately, and then everything works fine.
>>
>> Here is the stats dashboard just after the restart, when everything goes
>> back to normal:
>>
>> [image: image.png]
>>
>> You can see there is a huge spike, and it must be the reason why it gets
>> stuck.
>>
>> On similar occasions, I already saw such charts, with the number going
>> symetrically (like a sinus) in similar positive & negative values.
>>
>> It is not a huge pain, because restarting the server fixes the issue. But
>> maybe there is a way to avoid / fix this altogether? Or investigate? Or
>> it's not on pgadmin side?
>>
>> Best,
>> Michael
>>
>