Re: [scipion-users] Metadata loading

[Pablo <p.c...@gm...>]

Wow! We certainly need to have your case in mind (589 runs) when testing 
performance.

Collapsing some branches actually reduce "drawing time" but the whole 
project is loaded which takes most of the time.

Now I can't see any other workaround than deleting runs. If you want to 
keep them all, another option could be to copy the project folder (watch 
out for the disk space), and in the new project delete runs.
But I can understand this way you loose the whole picture of the project.

I've added your case into our issue, and will discuss your case in our 
next development meeting.

:-(

Pablo.

On 27/09/16 16:26, Сергей Назаров wrote:
> Dear Pablo,
>
> Thank you for a quick reply. Unfortunately, I don't have many 
> protocols running at the same time, usually from 1 to 3. But the 
> project is huge, the number of folders in Runs/ is 589. I keep most of 
> branches collapsed, hopefully it prevent them from loading.
>
> Best,
> Sergey
>
> 2016-09-26 19:00 GMT+02:00 Pablo <p.c...@gm... 
> <mailto:p.c...@gm...>>:
>
>     Dear Sergey,
>
>     Thanks a lot for your feedback, this is very important for us to
>     know important issues and also to give priority to new features.
>     We are aware of some performance issues related with the metadata
>     loading, which sometimes blocks the GUI, as you are describing. Is
>     this happening to  you when many protocols are in ‘running’ state?
>     Or just when the project has many protocols? We are working to
>     find a solution to this problem and make the processing with
>     Scipion more effective, sorry for the inconveniences.
>
>     All the best,
>     Pablo
>     Scipion team.
>
>
>
>     On 26/09/16 15:52, Сергей Назаров wrote:
>
>         Dear Scipion Community,
>
>         I constantly experience issue with loading metadata to my
>         projects.
>         It can get really annoying, I wait sometimes almost a minute
>         to be able to adjust the description of a protocol, choose
>         data (particles. models etc.). I found this behavior both on
>         local machines and on cluster through ssh. While I can
>         understand slow metadata loading through ssh, especially for a
>         big project with many protocols already executed, on local
>         machine this behavior appears after sometime after reboot. I
>         don't know if it starts after some threshold amount of
>         protocols or not, but after machine reboot everything runs
>         nice and smooth for certain time and then starts again.
>         I found one closed request on github concerning slow metadata,
>         but without any details.
>         Could you please advise me something about this issue? How to
>         behave when your project is on cluster and metadata is slow? I
>         am not superuser, so I am not able to reboot nodes everytime.
>
>         Best,
>         Sergey Nazarov
>         Biozentrum Basel
>
>
>