Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Improve HSL data getting #5

Open
antont opened this issue Dec 11, 2014 · 5 comments
Open

Improve HSL data getting #5

antont opened this issue Dec 11, 2014 · 5 comments
Assignees

Comments

@antont
Copy link
Member

antont commented Dec 11, 2014

Currently we get the HSL data for busses, trams and metros via CIE's POI server. It polls some HSL endpoint (I don't know which exactly, Stinkfist was asking on IRC but we didn't find the info) every 10 seconds now.

This is basically ok - we can show the objs and they move.

However there are two problems:

  1. The update is quite rare. 10secs is too long an interval for showing moving objects. Tapani is adding client-side interpolation to make it smooth but still it's a rare update.
  2. The HTTP get to the POI server from the client here stalls sometimes. Can take a minute or so to timeout then. Chrome's network debugging tool shows 'request stalled' or so. We've been unable to repeat the problem with manual use of the browser (just refreshing the POI link) -- use jquery's $.getJSON in the code where it happens. We haven't debugged this yet.

Possible solutions:

A. Use HSL service directly, either from Tundra server side or from the Web browser.

B. Keep using the POI intermediate server but debug the stalling, and investigate if the update freq could be e.g. 1 or 3 secs. Or live with 10.

A good benchmark is the existing 2d service with this data, it updates pretty quick and reliably: http://transport.wspgroup.fi/hklkartta/

We decided yesterday in the meet that Admino will work on this data retrieval so assigning to Ali as we agreed on IRC where just discussed this.

@antont
Copy link
Member Author

antont commented Dec 16, 2014

@Stinkfist0 are you on this btw? i mean the stalling prob is afaik still there.

@Stinkfist0
Copy link
Contributor

Haven't been actively researching this. Haven't noticed any problems though.

@antont
Copy link
Member Author

antont commented Dec 16, 2014

I got this immediately in the beginning now with the current version:

screen shot 2014-12-16 at 16 41 02

After taking the screenshot (at 20sec or so 'pending') it lasted exactly 1.0 minutes for the data to come.

And now the 2nd request is again 'pending' as well. Apparently took also 1.0 minutes .. and now 3rd one is pending:

screen shot 2014-12-16 at 16 43 10

You never get these?

Am on a slow mac with poor fps for the demo but Tapani reported this originally from a beefier win desktop.

@Stinkfist0
Copy link
Contributor

Aight, getting and noticing that myself too now.

@erno
Copy link
Contributor

erno commented Jan 8, 2015

After a ~40 requests I got this to replicate with vanilla wget as well,

for i in $(seq 10); do time wget -O/dev/null -q http://dev.cie.fi/FI-WARE/mlevanto/dynpoi/server/radial_search; sleep 1; done
...
wget -O/dev/null -q 0.00s user 0.01s system 0% cpu 1:01.78 total

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

3 participants