search

mess110 / coinmarketcap.northpole.ro

A straightforward JSON API for all the currencies and assets found on coinmarketcap.com crawled every 5 minutes
MIT License
41 stars 11 forks source link

Server does not update anymore. #23

Closed vanlooverenkoen closed 7 years ago

vanlooverenkoen commented 7 years ago

All responses have the same timestamp. They don't change.

1500980402

I don't know why? And how the timestamp is generated. Please take a look at this.

vanlooverenkoen commented 7 years ago

Also the data is never updated

ekerstein commented 7 years ago

Seems to have stopped on 7/25. For example if you look at http://coinmarketcap.northpole.ro/history.json?coin=bitcoin you'll see that's the most recent date.

vanlooverenkoen commented 7 years ago

Yes I know, but I can't restart them. So I'm stuck here.

mess110 commented 7 years ago

is this still the case? Am on holiday, but I did a quick run by hand and I didn't get errors. Will monitor

mess110 commented 7 years ago

Please leave any details you might have. Thx

vanlooverenkoen commented 7 years ago

There are no errors the data is just outdated. And hasn't been updated for 3 days.

iwok commented 7 years ago

I think there is a problem with the cronjob. Having issues getting the script running on my server. The command is executed by hand but not via the cronjob

jupiter50 commented 7 years ago

The cron job using the latest code seems to work fine with the v6 API and populating the last few days. However, looks like it is updating at 5am UTC

vanlooverenkoen commented 7 years ago

When did you pull from the server?

jupiter50 commented 7 years ago

A few days ago. Before that there was a bit of a data gap similar to what we are seeing on the main site so I tried upgrading the code on my server to see if it resolved it.

The history.json I dont have working but I raised that in another issue as its not in the source code

vanlooverenkoen commented 7 years ago

Can @mess110 change it back to the old version?

vanlooverenkoen commented 7 years ago

Where do you host your this website? And do you know where @mess110 hosts his server?

jupiter50 commented 7 years ago

Well the old version has its flaws (and v6 can still be accessed - http://coinmarketcap.northpole.ro/api/v6/history/BTC_2017.json). The new version using IDs is more robust so we should wait till he is back from his holidays to try see if he can fix it rather than revert back.

I believe he hosts it here: http://coinmarketcap.northpole.ro/

vanlooverenkoen commented 7 years ago

Is that one still updated till the latest 5 minutes?

jupiter50 commented 7 years ago

The latest code generates v6 and v8 of the api. Both sets update every 5 mins I believe but also store it on a daily basis under each respective "history"

vanlooverenkoen commented 7 years ago

http://coinmarketcap.northpole.ro/ticker.json?identifier=bitcoin

Hasn't been updated in 2 days

jupiter50 commented 7 years ago

Yes, I saw the same problem on my server. My server though has been updating the last 2 days (after my upgrade of the code). Probably need @mess110 to see why it sometimes fail as it seems random so far

vanlooverenkoen commented 7 years ago

@mess110 when do you think you can fix this?

vanlooverenkoen commented 7 years ago

I think it works again. What did you change?

mess110 commented 7 years ago

Nothing yet, just landed still in plane. Will catch up and fix this today

mess110 commented 7 years ago

Added some extra logging, but I can't explain the crash atm. Sorry for the inconveniences caused. I have been monitoring this for about 30 min and seems to work now. Please confirm if this is the case.

Will continue to keep an eye on this. More details will come once I have them

vanlooverenkoen commented 7 years ago

Yes it's working since yesterday morning. I don't know why but the data is updating since then.

vanlooverenkoen commented 7 years ago

Do you know what caused it?

mess110 commented 7 years ago

There were crashes related to writing/reading to/from disk. Now I am catching those exceptions. In addition to this, I modified some code on production part of the experiment which caused more crashes. The 2nd part was fixed after doing a fresh deploy a few days ago.

Sry about the inconveniences