[pardu]

Hi All,

I have did a fresh install with latest nightly from pre-eden on an ubuntu natty (unstable) x64.
After which I attempt to reconfigure xbmc from scratch since I intentionally removed the .xbmc directory there conf files are stored.

First thing I noticed is that repo is not working properly for the default scrapers:
allmusic, tmdb and the tvdb.

All three of them appear on the available updates list on the addons section in the configuration. But can't be updated either manually or automatically, they seems to sleep for a second when trying to manually update but they come back same as before.

Allmusic from 2.0.1 to 2.0.3
The Movie DB from 1.2.0 to 1.4.0
The TVDB: 1.0.8 to 1.2.3

Here is what happens when trying to manually one of them (same happens for the others):

15:53:50 T:140126762313696 M:2711973888 DEBUG: CFileCache::Open - opening <addons/eden-pre/metadata.tvdb.com/metadata.tvdb.com-1.2.3.zip.md5> using cache
15:53:50 T:140126762313696 M:2711973888 DEBUG: FileCurl::Open(0x2251a80) http://mirrors.xbmc.org/addons/eden-pre/....3.zip.md5
15:53:50 T:140126762313696 M:2711973888 INFO: easy_aquire - Created session to http://mirrors.xbmc.org
15:54:03 T:140126208067328 M:2711736320 DEBUG: GetImageHash - unable to stat url http://mirrors.xbmc.org/addons/eden-pre/...m/icon.png
15:54:03 T:140126208067328 M:2711736320 DEBUG: Thread Background Loader 140126208067328 terminating
15:54:06 T:140126762313696 M:2711793664 WARNING: FillBuffer: curl failed with code 28
15:54:06 T:140126762313696 M:2711793664 ERROR: CFileCurl::CReadState::Open, didn't get any data from stream.
15:54:06 T:140126762313696 M:2711793664 ERROR: Open - failed to open source <http://mirrors.xbmc.org/addons/eden-pre/metadata.tvdb.com/metadata.tvdb.com-1.2.3.zip.md5>

When I manually download the zip for the new plugin and attempt to install log says it is installed correctly however version is still shown as the same and same the "available updates" section.

Regards,

Pardu

[pardu]

Looking in depth seems like this could be a python or curl error?
I have an up-to-date natty version. (Python 2.7.1+)

[pardu]

Seem to be getting somewhere now with values of 40 or above into:
curlclienttimeout and curllowspeedtime
on the advanccedsettings.xml file.

Will keep posted if this ends up solving the issue. In that case question will be has there been seen any traffic speed drop in last couple of months for the site?

[pardu]

Ok, as said on my post above, issue was solved by the curl timeouts on the advanced settings file.

However I can't help to wonder why is mirrors.xbmc.org running so slow nowadays, my internet connection hasn't changed in over a year and I am only seeing this now.

Regards,

[theuni]

It could just be a congestion issue. The addon.xml and md5's are sent by the server directly, not mirrored out due to replication timing.

We'll be monitoring this more in the future to see if it needs a fix.

TheUni

[pardu]

Actually I had tested also with a plugin which also uses python functions to get some webpages. For some reason there seems to be a high delay on the ammount of time python takes to download the page compared with the time it takes for a normal request (like wget).

Here is an example of the addon (pelisalacarta):
02:15:28 T:139869085185792 M:2645381120 NOTICE: [scrapertools.py] petición GET
02:15:43 T:139869085185792 M:2645184512 NOTICE: [scrapertools.py] Descargado en 3 segundos

Notice the timings on the left, it takes 15 seconds to download the simple page but the counter only show 3 seconds have elapsed, that is because the counter is only taking the download time of the page which is 3 secs, but it is taking an extra 12 secs or so per python (and curl?) request which makes the overhead.

Have you heard of such behaviour?
Has anyone else tested the latest nightlies pre-eden on a fresh install on a up-2-date natty system?

Thanks

[pardu]

Ok... Feeling stupid now.

I found out today my ISP was having some problems with a DNS server (of course, the one I was using). I changed to the secondary, now it is working better.