|Re: [p2-dev] read timeout problems|
Hey Ian!Thanks for the quick reply and the explanations. That helps, but I still can't figure out what the exact problem is.
Re: the p2.index file. This only instructs p2 for which 'types' of repositories to load. The 'type' is indexed by a key. In the case of composite metadata repositories, the key is compositeContent.xml. This says nothing about the file names to load (these are handled by the particular extension that implements this repository type). So the file could bee foo.txt for all we know. Now, it happens that composite metadata repos check compositeContent.jar and composƒiteContent.xml (and it uses a particular search order, .jar first I think).
Ok, that makes sense to me and is very good to know. Thanks for the details a lot!!!
As for the access denied, yes, I know AWS does this (and some consider this a good security measure, since you can't figure out the difference between accessed denied and not available). The transport layer _should_ fail fast in this case and move on to the xml file.
It doesn't seem to be a general problem since I am not able to reproduce it, but we have a number of users reporting issues with the "check for updates" and that it is reporting read timeouts or is stalling on accessing some of those not-really-existing URLs (for which AWS gives you those access denied XML documents).
Is there an option to turn on for more debug info or so? So that we can find out what is really going on in this case? Or do you have any other hints or advice how to find out what is going wrong there (remotely)?
What version of p2 are you using? I know this was a problem a while ago, but it was fixed (IIRC 3.6.1).
The latest STS builds are based on Indigo 3.7.1, but it could be that people install it into older Eclipse installations. I will ask them.
Thanks a lot for your help!!! Really highly appreciated!!!!!!!! -Martin
Cheers, Ian On Fri, Jan 6, 2012 at 8:08 AM, Henrik Lindberg <henrik.lindberg@xxxxxxxxxxxxxx <mailto:henrik.lindberg@xxxxxxxxxxxxxx>> wrote: Odd that the information in the p2.index does not seem to be honoured. Maybe it continues if there is a problem delivering it (timeout), but I don't really know the logic around the p2.index file. This is Ian Bull's domain IIRC. Regarding delivering an access denied status when a not found (404) should have been delivered is never good as you are basically saying "the file is there but you are not allowed to read it". One potential source of problems is if the update site has changed over time, and user once got hold of index information (in jars or xml files) that are no longer present - I can imagine checks being made in those cases for those specific files even if the p2.index file says otherwise. In this case delivering an access denied instead of 404 does not give the p2 client side a chance to correctly update the cached information as it will continue to believed the file is there. (I am just speculating though). Hope that is of some help. Regards Henrik Lindberg henrik.lindberg@xxxxxxxxxxxxxx <mailto:henrik.lindberg@xxxxxxxxxxxxxx> On Jan 6, 2012, at 15:21, Martin Lippert wrote:Hi! I am facing an interesting problem with update sites and I hope you can help me understand what is going on. We have plenty of composite update sites that all contain just an "compositeArtifact.xml" and a "compositeContent.xml" file. Those files work nicely. But we are quite often getting reports from users that their installation stalls when they check for updates, complaining about "content.jar" could not be found (with a read timeout), same for "compositeArtifact.jar" or similar. Even if I put up a p2.index file like this: version=1 metadata.repository.factory.order=compositeContent.xml,! artifact.repository.factory.order=compositeArtifacts.xml,! it doesn't change much, the "check for updates" is again complaining about a missing file "compositeArtifact.jar". I don't know why and what to do about this. One thing that is special here is the fact that our server (Amazon S3) is serving an XML document (containing an access denied error) when a resource is not found (like it is the case for those resources that p2 seems to try to download. Any idea what I can do about this? Thank you very much in advance for your help!!! Cheers, -Martin _______________________________________________ p2-dev mailing list p2-dev@xxxxxxxxxxx <mailto:p2-dev@xxxxxxxxxxx> https://dev.eclipse.org/mailman/listinfo/p2-dev_______________________________________________ p2-dev mailing list p2-dev@xxxxxxxxxxx <mailto:p2-dev@xxxxxxxxxxx> https://dev.eclipse.org/mailman/listinfo/p2-dev -- R. Ian Bull | EclipseSource Victoria | +1 250 477 7484 http://eclipsesource.com | http://twitter.com/eclipsesource _______________________________________________ p2-dev mailing list p2-dev@xxxxxxxxxxx https://dev.eclipse.org/mailman/listinfo/p2-dev
Back to the top