Hi Noorul,

Just a follow up. I've been able to isolate (I think) the issue and
I've opened the corresponding JIRA [1]. I asked the chef-dev mailing
list [2] to see if someone can give us the details of how to properly
sign the URLs that are failing. It seems that some change may be
required in the signing filter [3].

Let's see if we can get this fixed!


[1] https://issues.apache.org/jira/browse/JCLOUDS-254
[2] http://lists.opscode.com/sympa/arc/chef-dev/2013-08/msg00034.html
[3] 
https://github.com/jclouds/jclouds-chef/blob/master/core/src/main/java/org/jclouds/chef/filters/SignedHeaderAuth.java

On 26 August 2013 16:54, Ignasi <[email protected]> wrote:
> Ok, I just added a few cookbooks more to my Hosted Chef account and
> the download is failing the same way.
> I still haven't figured out what can be failing in the signature (it
> shouldn't have changed since Chef 0.10), so any help would be
> appreciated!
>
> On 26 August 2013 15:55, Ignasi <[email protected]> wrote:
>> Hi again,
>>
>> The problem appears in the "getResourceContents" api call. After
>> debugging for a while I haven't found an explanation. It fails
>> randomly with random resources, and there is no pattern for that :(
>> Sometimes it fails to download a root file resource, sometimes not.
>> Sometimes it fails downloading a recipe file, sometimes it just works.
>>
>> I executed the live test suite against a Chef 0.10.8, and against
>> Enterprise Chef and all tests are passing, so I assume it might be an
>> issue in Chef 11; otherwise I can not explain why the same resource
>> sometimes is downloaded and sometimes it is not. The authentication
>> mechanism hasn't changed, and if it works for Chef 0.10 and Enterprise
>> Chef, it should work for Chef 11 too (or at least have some
>> deterministic behavior).
>>
>>
>> To properly reflect that in the live tests, I've separated the
>> "testListCookbooks" in two different tests: "testListCookbooks" and
>> "testDownloadCookbooks". If you update my branch, you should get those
>> changes.
>>
>> I'll search if there is any open issue and file a new one if appropriate.
>>
>> Anyway, let me know if you find something that could help getting rid of 
>> this!
>>
>>
>> Ignasi
>>
>>
>> On 26 August 2013 14:29, Noorul Islam K M <[email protected]> wrote:
>>> Ignasi <[email protected]> writes:
>>>
>>>> Thanks a lot for your feedback Noorul!
>>>>
>>>> You can run the 'core' live tests as follows:
>>>>
>>>> mvn clean install -pl core -Plive \
>>>>   -Dtest.chef.endpoint=<you chef server url> \
>>>>   -Dtest.chef.identity=<identity> \
>>>>   -Dtest.chef.credential=<path to the PEM private key>
>>>>
>>>
>>> I think, I get the same as you
>>>
>>>
>>> Test suite progress: tests succeeded: 41, failed: 1, skipped: 1.
>>> Tests run: 43, Failures: 1, Errors: 0, Skipped: 1, Time elapsed: 187.134 
>>> sec <<< FAILURE!
>>>
>>> Results :
>>>
>>> Failed tests:   testListCookbooks(org.jclouds.chef.ChefApiLiveTest): <?xml 
>>> version="1.0" encoding="UTF-8"?><Error><Code>SignatureDoesNotMatch
>>> </Code><Message>The request signature we calculated does not match the 
>>> signature you provided. Check your key and signing method.</Message><S
>>> tringToSignBytes></StringToSignBytes><RequestId>g2gCZAATYm9va3NoZWxmQDEyNy4wLjAuMWgDYgAABWFiAAef4WIAC6/x</RequestId><HostId></HostId><Signatu
>>> reProvided>hRkpN1RzN 
>>> /pGeMmyXnDtRZUNnI=</SignatureProvided><StringToSign>GET(..)
>>>
>>> Tests run: 43, Failures: 1, Errors: 0, Skipped: 1
>>>
>>> Thanks and Regards
>>> Noorul
>>>
>>>> And if you have a hosted chef account, you can run them like this:
>>>>
>>>> mvn clean install -pl enterprise -Plive \
>>>>   -Dtest.enterprisechef.org=<you organization name> \
>>>>   -Dtest.enterprisechef.identity=<identity> \
>>>>   -Dtest.enterprisechef.credential=<path to the PEM private key>
>>>>
>>>>
>>>>
>>>> On 26 August 2013 14:10, Noorul Islam K M <[email protected]> wrote:
>>>>> Ignasi <[email protected]> writes:
>>>>>
>>>>>> Noorul, I've just uploaded to my branch [1] a change that registers a
>>>>>> custom Map deserializer to allow repeated keys. It just keeps the last
>>>>>> one, which seems to be the one with the right content. You can take a
>>>>>> look at the commit [2] if you want to know the details.
>>>>>>
>>>>>> I've tried the api call that was failing to you, and now it works and
>>>>>> the cookbook is properly deserialized. However, the live test to list
>>>>>> all cookbooks is still failing due to a signature error (which is
>>>>>> weird, because the signature is generated in a generic filter [3] that
>>>>>> applies to all requests).
>>>>>>
>>>>>> I think it might be related to my Chef 11 installation. To make sure,
>>>>>> could you update the branch and run the live tests, or try to list the
>>>>>> cookbooks using the ChefApi and ChefServer, and share the results?
>>>>>>
>>>>>
>>>>> I tested these methods after pulling latest from your branch.
>>>>>
>>>>> getCookbook()
>>>>> listCookbooks()
>>>>> listEnvironmentCookbooks()
>>>>>
>>>>> Everything seems to be working fine.
>>>>>
>>>>> How do I run live tests? Is this documented somewhere?
>>>>>
>>>>> Thanks and Regards
>>>>> Noorul
>>>>>
>>>>>>
>>>>>> Thanks!
>>>>>>
>>>>>>
>>>>>> Ignasi
>>>>>>
>>>>>>
>>>>>>
>>>>>>
>>>>>> [1] https://github.com/nacx/jclouds-chef/tree/245-live-tests
>>>>>> [2] 
>>>>>> https://github.com/nacx/jclouds-chef/commit/59115b5a08197547cf74f6831e265fe67a3b9ab8
>>>>>> [3] 
>>>>>> https://github.com/nacx/jclouds-chef/blob/245-live-tests/core/src/main/java/org/jclouds/chef/filters/SignedHeaderAuth.java
>>>>>>
>>>>>>
>>>>>>
>>>>>> On 26 August 2013 12:24, Ignasi <[email protected]> wrote:
>>>>>>> As soon as possible.
>>>>>>>
>>>>>>> My branch depends on issue JCLOUDS-246 [1]. Since changes are not
>>>>>>> backward compatible (just a massive package and artifactId renaming),
>>>>>>> I sent a mail last week to let users express an opinion on the change
>>>>>>> [2]. I plan to merge it this week, since there are no opinions
>>>>>>> against, and as soon as it is merged, I'll merge my current branch.
>>>>>>>
>>>>>>> You could expect to have it merged by the end of the week, if I finish
>>>>>>> the fix by then (which I hope :)). I plan to upload to my branch the
>>>>>>> fix this afternoon; I'll let you know when I have something usable so
>>>>>>> you can test and validate it.
>>>>>>>
>>>>>>> Thanks for your help!
>>>>>>>
>>>>>>>
>>>>>>> Ignasi
>>>>>>>
>>>>>>>
>>>>>>> [1] https://issues.apache.org/jira/browse/JCLOUDS-246
>>>>>>> [2] http://markmail.org/message/5jm2ivpypewnov3h
>>>>>>>
>>>>>>> On 26 August 2013 12:02, Noorul Islam K M <[email protected]> wrote:
>>>>>>>> Ignasi <[email protected]> writes:
>>>>>>>>
>>>>>>>>> Just tested the openssl cookbook and found that the returned Metadata
>>>>>>>>> has a duplicate key in the recipes list. I've pasted the json here [1]
>>>>>>>>> so you can have a look.
>>>>>>>>>
>>>>>>>>> I'll ping Chef people to see if it is an issue in their side, and try
>>>>>>>>> to build a workaround in the meanwhile.
>>>>>>>>>
>>>>>>>>
>>>>>>>> I would like to know, usually how long it takes something to get into
>>>>>>>> upstream from the branch that you are working on?
>>>>>>>>
>>>>>>>> I can provide any help needed to make this working but need to be sure
>>>>>>>> that I can use this in production as soon as possible.
>>>>>>>>
>>>>>>>> Thanks and Regards
>>>>>>>> Noorul
>>>>>>>>
>>>>>>>>>
>>>>>>>>>
>>>>>>>>> [1] http://pastie.org/8270391
>>>>>>>>>
>>>>>>>>> On 26 August 2013 08:28, Ignasi <[email protected]> wrote:
>>>>>>>>>> Hi Noorul!
>>>>>>>>>>
>>>>>>>>>> There is definitely the intention to support Chef 11 as soon as 
>>>>>>>>>> possible. I
>>>>>>>>>> have a work in progress branch [1] where I fixed all Enterprise Chef 
>>>>>>>>>> live
>>>>>>>>>> tests and also made all live tests pass against a Chef 11 except the
>>>>>>>>>> "listCookbooks" one.
>>>>>>>>>>
>>>>>>>>>> I get the exact same error than you, and it happens when 
>>>>>>>>>> deserializing the
>>>>>>>>>> json for the cookbook into the CookbookVersion object. It seems that 
>>>>>>>>>> the
>>>>>>>>>> Chef Server is returning a duplicate key in the json structure 
>>>>>>>>>> (although it
>>>>>>>>>> shouldn't because we are forcing by default a 0.10 version header to 
>>>>>>>>>> make it
>>>>>>>>>> return the same responses than in previous versions), and the json 
>>>>>>>>>> parser
>>>>>>>>>> fails to build the corresponding map.
>>>>>>>>>>
>>>>>>>>>> I'm currently working on this but it's taking longer than expected, 
>>>>>>>>>> so any
>>>>>>>>>> help would be much appreciated. If you could build my branch and 
>>>>>>>>>> enable
>>>>>>>>>> jclouds wire logging [2], you would be able to see the json returned 
>>>>>>>>>> for
>>>>>>>>>> your cookbook. Could you share that json? Also, any help 
>>>>>>>>>> tracking/isolating
>>>>>>>>>> the error (or patch :)) is appreciated too!
>>>>>>>>>>
>>>>>>>>>> Ignasi
>>>>>>>>>>
>>>>>>>>>> [1] https://github.com/nacx/jclouds-chef/tree/245-live-tests
>>>>>>>>>> [2]
>>>>>>>>>> http://jclouds.incubator.apache.org/documentation/reference/jclouds-logging/
>>>>>>>>>>
>>>>>>>>>> El 26/08/2013 07:27, "Noorul Islam K M" <[email protected]> escribió:
>>>>>>>>>>
>>>>>>>>>>>
>>>>>>>>>>> We are planning to use jclouds-chef library to talk to community 
>>>>>>>>>>> Chef
>>>>>>>>>>> Server 11. In my initial testings, I get the following exception 
>>>>>>>>>>> when I
>>>>>>>>>>> try to get cookbook details using
>>>>>>>>>>>
>>>>>>>>>>> api.getCookbook("openssl", "1.0.2")
>>>>>>>>>>>
>>>>>>>>>>>
>>>>>>>>>>> |  java.lang.IllegalArgumentException: duplicate key: openssl
>>>>>>>>>>>         at
>>>>>>>>>>> com.google.common.base.Preconditions.checkArgument(Preconditions.java:119)
>>>>>>>>>>>         at
>>>>>>>>>>> com.google.common.collect.RegularImmutableMap.<init>(RegularImmutableMap.java:67)
>>>>>>>>>>>         at
>>>>>>>>>>> com.google.common.collect.ImmutableMap$Builder.fromEntryList(ImmutableMap.java:249)
>>>>>>>>>>>         at
>>>>>>>>>>> com.google.common.collect.ImmutableMap$Builder.build(ImmutableMap.java:235)
>>>>>>>>>>>         at
>>>>>>>>>>> org.jclouds.json.internal.NullFilteringTypeAdapterFactories$MapTypeAdapter.read(NullFilteringTypeAdapterFactories.java:313)
>>>>>>>>>>>         at
>>>>>>>>>>> org.jclouds.json.internal.NullFilteringTypeAdapterFactories$MapTypeAdapter.read(NullFilteringTypeAdapterFactories.java:278)
>>>>>>>>>>>         at
>>>>>>>>>>> com.google.gson.internal.bind.ReflectiveTypeAdapterFactory$1.read(ReflectiveTypeAdapterFactory.java:93)
>>>>>>>>>>>         at
>>>>>>>>>>> com.google.gson.internal.bind.ReflectiveTypeAdapterFactory$Adapter.read(ReflectiveTypeAdapterFactory.java:172)
>>>>>>>>>>>         at
>>>>>>>>>>> com.google.gson.internal.bind.ReflectiveTypeAdapterFactory$1.read(ReflectiveTypeAdapterFactory.java:93)
>>>>>>>>>>>         at
>>>>>>>>>>> com.google.gson.internal.bind.ReflectiveTypeAdapterFactory$Adapter.read(ReflectiveTypeAdapterFactory.java:172)
>>>>>>>>>>>         at com.google.gson.Gson.fromJson(Gson.java:795)
>>>>>>>>>>>         at com.google.gson.Gson.fromJson(Gson.java:761)
>>>>>>>>>>>         at com.google.gson.Gson.fromJson(Gson.java:710)
>>>>>>>>>>>         at
>>>>>>>>>>> org.jclouds.json.internal.GsonWrapper.fromJson(GsonWrapper.java:47)
>>>>>>>>>>>         at 
>>>>>>>>>>> org.jclouds.http.functions.ParseJson.apply(ParseJson.java:86)
>>>>>>>>>>>
>>>>>>>>>>> README.md says that
>>>>>>>>>>>
>>>>>>>>>>> It currently supports versions 0.9 and 0.10 of the standard Chef 
>>>>>>>>>>> server
>>>>>>>>>>> apis, and an initial and very basic (still in progress) 
>>>>>>>>>>> implementation
>>>>>>>>>>> of the user and organization api of the Hosted and Private Chef
>>>>>>>>>>> flavours.
>>>>>>>>>>>
>>>>>>>>>>> I think that the issue I am facing is because I am using Chef 
>>>>>>>>>>> server 11.
>>>>>>>>>>> Is there any plans to officially support Chef Server 11 in near 
>>>>>>>>>>> future?
>>>>>>>>>>>
>>>>>>>>>>> Thanks and Regards
>>>>>>>>>>> Noorul

Reply via email to