Print

Print


Unfortunately the security code is encoded so deeply into the server 
code, that even if you will submit unencrypted message, server will not 
be able to process it. User info will be null, and many decisions will 
not be possible to make.

So I suggest you get user grid credentials and use grid-proxy-init to 
create a proxy.

Thanks,
Timur
Fabrizio Furano wrote:

> Hi Artem,
>
>  well, I fear that I am not able to do what I need without installing 
> a world of packages.
>
>  Here is the latest output I get:
>
> ---------
> fabrizio@nbbbrrepro2 16:22:49 ~/Park/JavaSRM/srmclient>bin/srmcp 
> file:////bin/sh srm://nbbbrrepro2:8443//dir1/dir2/sh-copy
>
> org.globus.gsi.GlobusCredentialException: Proxy file 
> (/home/fabrizio/k5-ca-proxy.pem) not found.
>         at 
> org.globus.gsi.GlobusCredential.<init>(GlobusCredential.java:93)
>         at 
> org.dcache.srm.security.SslGsiSocketFactory.createUserCredential(SslGsiSocketFactory.java:305) 
>
>         at 
> org.dcache.srm.security.SslGsiSocketFactory.createUserCredential(SslGsiSocketFactory.java:351) 
>
>         at 
> gov.fnal.srm.util.SRMClient.getGssCredential(SRMClient.java:255)
>         at gov.fnal.srm.util.SRMClient.connect(SRMClient.java:203)
>         at gov.fnal.srm.util.SRMPutClient.connect(SRMPutClient.java:152)
>         at gov.fnal.srm.util.SRMDispatcher.work(SRMDispatcher.java:436)
>         at gov.fnal.srm.util.SRMDispatcher.main(SRMDispatcher.java:200)
> srm client error: org.globus.gsi.GlobusCredentialException: Proxy file 
> (/home/fabrizio/k5-ca-proxy.pem) not found.
> ---------
>
>
>  Is there any easy way to avoid all this? Since I am not interested in 
> testing the authentication stuff, cannot I send formatted get/put 
> requests to the server?
>
>  Fabrizio
>
>
> Artem Trunov wrote:
>
>> Hi, Fabrizio!
>>
>> it's srm's port for incoming requests.
>>
>> Artem.
>>
>> On Fri, 13 Jan 2006, Fabrizio Furano wrote:
>>
>>
>>> Hi Artem,
>>>
>>> Artem Trunov wrote:
>>>
>>>> hi, Fabrizio!
>>>>
>>>>
>>>>
>>>>>> Yuo can also do usefull stuff with default protocol. Yuo can test 
>>>>>> how your
>>>>>> srm interacts with xrootd(s) and what URLs it gives out for get/put
>>>>>> requests.
>>>>>
>>>>>
>>>>> Well, this is my intention but I have no ideas. How can I send a 
>>>>> request
>>>>> to the srm server? Possibly bypassing the scripts.
>>>>
>>>>
>>>>
>>>> yuo can try srmcp from srmclient package - this is the commandline srm
>>>> copy tool.
>>>>
>>>> srmcp srm://yourserver:8843/path file:/tmp/test1
>>>>
>>>
>>> Well, that's one of the scripts I'd like to bypass.
>>> Anyway, what's that 8843 port number? Is it needed to contact the srm
>>> server in the machine "yourserver" ?
>>>
>>>
>>>
>>> Fabrizio
>>>
>>>
>>>
>>>
>>>> I hope it could give some usefull result to you before bumping into 
>>>> luck
>>>> of grid infrastructure. Although, I am doubtfull, since it needs
>>>> authorization on the first place... Yuo can also try soap messages
>>>> directly :( . Timur may give more usufull advise.
>>>>
>>>> Artem.
>>>>
>>>>
>>>>
>>>>>> Yuo don't need to do the actual transfers? I guess. I will set
>>>>>> up a testbed with classical stogage element and validate the 
>>>>>> transfer.
>>>>>>
>>>>>
>>>>> In principle I wrote teh java code also for that. I totally ignore 
>>>>> if it
>>>>> works. I don't want to give away code written but *never* run.... 
>>>>> so I
>>>>> believe you have to wait some more time... sorry. Anyway the 
>>>>> deployment
>>>>> is not that hard. But definitely you need postgres.
>>>>>
>>>>> Fabrizio
>>>>>
>>>>>
>>>>>
>>>>>> Artem.
>>>>>>
>>>>>> On Fri, 13 Jan 2006, Fabrizio Furano wrote:
>>>>>>
>>>>>>
>>>>>>
>>>>>>
>>>>>>> Hi Timur,
>>>>>>>
>>>>>>> I just managed to get my custom srm server started!
>>>>>>>
>>>>>>> But now I have no idea about what to do. Is there a way to 
>>>>>>> inhoculate
>>>>>>> get/put requests directly to the server to debug it?
>>>>>>>
>>>>>>> I gave a look at the scripts in the srmclient directory, but I 
>>>>>>> don't
>>>>>>> believe that they are the answer. Moreover, the protocol 
>>>>>>> matchings are
>>>>>>> done inside the scripts, so I believe I'd need to modify them 
>>>>>>> all to
>>>>>>> include a new protocol.
>>>>>>>
>>>>>>> Thank you
>>>>>>>
>>>>>>> Fabrizio
>>>>>>>
>>>>>