Hi,

arcproxy doesn't do any randomisation, it just picks the servers in the order they are specified in the vomses list. I suppose randomisation can be added, if really needed. With arcproxy it should be possible to specify a different vomses file, for your testing purposes, though I didn't try it myself (option -s, I believe).

Cheers,
Oxana

On 11.12.2013 11:39, Sam Skipsey wrote:
> I don't know if arcproxy has better randomness in its VOMS server
> selection (obviously, it's a less standard tool for those of us not
> using the NorduGrid middleware)?
>
> Sam
>
> On 11 December 2013 10:30, Kashif Mohammad <[log in to unmask]> wrote:
>> Hi Steve
>>
>> Thanks,  I tried it myself before posting on LCG-ROLLOUT and it was picking the same VOMS server.  It means that it is supposed to pick random VOMS server but in practice it is not happening.
>> So I should not rely on randomness of voms-client.
>> Cheers
>> Kashif
>>
>>> -----Original Message-----
>>> From: LHC Computer Grid - Rollout [mailto:[log in to unmask]]
>>> On Behalf Of Stephen Jones
>>> Sent: 11 December 2013 10:16
>>> To: [log in to unmask]
>>> Subject: Re: [LCG-ROLLOUT] voms-proxy-init default order for vomses
>>>
>>> On 12/10/2013 09:07 PM, Maarten Litmaath wrote:
>>>> The selection is randomized, but I have often seen long sequences of
>>>> the same endpoint getting picked during ~1 minute. Also check if you
>>>> do not have a ~/.glite/vomses or ~/.voms/vomses file or directory that
>>>> might be taking precedence.
>>>
>>> One last thing. The randomising works fine on C++ version of voms-proxy-init
>>> (e.g. voms-clients-2.0.9-1...). But it's very poor/totally useless on the Java
>>> version (voms-clients3-3.0.2-1...) .
>>> You can get twenty++ runs without any changes to the vomses file selected.
>>> In fact, it always prefers one, and never selected the other when I tried it
>>> out. Plus the java takes 10 seconds just to load.
>>>
>>> My advice: avoid random selection on Java version.
>>>
>>> Steve
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>> --
>>> Steve Jones                             [log in to unmask]
>>> System Administrator                    office: 220
>>> High Energy Physics Division            tel (int): 42334
>>> Oliver Lodge Laboratory                 tel (ext): +44 (0)151 794 2334
>>> University of Liverpool                 http://www.liv.ac.uk/physics/hep/