> -----Original Message-----
> From: Testbed Support for GridPP member institutes [mailto:TB-
> [log in to unmask]] On Behalf Of Winnie Lacesso
>
> I can't replicate this on pplxint8 (Oxford UI - the only one I have access
> to) as it *always* asks lcg-voms.cern.ch never lcg-voms2.cern.ch
> (which seems odd too)
>
Hmm. Looks like it's got them all configured, though I guess we could probably drop the old ones now. I suspect it will actually fail over once they stop responding to it though.
> lacesso@pplxint8> voms-proxy-init --voms lcg-voms2.cern.ch:cms
> /* is that syntax wrong? What's the right syntax? */
> Enter GRID pass phrase for this identity:
> VOMS server for VO lcg-voms2.cern.ch is not known! Check your vomses
> configuration.
>
That's not the same failure, and I don't think that is the right syntax. I think to force it you'd have to create a new vomses directory with just the config for the one you want to try and point it at that, so something like:
mkdir testvomses
cp /etc/vomses/cms-lcg-voms2.cern.ch testvomses/
voms-proxy-init -vomses ./testvomses/ --voms cms
That appears to work for me, including actually contacting the new VOMS server, which then (of course) correctly refuses to give me a CMS proxy.
> Anyone else seeing this? I'm wondering if problem is in my UI config
> (vo.d/cms) or with lcg-voms2.cern.ch.
>
Is it still doing this? Id there any chance that lcg-voms2 was just down when you tried it, but back now?
Ewan
|