On 3 August 2012 12:14, Govind Songara <[log in to unmask]> wrote:
> Hi Sam,
>
> Yes , we do have some CV stuffs.
> I thought if munge key same accross CE & Batch server (which is ce3 here)
> and WN, then there should not be any issue.?
Yes, as you note, your batch system is working :D
The "errors" appear to be coming from *Bright* trying to talk to your
batch system; the fact that it can't, because it doesn't have a munge
key, is probably a feature, given what I've heard about Bright.
(Bright *does* try to talk to the batch system to do status updates,
amongst other things. If you want it to be able to do that, then you
need to give the Bright server the shared key. If you don't then just
put up with the error messages. Or uninstall Bright. One of those
options.)
Sam
> So how to get rid of this?
>
> BTW, why do we need pbsnodes (i.e torque-client) on WN ?
>
> Thanks
> Govind
>
>
> On Fri, Aug 3, 2012 at 11:59 AM, Sam Skipsey <[log in to unmask]>
> wrote:
>>
>> Hi Govind,
>>
>> On 3 August 2012 11:53, Govind Songara <[log in to unmask]> wrote:
>> > Hi,
>> >
>> > I have setup a EMI-cream with glite3.2 worker nodes.
>> > Jobs are running fine.
>> >
>> > From worker nodes cron (/opt/edg/sbin/edg-pbs-knownhosts), keep getting
>> > emails
>> > with message /usr/bin/pbsnodes: Invalid credential
>> >
>> > WN & CE sharing same key
>> >
>> > debug command says
>> > [root@node002 ~]# pbsnodes
>> > pbs_connect using default server name list "ce3.ppgrid1.rhul.ac.uk"
>> > pbs_connect attempting connection to server "ce3.ppgrid1.rhul.ac.uk"
>> > pbs_connect: Successful connection to server "ce3.ppgrid1.rhul.ac.uk",
>> > fd =
>> > 1
>> > pbsnodes: Invalid credential
>> >
>> > debug server_log says
>> > 08/02/2012 19:24:01;0100;PBS_Server;Req;;Type
>> > AlternateUserAuthentication
>> > request received from [log in to unmask], sock=11
>> > 08/02/2012 19:24:01;0100;PBS_Server;Req;;Type StatusNode request
>> > received
>> > from [log in to unmask], sock=11
>> > 08/02/2012 19:24:01;0080;PBS_Server;Req;req_reject;Reject reply
>> > code=15021(Invalid credential), aux=0, type=StatusNode, from
>> > [log in to unmask]
>> >
>> > Any sugesstions?
>> >
>>
>> So, what is "master.bright.office"? Are you, by any chance, running
>> the ClusterVision Bright management system? It looks like that service
>> is trying to auth to the batch system and can't (presumably because it
>> doesn't have a Munge key?)
>>
>> Sam
>>
>> >
>> >
>> > Thanks
>> > Govind
>
>
|