FW: [bug #20804] GlueCEInfoContactString should be a real URL
Hi, Does anyone (OSG particularly) have a view on what you would like to see in the ContactString for a glite CE? Stephen -----Original Message----- From: Apache [mailto:apache@lcgapp01.cern.ch] On Behalf Of noreply [Gergely Debreczeni] Sent: 27 March 2007 16:13 To: Robert Harakaly; Laurence; Burke, S (Stephen); Gergely Debreczeni; project-eu-egee-middleware-iteam@cern.ch; project-eu-egee-glite-all-bugs@cern.ch Subject: [bug #20804] GlueCEInfoContactString should be a real URL This is an automated notification sent by LCG Savannah. It relates to: bugs #20804, project gLite Middleware ======================================================================== ====== LATEST MODIFICATIONS of bugs #20804: ======================================================================== ====== Update of bug #20804 (project jra1mdw): Status: None => Accepted _______________________________________________________ Follow-up Comment #1: So, is there any agreement on the format ? Gergo ======================================================================== ====== OVERVIEW of bugs #20804: ======================================================================== ====== URL: <http://savannah.cern.ch/bugs/?20804> Summary: GlueCEInfoContactString should be a real URL Project: gLite Middleware Submitted by: sburke Submitted on: 2006-10-17 16:36 Status: Accepted Open/Closed: Open Category: Installation and configuration (deprecated) Severity: 4 - Normal Baseline Release: gLite 3.0.0 OS: None Architecture: None Bug detection area: Production Assigned to: harakaly Priority: 5 - Enhancement GGUS reference URL: Component tag(s): Subsystem tag(s): Discussion Lock: Any Build environment: None Release: _______________________________________________________ The ContactString was introduced in the schema at the request of OSG as they would like a "real" gram URL and not just the CEId with e.g. the queue name on the end. At the moment it seems that we are in fact just duplicating the CEId. For LCG CEs at least it would be easy enough to manipulate this into a proper URL. For glite CEs it's less obvious as there is no real URL for condor-C, but we could invent a format. _______________________________________________________ Follow-up Comments: ------------------------------------------------------- Date: 2007-03-27 15:13 By: Gergely Debreczeni <gdebrecz> So, is there any agreement on the format ? Gergo _______________________________________________________ Carbon-Copy List: CC Address | Comment ------------------------------------+----------------------------- 857 | -COM- 570 | -UPD- 761 | -SUB- ======================================================================== ====== This item URL is: <http://savannah.cern.ch/bugs/?20804> _______________________________________________ Message sent via/by LCG Savannah http://savannah.cern.ch/
Hi, I would like for the GlueCEInfoContactString to be the GRAM-URL for the resource. This makes it possible to use this value out-of-the-box to submit jobs with globus-job-run and condor-g (GlobusResource attribute). The current format of GlueCEInfoContactString is <GRAM-URL>-<Batch-Queue-Name>. I understand that this string is parsed by the EGEE software stack and passed correctly to the underlying condor layers. Today, we need to extract <GRAM-URL> from the GIP value GlueCEInfoContactString in order to make the value useful in OSG. Ideally, I would not want to have to do this parsing. Then again, we have to make sure that we do not break interoperability. Cheers Gabriele Burke, S (Stephen) wrote:
Hi,
Does anyone (OSG particularly) have a view on what you would like to see in the ContactString for a glite CE?
Stephen
-----Original Message----- From: Apache [mailto:apache@lcgapp01.cern.ch] On Behalf Of noreply [Gergely Debreczeni] Sent: 27 March 2007 16:13 To: Robert Harakaly; Laurence; Burke, S (Stephen); Gergely Debreczeni; project-eu-egee-middleware-iteam@cern.ch; project-eu-egee-glite-all-bugs@cern.ch Subject: [bug #20804] GlueCEInfoContactString should be a real URL
This is an automated notification sent by LCG Savannah. It relates to: bugs #20804, project gLite Middleware
======================================================================== ====== LATEST MODIFICATIONS of bugs #20804: ======================================================================== ======
Update of bug #20804 (project jra1mdw):
Status: None => Accepted
_______________________________________________________
Follow-up Comment #1:
So, is there any agreement on the format ?
Gergo
======================================================================== ====== OVERVIEW of bugs #20804: ======================================================================== ======
URL: <http://savannah.cern.ch/bugs/?20804>
Summary: GlueCEInfoContactString should be a real URL Project: gLite Middleware Submitted by: sburke Submitted on: 2006-10-17 16:36 Status: Accepted Open/Closed: Open Category: Installation and configuration (deprecated) Severity: 4 - Normal Baseline Release: gLite 3.0.0 OS: None Architecture: None Bug detection area: Production Assigned to: harakaly Priority: 5 - Enhancement GGUS reference URL: Component tag(s): Subsystem tag(s): Discussion Lock: Any Build environment: None Release:
_______________________________________________________
The ContactString was introduced in the schema at the request of OSG as they would like a "real" gram URL and not just the CEId with e.g. the queue name on the end. At the moment it seems that we are in fact just duplicating the CEId. For LCG CEs at least it would be easy enough to manipulate this into a proper URL. For glite CEs it's less obvious as there is no real URL for condor-C, but we could invent a format.
_______________________________________________________
Follow-up Comments:
------------------------------------------------------- Date: 2007-03-27 15:13 By: Gergely Debreczeni <gdebrecz> So, is there any agreement on the format ?
Gergo
_______________________________________________________
Carbon-Copy List:
CC Address | Comment ------------------------------------+----------------------------- 857 | -COM- 570 | -UPD- 761 | -SUB-
======================================================================== ======
This item URL is: <http://savannah.cern.ch/bugs/?20804>
_______________________________________________ Message sent via/by LCG Savannah http://savannah.cern.ch/
_______________________________________________ glue-wg mailing list glue-wg@ogf.org http://www.ogf.org/mailman/listinfo/glue-wg
Gabriele Garzoglio [mailto:garzogli@fnal.gov] said:
I would like for the GlueCEInfoContactString to be the GRAM-URL for the resource.
For the LCG CE that's OK, but the glite CE doesn't have a gram interface (apart from fork on the CE headnode). We can either invent a URL for that or leave it unset. In the future it's possible that some glite CEs will have gram as well, but at the moment it isn't supported. Stephen
Hi, having a URL for each different syntax is probably a good approach. Cheers Gabriele Burke, S (Stephen) wrote:
Gabriele Garzoglio [mailto:garzogli@fnal.gov] said:
I would like for the GlueCEInfoContactString to be the GRAM-URL for the resource.
For the LCG CE that's OK, but the glite CE doesn't have a gram interface (apart from fork on the CE headnode). We can either invent a URL for that or leave it unset. In the future it's possible that some glite CEs will have gram as well, but at the moment it isn't supported.
Stephen
Hi, I seem to have lost track of this ... we still need to say what the (pseudo-)URL should be for the glite CE. Something like glitece://host.name/blah-pbs? We need a scheme name which isn't going to clash with anything else, and offhand I can't think of anything better - the interface is actually condor-C but that might be needed for other things. Stephen ________________________________ From: glue-wg-bounces@ogf.org [mailto:glue-wg-bounces@ogf.org] On Behalf Of Gabriele Garzoglio Sent: 28 March 2007 17:03 To: Burke, S (Stephen) Cc: glue-wg@ogf.org Subject: Re: [glue-wg] FW: [bug #20804] GlueCEInfoContactString should be a real URL Hi, having a URL for each different syntax is probably a good approach. Cheers Gabriele Burke, S (Stephen) wrote: Gabriele Garzoglio [mailto:garzogli@fnal.gov] said: I would like for the GlueCEInfoContactString to be the GRAM-URL for the resource. For the LCG CE that's OK, but the glite CE doesn't have a gram interface (apart from fork on the CE headnode). We can either invent a URL for that or leave it unset. In the future it's possible that some glite CEs will have gram as well, but at the moment it isn't supported. Stephen
Hi, I don't think that we are very resilient to any changes, at this point, but we have to take a decision. By the way, is the lcg broker going to distinguish the new attribute value and not crash? glitece:// seams appropriate to me. Cheers Gabriele Burke, S (Stephen) wrote:
Hi,
I seem to have lost track of this ... we still need to say what the (pseudo-)URL should be for the glite CE. Something like glitece://host.name/blah-pbs? We need a scheme name which isn't going to clash with anything else, and offhand I can't think of anything better - the interface is actually condor-C but that might be needed for other things.
Stephen
------------------------------------------------------------------------ *From:* glue-wg-bounces@ogf.org [mailto:glue-wg-bounces@ogf.org] *On Behalf Of *Gabriele Garzoglio *Sent:* 28 March 2007 17:03 *To:* Burke, S (Stephen) *Cc:* glue-wg@ogf.org *Subject:* Re: [glue-wg] FW: [bug #20804] GlueCEInfoContactString should be a real URL
Hi, having a URL for each different syntax is probably a good approach. Cheers Gabriele
Burke, S (Stephen) wrote:
Gabriele Garzoglio [mailto:garzogli@fnal.gov] said:
I would like for the GlueCEInfoContactString to be the GRAM-URL for the resource.
For the LCG CE that's OK, but the glite CE doesn't have a gram interface (apart from fork on the CE headnode). We can either invent a URL for that or leave it unset. In the future it's possible that some glite CEs will have gram as well, but at the moment it isn't supported.
Stephen
Gabriele Garzoglio [mailto:garzogli@fnal.gov] said:
By the way, is the lcg broker going to distinguish the new attribute value and not crash?
I believe it uses the CEUniqueID - which is of course a mistake since UniqueIDs aren't supposed to be parsed ...
glitece:// seams appropriate to me.
I'll pass that back in the bug as a suggestion unless anyone else has any thoughts. Stephen
participants (2)
-
Burke, S (Stephen) -
Gabriele Garzoglio