[log in to unmask] [mailto:[log in to unmask]] said:
> Despite appearances, the GlueSEType is _not_ used.
Oops, yes - and in fact this is a bit odd, because the old glue schema
doesn't seem to have an SEType defined, and in the new Glue it will be
moved into the new ControlProtocol object, so I don't know where that
came from or what it's supposed to be for ...
> Instead, the type is derived from the GlueSEName, which must
> be of the form:
>
> SITE_NAME:SE_TYPE
>
> I suppose this is for historical reasons?
I think it was introduced as a fudge when we decided we needed classic
SEs as a short-term stopgap and the RM needed to treat them differently
from the EDG SE. Like many fudges it lasted quite a while ... anyway:
ldapsearch -x -h lcgbdii02.gridpp.rl.ac.uk -p 2170 -b o=grid | grep -i
sename | cut -d: -f3 | sort | uniq -c
153 disk
15 srm_v1
I'm a bit surprised there are 15 SRMs out there ...
ldapsearch -x -h lcgbdii02.gridpp.rl.ac.uk -p 2170 -b o=grid | grep -i
sename |grep srm_v1 | cut -d: -f2
PIC-CASTOR-SRM
PIC-CASTOR-SRM
zibpro
DESYPRO
CERNDCACHE
PIC-CASTOR-SRM
GVMUAM-LCG2
CERN-CASTOR-SRM
DESYPRO
IFIC-CASTOR-SRM
USCMS-FNAL-WC1
CNAF-CASTOR-SRM
desypro
CNAF-CASTOR-SRM
RU-Protvino-IHEP
Stephen
|