Hi,
OK, strangly enough I just found the same error in a completely
different bit of code I'm evaluating.
Line 51 of ... needs to be changed to:
open DF, "df -P -k '$path' |" || die "Cannot run df: $!,";
otherwise linux splits the df output over two lines if $path or more
likely where it's mounted from is too long.
doing that has populated those keys and I can copy and register the
file.
heplnx102 - ~ $ lcg-cr --vo dteam -v -d heplnx132.pp.rl.ac.uk -l
lfn:brew-test-20050803-0935 file:`pwd`/pi.out
Using grid catalog type: edg
Source URL: file:/home/hephp/brew/pi.out
File size: 85366
Destination specified: heplnx132.pp.rl.ac.uk
Destination URL for copy:
gsiftp://heplnx132.pp.rl.ac.uk/raid/dteam/generated/2005-08-03/filedd4f3
127-0e5e-4ee8-b525-fe30159a70dc
# streams: 1
Alias registered in Catalog: lfn:brew-test-20050803-0935
Transfer took 1190 ms
Destination URL registered in Catalog:
sfn://heplnx132.pp.rl.ac.uk/raid/dteam/generated/2005-08-03/filedd4f3127
-0e5e-4ee8-b525-fe30159a70dc
guid:ffcd4d80-9696-4c47-b188-370762052b7f
Thanks for the help. Now onto the Mon box.
Chris.
> -----Original Message-----
> From: LHC Computer Grid - Rollout
> [mailto:[log in to unmask]] On Behalf Of
> Maarten Litmaath, CERN
> Sent: 02 August 2005 23:19
> To: [log in to unmask]
> Subject: Re: [LCG-ROLLOUT] More lcg-cr 2_6_0 SE-classic woes
>
> On Tue, 2 Aug 2005, Brew, CAJ (Chris) wrote:
>
> > Hi,
> >
> > I've got part way through the 2_6_0 upgrade today but this
> one's got
> > me stumped.
> >
> > I've upgraded my CE and both SE's then had a lot of
> problems getting
> > the second one to show up in the info system (the bdii
> update file's
> > moved
>
> We renamed lcg-bdii to bdii because VDT expressed interest to
> include it in their distribution, which makes the package no
> longer LCG-specific.
>
> > again). I know pass the GIIS sanity check but I cannot get
> a file onto
> > my second SE.
> >
> > heplnx102 - ~ $ lcg-cr --vo dteam -v -d heplnx132.pp.rl.ac.uk -l
> > lfn:brew-test-20050802-1931 file:`pwd`/pi.out Using grid
> catalog type:
> > edg SA Root not found
> > lcg_cr: Invalid argument
> >
> > It works ok to the first one.
> >
> > heplnx102 - ~ $ lcg-cr --vo dteam -v -d heplnx202.pp.rl.ac.uk -l
> > lfn:brew-test-20050802-1930 file:`pwd`/pi.out Using grid
> catalog type:
> > edg Source URL: file:/home/hephp/brew/pi.out File size: 85366
> > Destination specified: heplnx202.pp.rl.ac.uk Destination
> URL for copy:
> >
> gsiftp://heplnx202.pp.rl.ac.uk/storage/dteam/generated/2005-08-02/file
> > 61
> > 13d21e-1ca3-46b4-a49a-0361447b784e
> > # streams: 1
> > Alias registered in Catalog: lfn:brew-test-20050802-1930
> Transfer took
> > 370 ms Destination URL registered in Catalog:
> >
> sfn://heplnx202.pp.rl.ac.uk/storage/dteam/generated/2005-08-02/file611
> > 3d
> > 21e-1ca3-46b4-a49a-0361447b784e
> > guid:7f66f577-d440-4f60-b732-21301e66ba59
> >
> > As far as I can tell both SEs seem to but publishing the
> same info so
> > I suspect the problem is actually on the CE. Qualitatively
> I see the
> > same entries for both SEs on the site BDII but there seem
> to be many
> > more for
> > heplnx202 than heplnx132.
> >
> > Does anyone have any ideas what I've missed/misconfigured.
>
> Your site BDII on heplnx201.pp.rl.ac.uk is rejecting various
> records that SE heplnx132.pp.rl.ac.uk is publishing.
>
> Look into /opt/bdii/var/tmp/stderr.log on the CE (beware that
> the file is rewritten every two minutes):
>
> --------------------------------------------------------------
> ---------------
> [...]
> str2entry: invalid value for syntax 1.3.6.1.4.1.1466.115.121.1.27
> slapadd: could not parse entry (line=399) [...]
> --------------------------------------------------------------
> ---------------
>
> Now look at what the SE is publishing:
>
> --------------------------------------------------------------
> ---------------
> $ ldapsearch -x -h heplnx132.pp.rl.ac.uk:2135 -b
> mds-vo-name=local,o=grid [...] # alice,
> heplnx132.pp.rl.ac.uk, local, grid
> dn:
> GlueSALocalID=alice,GlueSEUniqueID=heplnx132.pp.rl.ac.uk,Mds-V
o-name=local
> ,o=grid
> objectClass: GlueSATop
> objectClass: GlueSA
> objectClass: GlueSAPolicy
> objectClass: GlueSAState
> objectClass: GlueSAAccessControlBase
> objectClass: GlueKey
> objectClass: GlueSchemaVersion
> GlueSARoot: alice:alice
> GlueSAPath: /raid/alice
> GlueSAType: permanent
> GlueSALocalID: alice
> GlueSAPolicyMaxFileSize: 10000
> GlueSAPolicyMinFileSize: 1
> GlueSAPolicyMaxData: 100
> GlueSAPolicyMaxNumFiles: 10
> GlueSAPolicyMaxPinDuration: 10
> GlueSAPolicyQuota: 0
> GlueSAPolicyFileLifeTime: permanent
> GlueSAStateAvailableSpace:
> GlueSAStateUsedSpace:
> GlueSAAccessControlBaseRule: alice
> GlueChunkKey: GlueSEUniqueID=heplnx132.pp.rl.ac.uk
> GlueSchemaVersionMajor: 1
> GlueSchemaVersionMinor: 2
> [...]
> --------------------------------------------------------------
> ---------------
>
> The GlueSAStateAvailableSpace and GlueSAStateUsedSpace
> attributes must not have empty values.
>
> This means the info provider
> /opt/lcg/libexec/lcg-info-dynamic-classicSE
> on heplnx132.pp.rl.ac.uk fails to parse the output of the
> "df" commands it executes. To debug this, first run the
> command manually:
>
> /opt/lcg/libexec/lcg-info-dynamic-classicSE \
> /opt/lcg/var/gip/lcg-info-generic.conf
>
> Does it complain? If not, try adding a few debug print statements.
>
