Dear Matt,
From our configuration, on the dcache repoitory
[root@ford ~]# rpm -q -R desy-SE_dcache_info
bdii >= 3.9.1
glite-info-generic >= 2.0.1
glite-info-provider-ldap >= 1.0.0
glite-info-templates >= 1.0.0
desy-SE_dcache_pool
glue-schema >= 1.3.0
openldap-servers
rpmlib(PayloadFilesHavePrefix) <= 4.0-1
rpmlib(CompressedFileNames) <= 3.0.4-1
YAIM just works with SL3/sl4 for the static information provider. So
please do make sure that you remove all LCG info provider components.
Particularly check your BDII version.
If you are using YAIM use the desy branch as the fixed info provider
branch for SL4 is still in PPS.
Regards
Owen
On Wed, 5 Mar 2008 14:04:35 +0000
Greig Alan Cowan <[log in to unmask]> wrote:
> Matt,
>
> Can you remove all of the lcg-info packages and re-run YAIM. There's
> clearly been some confusion between /opt/lcg and /opt/glite . Which
> version of the BDII are you running? Make sure that it is the SL3 one.
> You should have this stuff:
>
> bdii >= 3.9.1
> glite-info-generic >= 2.0.1
> glite-info-provider-ldap >= 1.0.0
> glite-info-templates >= 1.0.0
> glue-schema >= 1.3.0
> openldap-servers
>
> Greig
>
> On 05/03/08 13:18, Matt Doidge wrote:
> > Thanks Greig and Chris,
> >
> > Ldapsearching against our SE gives us a "No ldap server" type error.
> > Looking at slapd processes on my SE I see nothing on port 2170 but
> > lots listening near it:
> >
> > edguser 11689 0.3 0.2 74732 5004 ? S 13:08 0:00
> > /usr/sbin/slapd -f /opt/bdii//var/2173/bdii-slapd.conf -h
> > ldap://localhost:2173 -u edguser
> > edguser 12303 0.6 0.2 76772 5000 ? S 13:08 0:00
> > /usr/sbin/slapd -f /opt/bdii//var/2171/bdii-slapd.conf -h
> > ldap://localhost:2171 -u edguser
> > edguser 13307 2.0 0.2 76780 4996 ? S 13:09 0:00
> > /usr/sbin/slapd -f /opt/bdii//var/2172/bdii-slapd.conf -h
> > ldap://localhost:2172 -u edguser
> >
> > Looking at /opt/glite/etc/gip/glite-info-generic.conf I see;
> > temp_dir = /opt/glite/var/tmp/gip
> > cache_dir = /opt/glite/var/cache/gip
> > lock_dir = /opt/glite/var/lock/gip
> > plugin_dir = /opt/glite/etc/gip/plugin
> > static_dir = /opt/glite/etc/gip/ldif
> > provider_dir = /opt/glite/etc/gip/provider
> > freshness = 60
> > cache_ttl = 300
> > response = 110
> > timeout = 150
> >
> > and looking at my packages I see:
> > glite-info-provider-ldap-1.1.0-1
> > glite-info-templates-1.0.0-8
> > glite-info-update-endpoints-1.0.0-5
> > glite-info-generic-2.0.2-3
> > glite-info-plugin-fcr-1.0.0-3
> > lcg-info-1.11.0-1
> > lcg-info-templates-1.0.15-1
> > lcg-info-provider-software-1.0.6-1
> > lcg-info-dynamic-software-1.0.3-3
> >
> > I've ran all the yaim config scripts that I thought were relevent-
> > config_gip, config_bdii and config_gip_dcache.
> >
> > cheers,
> > Matt
> >
> > On 05/03/2008, Greig Alan Cowan <[log in to unmask]> wrote:
> >> What are the contents of: /opt/glite/etc/gip/glite-info-generic.conf
> >>
> >> What packages do you have installed?
> >>
> >> glite-info-generic-2.0.2-2.noarch
> >> glite-info-templates-1.0.0-4.noarch
> >> lcg-info-dynamic-dpm-2.2-2.noarch
> >> lcg-info-generic-1.0.22-1_sl3.noarch
> >> lcg-info-templates-1.0.15-1.noarch
> >>
> >> You should also run the YAIM config_gip and config_bdii functions on the
> >> dCache head node.
> >>
> >>
> >> Greig
> >>
> >>
> >> On 05/03/08 12:55, Matt Doidge wrote:
> >> > Hello,
> >> > /opt/glite/var/tmp/gip/infoDynamicSE-plugin-dcache.ldif.5079.3861
> >> > doesn't exist. The directory /opt/glite/var/tmp/gip/ contains more
> >> > empty directories; ldif plugin provider
> >> >
> >> > In /opt/bdii/var/tmp there's a GIP.ldif which seems to contain the
> >> > site infomation and the error file:
> >> > [root@fal-pygrid-20 ~]# cat /opt/bdii/var/tmp/stderr.log
> >> > bdb_initialize: Sleepycat Software: Berkeley DB 4.2.52: (December 3, 2003)
> >> > bdb_initialize: Sleepycat Software: Berkeley DB 4.2.52: (December 3, 2003)
> >> > could not open config file "/opt/lcg/schema/ldap/SiteInfo.schema": No
> >> > such file or directory (2)
> >> > slapadd: bad configuration file!
> >> >
> >> > Thanks for the help,
> >> > Matt
> >> > On 05/03/2008, Greig Alan Cowan <[log in to unmask]> wrote:
> >> >> So, does
> >> >>
> >> >>
> >> >> /opt/glite/var/tmp/gip/infoDynamicSE-plugin-dcache.ldif.5079.3861
> >> >>
> >> >>
> >> >> exist? If so, what are it's contents? What else is in that directory?
> >> >>
> >> >> I meant to say /opt/bdii/var/tmp. There should be a GIP.ldif and
> >> >> sterr.out file in there.
> >> >>
> >> >>
> >> >> Greig
> >> >>
> >> >>
> >> >> On 05/03/08 12:10, Matt Doidge wrote:
> >> >> > Things look like their running but not properly. From the bdii logs:
> >> >> > ----------------------------------------------------
> >> >> > Wed Mar 5 12:01:46 GMT 2008
> >> >> > Sleeping for 30
> >> >> >
> >> >> > Updating DB on port 2172
> >> >> > Waiting 115 s for query results.
> >> >> >
> >> >> > Time for searches: 0 s
> >> >> > current port: 47196 - OK
> >> >> > skipping 'dn: GlueServiceUniqueID=httpg://fal-pygrid-20.lancs.ac.uk:8443/srm/managerv2,mds-vo-name=resource,o=grid'
> >> >> > Time to update DB: 1 s
> >> >> > current port: 47197 - OK
> >> >> > ldap_bind: Can't contact LDAP server (-1)
> >> >> > Time to load DB: 60 s
> >> >> > Grabbing port 2170 for 2172
> >> >> > Error for sh: /opt/glite/var/tmp/gip/infoDynamicSE-plugin-dcache.ldif.5079.3861:
> >> >> > No such file or directory
> >> >> > ==> slapadd: bad configuration file!
> >> >> > -------------------------------------------------
> >> >> >
> >> >> > I don't seem to have a /opt/bdii/tmp dir. I also looked in the
> >> >> > bdii-fwd.log and saw entries like:
> >> >> > ----------------------------------
> >> >> > remote server: IO::Socket::INET: connect: Connection refused at
> >> >> > /opt/bdii//sbin/bdii-fwd line 167.
> >> >> > 20080305_120520 [Connecting to localhost...20080305_120520 Forked
> >> >> > process 4094 -> 2171
> >> >> > 20080305_120520 Reaped process 4094 (genNr 27)
> >> >> > 20080305_120551 [Connect from 194.80.35.23:53175]
> >> >> > -----------------------------------
> >> >> > 194.80.35.23 is our site-bdii (fal-pygrid-17.lancs.ac.uk).
> >> >> >
> >> >> > cheers,
> >> >> > Matt
> >> >> >
> >> >> > On 05/03/2008, Greig Alan Cowan <[log in to unmask]> wrote:
> >> >> >> Hi Matt,
> >> >> >>
> >> >> >> Are you sure that the BDII is running on the dCache node?
> >> >> >>
> >> >> >> $ ldapsearch -LLL -x -H ldap://fal-pygrid-20.lancs.ac.uk:2170 -b
> >> >> >>
> >> >> >> mds-vo-name=resource,o=grid
> >> >> >>
> >> >> >> ldap_bind: Can't contact LDAP server (-1)
> >> >> >>
> >> >> >> $ telnet fal-pygrid-20.lancs.ac.uk 2170
> >> >> >> Trying 194.80.35.12...
> >> >> >> Connected to fal-pygrid-20.lancs.ac.uk (194.80.35.12).
> >> >> >> Escape character is '^]'.
> >> >> >> Connection closed by foreign host.
> >> >> >>
> >> >> >> Compare with:
> >> >> >>
> >> >> >> $ telnet srm.glite.ecdf.ed.ac.uk 2170
> >> >> >> Trying 129.215.95.134...
> >> >> >> Connected to srm.glite.ecdf.ed.ac.uk (129.215.95.134).
> >> >> >> Escape character is '^]'.
> >> >> >> ^]
> >> >> >> telnet> Connection closed.
> >> >> >>
> >> >> >> What do the BDII logs in /opt/bdii/var/bdii.log and in the bdii/tmp dir say?
> >> >> >>
> >> >> >> Cheers,
> >> >> >>
> >> >> >> Greig
> >> >> >>
> >> >> >>
> >> >> >> On 05/03/08 11:48, Matt Doidge wrote:
> >> >> >> > Hello,
> >> >> >> >
> >> >> >> > Thanks for all the pointers, it really clarified things. I'm almost
> >> >> >> > there now. Running /opt/glite/libexec/glite-info-wrapper gives me all
> >> >> >> > a lot of text that looks right, plus all the bdii gubbins seem to be
> >> >> >> > running. We just don't seem to be pushing info to the site bdii. I
> >> >> >> > poked relevant holes in both firewalls, and editied the
> >> >> >> > bdii-update.conf on the site bdii to have the line:
> >> >> >> > SRM ldap://fal-pygrid-20.lancs.ac.uk:2170/mds-vo-name=resource,o=grid
> >> >> >> >
> >> >> >> > One clue from the logs on our site bdii:
> >> >> >> > SRM: ldap_bind: Can't contact LDAP server
> >> >> >> >
> >> >> >> > Is there something else I should be running on the SRM node? Have I
> >> >> >> > missed something obvious?
> >> >> >> >
> >> >> >> > cheers,
> >> >> >> > Matt
> >> >> >> > On 04/03/2008, Greig Alan Cowan <[log in to unmask]> wrote:
> >> >> >> >> Hi Matt,
> >> >> >> >>
> >> >> >> >> Yes, things have all changed. Most things should be under /opt/glite now
> >> >> >> >> and you should be using the BDII on port 2170 as your information
> >> >> >> >> provider; globus-mds has now been retired. You should be using
> >> >> >> >> mds-vo-name=resource (making sure that your site BDII now points to this
> >> >> >> >> new endpoint).
> >> >> >> >>
> >> >> >> >> To understand the new system, you just need to follow the configuration
> >> >> >> >> files. Everything is controlled by
> >> >> >> >> /opt/glite/libexec/glite-info-wrapper. This calls all of the information
> >> >> >> >> providers and plugins. The configuration file that the info-wrapper uses
> >> >> >> >> contains pointer to where all of the plugins and providers live.
> >> >> >> >>
> >> >> >> >> Hope that all makes sense.
> >> >> >> >>
> >> >> >> >> Cheers,
> >> >> >> >>
> >> >> >> >> Greig
> >> >> >> >>
> >> >> >> >>
> >> >> >> >> On 04/03/08 21:21, Matt Doidge wrote:
> >> >> >> >> > 11 hours after putting the SL4 CD in the tray... (okay I did treat
> >> >> >> >> > myself to a long lunch break)
> >> >> >> >> >
> >> >> >> >> > And I still haven't cracked it. I *think* we just have the infosystem
> >> >> >> >> > to crack now. But I've been trying to crack it for an age. I think
> >> >> >> >> > I've got the similar problems to what the chaps at Liverpool had, this
> >> >> >> >> > SL4 beast feels very different from the SL3 critter that died
> >> >> >> >> > yesterday. What happened to globus-mds? Should I have edguser or
> >> >> >> >> > edginfo? Why did I end up installing a bdii? Why is nothing in
> >> >> >> >> > /opt/lcg/ anymore? I don't like change.
> >> >> >> >> >
> >> >> >> >> > I'm suffering from staring at the same problem for too long, but I
> >> >> >> >> > sure could do with some help tomorrow. But first food, and precious
> >> >> >> >> > sleep.
> >> >> >> >> >
> >> >> >> >> > cheers,
> >> >> >> >> > Matt
> >> >> >> >> >
> >> >> >> >> > On 04/03/2008, Greig Alan Cowan <[log in to unmask]> wrote:
> >> >> >> >> >> Arrrgh. Too bad Matt - there's always something with you guys!
> >> >> >> >> >>
> >> >> >> >> >> At least things were backed up so nothing much will have been lost. Let
> >> >> >> >> >> me know if you need a hand.
> >> >> >> >> >>
> >> >> >> >> >> Cheers,
> >> >> >> >> >>
> >> >> >> >> >> Greig
> >> >> >> >> >>
> >> >> >> >> >>
> >> >> >> >> >> On 04/03/08 10:48, Matt Doidge wrote:
> >> >> >> >> >> > Hello,
> >> >> >> >> >> >
> >> >> >> >> >> > This is just a heads up to the powers that be- we had an unexpected
> >> >> >> >> >> > power cut yesterday that ended up crunching the system disk of our
> >> >> >> >> >> > dcache SRM. We've tried to bring back the lost data from beyond the
> >> >> >> >> >> > grave but without success. Luckily the majority of our configs are
> >> >> >> >> >> > backed up, but we're likely to be down a fair chunk of the day.
> >> >> >> >> >> >
> >> >> >> >> >> > On the plus side, we finally got a long overdue upgrade to SL4 for our
> >> >> >> >> >> > head node... You've got to think positive in this business!
> >> >> >> >> >> >
> >> >> >> >> >> > cheers,
> >> >> >> >> >> > Matt
> >> >> >> >> >>
> >> >> >> >>
> >> >> >>
> >> >>
> >>
|