DCache FAQs

Please note: This documentation is for OSG 1.2. While we still provide critical security updates for OSG Software 1.2, we recommend you use OSG Software 3 for any new or updated installations. We are considering May 31, 2013 as possible OSG 1.2 End of Life (EOL).

Review Passed
by NehaSharma
by NehaSharma

Can I use dccp to copy files in to reserved area?

Not at present. Developers are working on it.

Is there a command line tool available for reading a file stored on dCache or concatenating individual files stored on dCache?

No. In order to perform file operations (reading, concatenating, etc), you must copy files locally, manipulate/use them and copy them back to storage location.

Can I find out pool name for a file using dCap?

No. There is no way to find out pool name on client side.

How do I clean up files belonging to non-existent pools?

Recommended way is to clean out all the replicas from the companion DB via postgres, then run "pnfs register" on each pool via the admin interface.

I accidently deleted the parent directory,how can I change parent id of files in PNFS?

Send an email to osg-storage@opensciencegrid.org to get the solution. It is not posted here because of its critical nature.

Can I create soft links on SRM server as a non-admin?

No. Only Storage Element Admin can do so.

How can I install Fermi srmclient package?

Use the "pacman -get OSG:client" command.

As a SE admin, how can I set up group access on a storage location?

This can be obtained by having a few users belong to a special group/role within your VO and then using standard unix like permissions on the storage locations in question for the user to which this VO/group/role gets mapped to.

What happens to data stored under an expired Space Reservation?

At present, nothing UNLESS you explicitly run a space cleaner tool which cleanups expired space reservations.

Do pool nodes need host certificates? Why?

Yes. Credentials are delegated to pools so that they can act as clients when receiving data in pull mode or delivering
data in push mode. See slides 14 and 16 of https://srm.fnal.gov/twiki/pub/SrmProject/SrmPresentations/SRMDesign.pdf. Therefore in these cases they
need host certificates.

Some of the gridftp doors (which are on both public and private network) are returning private ip address to SRM server. How can I get around it?

Following different approaches have worked for admins in the past. However, you are responsible for making sure that this does not break other things on the system

  • In /etc/hosts, map both the public and private host names to the public IP address.
  • Swap IP address definitions for eth0 and eth1.
  • Check that the public/private NICs occupy the same slots of the PCI bus on the doors that don't work as they do on the doors that do work and physically swap them if different.

What do the terms "Free Space","Available Space" and "Reserved Space" mean?

Free Space - amount of available pool space for link group
Reserved Space - amount of space you reserved
Available Space - amount of space available for reservation
(Available space=Free Space-Reserved Space)

How can I setup Opportunistic Storage on my SE?

Two links may be useful to you: Opportunistic Storage Use and Opportunistic Storage Model for USCMS

Do I need to maintain same dCache version on all my SE nodes (admin, PNFS, pool etc)?

Yes, we highly recommend it. There have been cases of slow response from commands like srmcp, srmls when nodes were running different dCache versions.

How do I get around "DB Encoding" problem when trying to upgrade Postgres Chimera DB from 8.3.9 to 9.0.2 ? (Contributed by Shawn McKee)

Referring to the following sections of the Postgresql 9.0 documentation:

http://www.postgresql.org/docs/9.0/interactive/runtime-config-client.html (scroll to “bytea_output(enum)”) http://www.postgresql.org/docs/9.0/interactive/datatype-binary.html

“The bytea type supports two external formats for input and output: PostgreSQL?'s historical "escape" format, and "hex" format. Both of these are always accepted on input. The output format depends on the configuration parameter bytea_output; the default is hex. (Note that the hex format was introduced in PostgreSQL? 9.0; earlier versions and some tools don't understand it.) “

So try this on the Postgresql 9.0.2 version of the Chimera DB:

[root@head02 ~]# env PGOPTIONS='-c bytea_output=escape' /usr/pgsql-9.0/bin/psql -p 5431 -U postgres

psql (9.0.2)

Type "help" for help.

postgres=# \c chimera

You are now connected to database "chimera".

chimera=# select * from t_tags_inodes limit 2;

itagid | imode | inlink | iuid | igid | isize | ictime | iatime | imtime |ivalue E33177DF?-28CB-4A23-9950-41E37A6F0B3A | 33188 | 1 | 0 | 0 | 18 | 2 009-05-04 18:51:41.628-04 | 2009-05-04 18:51:41.628-04 | 2009-05-04 18:51:41.628-04 | StoreName? usatlas\012 F6B24796?-DED1-4295-81CF-673A4099D716 | 33188 | 1 | 0 | 0 | 18 | 2 009-05-04 18:52:03.467-04 | 2009-05-04 18:52:03.467-04 | 2009-05-04 18:52:03.467 -04 | StoreName? usatlas\012

(2 rows)

Things seem to work now. I just need to find out how to set this “DB” wide (perhaps in postgresql.conf). Yes there is a setting:

#bytea_output = ‘hex’

Changing this to

bytea_output = ‘escape’

And restarting Postgresql 9.0.2 seems to work

Topic revision: r14 - 15 Jun 2016 - 17:02:32 - ElizabethChism
Hello, TWikiGuest


Installation and Update Tools


Compute Element

Storage Element

Other Site Services

VO Management

Software and Caches

Central OSG Services

Additional Information

FaceBook_32x32.png Facebook

TWiki | Report Bugs | Privacy Policy

This site is powered by the TWiki collaboration platformCopyright by the contributing authors. All material on this collaboration platform is the property of the contributing authors..