[part of a series of meetings]

Participants

Patrick, Antje, Daniel, Paul, Tigran, Christian, Karsten, Tanja, Al

Agenda

[see box on the right-hand side]

Postcards

Up to two minutes (uninterrupted) per person where they can answer two questions:

• What I did last week (since the last meeting),
• What I plan to do in the next week.

No questions until we get through everyone :)

Tigran: 1.9.12-16 release, etc ...

Christian: Vacation (nice), getting Debian building to work locally, ETICS is as usual.

Karten: IPv6 work (with some success!)

Tanja: "Vacation" (not so nice :-( Patch to display billing plots .. plan this week is to get it to run without httpd.

Al: Wrapped up loose ends on XACML and committed it; will commit other patch soon. investigating processing checksum alarms.

Gene: nothing to report.

Dmitry: same as last week: preparing Enstore for Chimera migration. Prep. work is going fine. Ancillary parts have hard-coded PNFS utility. Plan is to do the migration in March. Project on event-based monitoring. Try it out with checksum mismatch on the pool. Also got an account on GGUS .. spent a little bit of time on SRM scheduling. Forwarded announcement to Tanja of workshop to OSG for redistribution.

Patrick: mostly organising dCache workshop in Zuiten, program committee working ...

Daniel: Preparing to leave for CERN on Monday.

Antje: We have the new 2.1 release --- headlines had to be updated .. 16--18 files. Scripted up this process. Release notes generated.

Paul: [...]

Plans for patch-releases

Should we make a new patch release?

1.9.12-16 just released (today !)

We have 2.0.3 on test instance.

Then we have a 2.1

1.9.5 ? Still in support, so we can do one (last ?) build.

Starting 1st March there will be no further updates to 1.9.5-series.

Trunk activity

Progress with new features...

backwards compat: IPv6 =

There is a small problem uncovered by Karsten .. we have 'world-net' (with IPv4 format). If you come with an IPv6 address then this doesn't match.

One solution is to introduce a rule that a "match-all" rule for IPv4 also works for IPv6.

0.0.0.0/0.0.0.0 or 0.0.0.0/0 or 1.1.1.1/0.0.0.0 or ... all are catch-all, so we would need to be careful.

psu create unit world-net -net 0.0.0.0/0.0.0.0

One approach is to tell users that they need to update their configure when enabling IPv6.

Another is that we do our best to provide same functionality with the old syntax.

IPv6 catch-all is "::/0" this gave problems inside the batch file (default pool-manager configuration).

dCache compiled with Java v7 --- "uncompilable source code" -- could be to do with Enums -- but it looks like it's something to do with Berkeley DB code.

JMS and event-based monitoring

Idea is to monitor the checksum calculating code in the pool. This event is sent somewhere.

The administrator can clear the alarm in a web-admin.

Why not build in in the logging level? Just use log appender.

The underlying concern is that there are a bunch of scripts that look into pool messages. We want to make these internal.

pool + PNFS-ID + string (human-readable msg).

This could be done using current logging (in logback.xml).

Currently the pool runs checksum and logs these to a file.

Dmitry & Al to investigate using logback to handle the messages.

What is the destination of these messages? A new component that would accept these; please look into adding this to our new web-admin.

Daniel

Patrick

Work-group = Storm + DPM + dCache. Trying to implement a redirector using HTTP.

Daniel, is going to CERN for 5 weeks to work with Ricardo and Fabritico to work on the redirector part.

News

Antje

Every week: asking for news and announcements because she is generating a news-letter and post news on the web-pages (RSS google+ twitter).

Reoccurring news item.

Topic is anything that customers might find interesting.

For example:

1.9.5 is dropping out of maintenance new feature in a feature release. dCache workshop developers having a workshop and want feedback.

Issues from [FIXME: Add link to yesterday's Tier-1 meeting]

Very quiet.

Issues from EMI

srmclient: will we provide a Debian package? Yes.

License

Can we use dCache.org or "the dCache collaboration" as the owner of the code? Patrick to investigate this.

Patrick would like to have an agreement (between Fermi and DESY) on licensing. A proposal: SRM stays BSD with Fermi-lab as owner, for other code we come to an agreement on a common license.

Gene to poke around on this issue: two questions:

is the BSD policy mandatory? is it OK to come up with a common statement?

"Would you be allows to contribute to a file that is AGPL without insisting that it is BSD licensed?"

Outstanding RT Tickets

[This is an auto-generated item. Don't add items here directly]

Oleg visited us .. he asked about webdav without KPWD file. He wants to use vo-rolemap file, so

kpwd direct hooks for webadmin.

Access with a role.

Tanja to ask Oleg to open ticket for further information.

Review of RB requests

DTNM

Same time, next week.