wiki:developers-meeting-20140806
Last modified 3 years ago Last modified on 08/06/14 15:59:39

[part of a series of meetings]

Participants

Karsten, Christian, Paul, Gerd, 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 :)

Karsten: Mostly working on LSDMA Article, some user support regarding gPlazma2,

Requested and got a renewal of our open-source license for IntelliJ Helped Summer Students

Christian: GridKA tutorial

Paul: Still working through email after vacation. Tried to get release out with Jenkins (s.b.)

Code reviews. LSDMA work coming, GridKA talk

Al: Wathcing public dCache, some issues with NFS, put some time into fixing and testing

ReplicaManager. Getting deeper into Wicket.

Gerd: Debugging issues and fixing things on "all" branches. Developing ideas how to handle lost

files on tape.

Plans for patch-releases

Should we make a new patch release?

Decided to work through pending tasks before releasing. Currently all is merged and the tests are running.

2.8-cloud is for cloud service at DESY.

2.2 is no longer supported, but the build mechanisms are still there for emerency cases

Trunk activity

Progress with new features...

Paul: Added autogeneration of SRM classes from WSDL as part of the build process.

There is still some fixing needed. Q1: Would it make sense to have to autogeneration code in SRM common? Or should

we split it with a little duplication to server and client.

Q2: Wrote a SRM cli client. This allows for a lot faster response times. This

does not work with Axis1. We could move to Axis2, but the problem is that Axis2 does not support rpc-encoded.

As a first step we could use Axis2 on the client side. Our server should understand a clients doc/literal. Haven't check gsoap, yet.

  • Gerd: Could we hack the axis1 to allow keeping open connections?
  • Paul: It could be that they close the http1 connection for a reason...

It would be good to know what the future of SRM actually is. Currently the time "after" SRM is officially said to be solved by the experients.

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

Gerd: A tape system was broken and stage requests piled up, but since the replies never arrived, the entries in the

map were never cleaned and may have caused the OOM.

  • Paul: We could kill the original message on retry

We already had something similar in other places.

Gerd already provided patches for this.

Paul: Current versions of dCache have a clumping-limit of 1. That means that only the first request to a file is successful

and subsequent requests are rejected. Gerd's patches fixes this by triggering a refresh from the door and by increasing the clumping limit to >1 (20) This patch has been backported to the stable branches

Issues from EMI

New or noteworthy

Outstanding RT Tickets

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

Review of RB requests

DTNM

Proposed: same time, next week.