wiki:developers-meeting-20091209
Last modified 11 years ago Last modified on 12/16/09 16:01:24

[part of a series of meetings]

Participants

Agenda

[see box to the 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 :)

Owen

  • Been on Holiday, cleaned up patches for resubmission, and provided patch for autotools dcap.
  • Clean up more patches and get certification ready to start for new dCache Server.

Paul

  • A few days off, fixing pnfsDump, working on advisory for pnfs-Chimera migration nlink issue.
  • Final handling of advisory, finalise pnfsDump multithreaded support.

Timur

  • Fix to PinManager? to unpin files were the sticky bit is confused, testing Terracotta with S2, thinking of upgrading to 1.9.5-10.
  • next week working with gPlazma working with Kerberos.

Dmitry

  • Regexp work in auth file, SRM testing.

Vejay

  • Worked on automatically detect underlying namespace impl. for enstore. Review some patches. Test whether PNFS is running before
  • Work on the making the Fermi-dCache package work with Chimera.

Gerd

  • Worked on read-permission checks, two days doing performance testing on the patches, debugging jglobus problem with CRL race-conditions.
  • Work on work-around in BC for jglobus , ...

Irina

  • Improvements to HSM cleaner.
  • Trying to find out a possibility to set ACLs automaticaly on already existing files/directories. (As requested by Yvan Calas.) ACLs can only be inherited by newly created objects.

Jan

  • Got failing unit tests down to two. Focused on web-admin: looking at wikit
  • Continuing wikit work.

Tigran

  • Trying to find out why 1.9.6 and 1.9.5 doesn't work on test instance. 1.9.5 declared "pilot error", 1.9.6 still under investigation. Partially fixed tests and found an minor issue with Chimera and HSQL DB.
  • GSI dcap with multiple roles then new mover.

Patrick

  • (Enjoying Cuba) finalising European Middleware Initiative, 27M euro coming from the project. Patrick is part of the data-management team. We wait for the result. If all goes well, dCache will get 1 (+1) people for dCache. EMI is basically the follow-up from the EGEE projects. This will establish a European centre in Amsterdamn. NGIs are non-competative, EMI is competetive, though.
  • CERN is starting a new approach to solve the storage problems. They're trying to solve it without Castor, dCache, etc... a completely new approach.

Tanja

  • Off sick for most of last week, but when back looked at the CSS for the web-dav.
  • Continue working on pool manager.

Status of work for 1.9.5 (5 Min)

A (quick?) review of activity needed for the next 1.9.5 release

Outstanding issue is the GSI /JGlobus / BC issue.

Should we move back to the latest JGlobus version? Yes.

Back-porting fixes to 1.9.4

A couple of things fixed in 1.9.5, and BNL are not planning to move to 1.9.5 yet. Prestage on directories, PoolManager stuff (magic)-but this was also committed to 1.9.4.

When do we plan to end support for 1.9.4?

With the release of 1.9.7, we plan to drop support for 1.9.4.

Status of CERN release?

Small bug (in Hudsun)

CERN are closing their queue for their certification. We may have missed the opportunity to get the upgrade in before the new procedure is underway.

OSG release and 1.9.2?

Timur: have notified Tanya, but

Status of work for 1.9.6 (5 Min)

A (quick?) review of activity needed for the 1.9.6-1 release

Things that we would like to go in:

  • WebDAV CSS patch from Tanya,
  • SRM-ls from Dmitry.

Issues from yesterday's Tier-1 meeting (5 Min)

Nothing needed discussing

SRM issue with logging

It seems that some logging targets are not routed to our log4j configuration file.

Tomcat server runs with different class loader, so log4j configuration is not visible. Need to move this log4j lib to the common directory.

Should somehow try to fix this: merging tomcat and dCache logging, if possible, otherwise log events to another file.

Another issue is with log file rotation.

This is possible: two issues:

  • not always logging through log4j, there are some outstanding println()s in the code.

Potentially an issue if sites do their own rotation using external tools. log4j configuration file

Outstanding RT Tickets (5 Min)

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

RT 5297: 1.9.5-8 -> 1.9.5-9 "readline error"

RT 5298: Error message in pnfsDomain.log after upgrade

RT 5313: gsidcap control line errors

RT 5314: Few percentage dccp got No call back connection from dcap door at 1.9.3-3

Review of RB requests (10 Min)

DTNM

Proposed: same time, next week.