Last modified 4 weeks ago Last modified on 09/18/18 17:15:32

[part of a series of meetings]



About one minute (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 :)

Marina: Several bug fixes, releases, investigating Kafka appender. Investigating why Kafka killed all doors. Last week had the XDC All-Hand Meeting. Femi: XDC AllHands? Meeting; bug fixes going through RB. Dmitry: Preparing 4.2 dCache upgrade. Identified problems: 3rd-party HTTP failures, delegation, ... Tweak for gplazma fermi plugin. Remote transfer retrying due to staging. On Thursday upgrade to 2.16; October 4.2. DUNE are taking data, don't want downtime -- xrootd padding problem. Tigran: reading through emails, coming back. DESY cloud to 4.2 -- one problem: bad feeling that HA is the cause. Seem to have tunnels that are not working correctly. pool gets disabled when file is deleted. Update from Adrian/IN2P3 before fix, they get pool crashing couple of times per day, now a couple of times per week. Happens when removing a file in the flush queue -- all sites are LHCb sites. Vincent: XDC on panoptes; NDGF did the upgrade 4.2.6 -- found some problems -- hit by xrootd issue. Tape plugin to flush files on tape -- manage to reproduce it on pre-production. But no error on compilation, and no error at run time. Enabled macaroon and enable frontend. Oslo would like to enable CEPH using dCache CEPH support. Currently 1 PiB, should grow up to 3 PiB. Max file descriptor is not honoured by systemd -- patch coming soon. Paul:

Special topics

Please list special topics you would like to bring up here *before* the meeting, and, if posssible, add a time estimate. We should try to limit this segment to 45 minutes, and that only in exceptional cases.

1) xrootd4j "pad block corrupted" exception

Xavier reported this as being a problem with 1.9.x

US-CMS doesn't show any failures when tested against GSI. This is running Java v1.8.0_111b15. On the test instance, this is Java v1.8.0_141b16.

Demonstrate the problem with SLAC xrootd.

Should be an if-statement that triggers the exception. Can we figure out what triggers this, and update the code so we can understand that is causing this.

March 6th BouncyCastle? had a patch that fixes some incompatibility with OpenSSL and padding.

2) Inability to query RESTFul API w/o assuming admin role

Paul to write a document on how to implement

3) 3rd party HTTP(s) transfers.

1) davix-cp seem to be mixing interface version number

and delegation version number

2) There is problem w/ how delegation is handled after stored

Release news

Keeping with the weekly release cycle.

Paul mentioned the possibility of making a dCache release to satisfy XDC requirements. An alternative would be to create a XDC-specific branch, with XDC-specific releases, but we may need to support this.

Trunk activity

Progress with new features...

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

Nothing major -- Femi is aware of the Swestore issues (X.509 certificate authn)

Outstanding RT Tickets

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

Review of RB requests

New, noteworthy and other business


Same time next week, again on