Last modified 3 years ago Last modified on 04/02/19 15:10:47

dCache Tier I meeting APril 2, 2019

[part of a series of meetings]

Present, Paul), IN2P3(), Sara(), Triumf(), BNL(), NDGF(), PIC(), KIT(), Fermi(), CERN()


(see box on the other side)

Site reports


No major issues.

Accidentally killed a head node -- minor outage, but recovered.

No strange activity.


Everything is running fine.

httpd out-of-memory

Ticket 9537

Similar issue at Vancouver.

Is there a correlation between number of active transfers and the domain running out of memory.

Host memory or Java memory?

Host memory

Could you investigate whether you can monitor the Java memory?

RT ticket 9648

Two issues:

  1. org.dcache.uuid missing or reused.
  1. Slow GPFS causing xrootd client to time-out on kXR_read operations.

Active transfers

In the web admin.

Could this be updated to include the pool-to-pool in the "active transfer" JSON?

No, since we have thrown the web-admin interface away.

We need to check whether this is present in the new REST API.

Support tickets for discussion

[Items are added here automagically]


Proposed: same time, next week.