Last modified 4 years ago Last modified on 09/12/17 14:53:00

dCache Tier I meeting MONTH DATE, 2013

[part of a series of meetings]

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


(see box on the other side)

Site reports


OK at PIC, no issues during the last week.

Opened a ticket regarding the missing information in the pcells

Problem with IPv6 and IPv4

Upgrading development instance to 3.1

For some reason, pool was not able to connect to the core domains.

This pool has no IPv6 address; IPv4 only. Core domains are running dual-stack: IPv4 & IPv6. There is the "prefer IPv6" option enabled through the JVM command-line.

The pool tries to connect to the core domains, but fails with the following error message

12 Sep 2017 12:41:12 (c-dcache-dccore01-devDomain-AAVY-47-N_A) [] Failed to connect to dcache-dccore01-devDomain: Failed to connect to dccore01-dev. Network is unreachable

Further information:

[root@dc002 dcache]# ifconfig bond0
inet netmask broadcast
inet6 fe80::21b:21ff:fe91:aa17 prefixlen 64 scopeid 0x20<link>
ether 00:1b:21:91:aa:17 txqueuelen 50000 (Ethernet)
RX packets 78498 bytes 9111764 (8.6 MiB)
RX errors 0 dropped 0 overruns 0 frame 0
TX packets 35074 bytes 15377487 (14.6 MiB)
TX errors 0 dropped 0 overruns 0 carrier 0 collisions 0

[root@dc002 dcache]# host dc002 has address

The CMS preproduction testbed is IPv6 only, and is working fine.

The production instance are either dual-stack (if client-facing) or IPv4 only. All core domains are IPv4, and it is working fine.

After removing the "prefer IPv6" option java_options.extra, the problem went away.

Marc will open a ticket.

Also try removing the IPv6 link-local address to see if that fixes the problem.

Support tickets for discussion

[Items are added here automagically]


Same time, next week.