[dns-wg] Service outage at two K-root nodes

Dear colleagues,

We have been upgrading the operating system and name server software on
all K-root servers. The process has mostly gone quite well, but we have
had an unfortunate incident. Two of the newly upgraded nodes had been
accidentally announcing the K-root prefixes even though the name server
software on them was not running. This means that queries sent to these
two servers between 12 and 19 January were not answered.

The affected nodes are in Belgrade and Reykjavik. We estimate the
percentage of queries lost to be 0.3%.

Normally, failure of the name server would result in automatic
withdrawal of the prefix announcements. However, this failure was caused
by a flaw in the deployment process, which did not activate the
correct software components in the right sequence. We have identified
this issue and fixed it, so that this failure cannot happen with any
future installations or upgrades.

We apologise for any inconvenience caused. If you have any specific
questions please send email to <dns at ripe.net>.

Regards,
Anand Buddhdev
RIPE NCC

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 898 bytes
Desc: OpenPGP digital signature
URL: </ripe/mail/archives/dns-wg/attachments/20170120/802c7055/attachment.sig>