We will be deploying a series of upgrades to the infrastructure
underlying the datatracker on 4 June roughly between 1800 and 2200 UTC.
The first upgrade will be to move datatracker from memcached to Redis.
This upgrade will require a short (few minutes) of actual outage, and
will result in everyone being logged out. After this upgrade, persistent
login will be more durable. This upgrade will provide more opportunities
for future performance improvements and will provide a more stable
platform for future upgrades.
The remaining upgrades are to the underlying infrastructure. We will be
upgrading the Kubernetes cluster hosting datatracker to the latest
version. We believe we will be able to make these upgrades with zero
downtime.
Note however, that some operations involving asynchronous processing,
like I-D submission, may take longer to complete during these upgrades.
This upgrade technique brings us closer to zero-downtime deployments of
new versions of the datatracker itself, as it can be used for those
deployments that do not change the database schema. New versions in the
immediate future that dooperate on the schema will still require a small
number of minutes of downtime to deploy. We continue to work towards
being able to deploy such versions with zero downtime.
Robert Sparks
Senior Director of IT for the IETF LLC.
_______________________________________________
IETF-Announce mailing list -- [email protected]
To unsubscribe send an email to [email protected]
