Hi,
here are the minutes from todays infra call:
• Pootle migration status
• Migration takes much longer than expected
• For the future, evaluate migration paths that shorten downtime
• Export -> Modify SQL Dump, Migrate -> Import as example
• Jan: Postgres is not fast with Pootle
•
Hi,
the migration of dauntless took a bit longer than expected, because we ran into
the following problems:
- During reboot, the RAID was not assembled as fast as expected by systemd.
Fixed by adding rootdelay=10 to /etc/default/grub
- The network configuration with bridges was hard to adapt
The database migration is still in progress as of Wednesday, 20 April, 1230 UTC.
It is taking longer than expected by the pootle experts.
Details about the ongoing migration:
- The current database migration is the addition of a column to the largest
table of pootle
- The size of the tablespace
Hello,
Alexander Werner wrote on 2016-04-20 at 14:59:
The database migration is still in progress as of Wednesday, 20 April, 1230 UTC.
It is taking longer than expected by the pootle experts.
thanks so much to you and Cloph for all your hard work on this and the
long night shift - kudos!
Hi,
for the migration of our virtualization servers to our own rack and the upgrade
to 10G based Intranet, we plan the downtime of excelsior on April 26, 0730UTC,
and the downtime of falcon on April 27, 0730UTC. The downtime should last about
one hour.
The vms that will be affected are: