2.6 KiB
:orphan:
Running expensive migrations early
Zulip 1.7 and 1.9 each contain some significant database migrations that can take several minutes to run.
The upgrade process automatically minimizes disruption by running these first, before beginning the user-facing downtime. However, if you'd like to watch the downtime phase of the upgrade closely, you can run them manually before starting the upgrade:
-
Log into your Zulip server as the
zulip
user (or asroot
and then runsu zulip
to drop privileges), andcd /home/zulip/deployments/current
-
Run
./manage.py dbshell
. This will open a shell connected to the PostgreSQL database. -
In the PostgreSQL shell, run the following commands:
CREATE INDEX CONCURRENTLY zerver_usermessage_is_private_message_id ON zerver_usermessage (user_profile_id, message_id) WHERE (flags & 2048) != 0; CREATE INDEX CONCURRENTLY zerver_usermessage_active_mobile_push_notification_id ON zerver_usermessage (user_profile_id, message_id) WHERE (flags & 4096) != 0;
(These first migrations are the only new ones in Zulip 1.9).
CREATE INDEX CONCURRENTLY
zerver_usermessage_mentioned_message_id
ON zerver_usermessage (user_profile_id, message_id)
WHERE (flags & 8) != 0;
CREATE INDEX CONCURRENTLY
zerver_usermessage_starred_message_id
ON zerver_usermessage (user_profile_id, message_id)
WHERE (flags & 2) != 0;
CREATE INDEX CONCURRENTLY
zerver_usermessage_has_alert_word_message_id
ON zerver_usermessage (user_profile_id, message_id)
WHERE (flags & 512) != 0;
CREATE INDEX CONCURRENTLY
zerver_usermessage_wildcard_mentioned_message_id
ON zerver_usermessage (user_profile_id, message_id)
WHERE (flags & 8) != 0 OR (flags & 16) != 0;
CREATE INDEX CONCURRENTLY
zerver_usermessage_unread_message_id
ON zerver_usermessage (user_profile_id, message_id)
WHERE (flags & 1) = 0;
These will take some time to run, during which the server will continue to serve user traffic as usual with no disruption. Once they finish, you can proceed with installing Zulip 1.7.
To help you estimate how long these will take on your server: count
the number of UserMessage rows, with select COUNT(*) from zerver_usermessage;
at the ./manage.py dbshell
prompt. At the time these migrations
were run on chat.zulip.org, it had 75M UserMessage rows; the first 5
indexes took about 1 minute each to create, and the final,
"unread_message" index took more like 10 minutes.