gaogaotiantian commented on code in PR #57083:
URL: https://github.com/apache/spark/pull/57083#discussion_r3581569311


##########
.github/workflows/build_infra_images_cache.yml:
##########
@@ -20,6 +20,13 @@
 name: Build / Cache base image
 
 on:
+  # Rebuild cache on a schedule to keep it warm despite base image drift.
+  # The `ubuntu:noble` tag is re-pointed to a new image roughly monthly,
+  # which invalidates all cached layers built on top of it. Without a
+  # scheduled rebuild, the cache goes stale between Dockerfile changes
+  # and every commit on every PR spends ~50 minutes building base images!
+  schedule:
+    - cron: '0 4 * * *'

Review Comment:
   Pinning versions has its issues too. We need to monitor and manually update 
it. We could miss security updates. 24 hours is much better than a week. If 
this is cheap enough we can even make it more frequent - 2 or 4 times a day. I 
think we just need a tradeoff here. 24 hour per month is not too bad - and it's 
not always 24, on average it should be 12. Considering that we could update 
docker image for other reasons during that time, the gap is even smaller. We 
can check the cost vs waste and find a good balance.



-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: [email protected]

For queries about this service, please contact Infrastructure at:
[email protected]


---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to