You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Recovery tasks are created based on message counts, which is a static config value. If this value is not correctly optimized, it can cause the cache nodes to download too much data and crash with an out-of-disk.
Expected behavior
There are several ways we could resolve this - the cache nodes could dynamically calculate and advertise slots based on available disk space; the recovery process could use disk size instead of message counts (somewhat complex to create tasks based on an anticipated size).
Of these options the former may be the most attractive, as it also accounts for the inverse issue where lots of small recovery tasks take up slots but not much disk space.
The text was updated successfully, but these errors were encountered:
Describe the bug
Recovery tasks are created based on message counts, which is a static config value. If this value is not correctly optimized, it can cause the cache nodes to download too much data and crash with an out-of-disk.
Expected behavior
There are several ways we could resolve this - the cache nodes could dynamically calculate and advertise slots based on available disk space; the recovery process could use disk size instead of message counts (somewhat complex to create tasks based on an anticipated size).
Of these options the former may be the most attractive, as it also accounts for the inverse issue where lots of small recovery tasks take up slots but not much disk space.
The text was updated successfully, but these errors were encountered: