Constant polling of metadata/content queue causes high CPU usage (indicator & syncdaemon) on large queues
Bug #692263 reported by
Roman Yepishev
This bug affects 1 person
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
Ubuntu One Indicator |
Fix Released
|
Undecided
|
Roman Yepishev |
Bug Description
In case there are 8000 items in metadata queue every MetaQueueChanged causes the indicator to request syncdaemon to provide the current metadata queue. This causes both performance degradation of syncdaemon and high CPU usage.
I belive metadata items should be polled based on the amount of the data in the queue:
for < 1000 - every time MetaQueueChanged is sent
for > 1000 - every 10th time MetaQueueChanged is sent
Related branches
Changed in one-indicator: | |
assignee: | nobody → Roman Yepishev (rye) |
Changed in one-indicator: | |
status: | New → Fix Committed |
summary: |
- Constant polling of metadata queue causes high CPU usage (indicator & - syncdaemon) on large queues + Constant polling of metadata/content queue causes high CPU usage + (indicator & syncdaemon) on large queues |
Changed in one-indicator: | |
status: | Fix Committed → Fix Released |
To post a comment you must log in.
Unfortunately, this bug is still there. When first syncing a large directory, the indicator lists meta > 5000, content > 20000, and the ubuntuone-indicator process is taking 50% cpu, with ubuntuone- syncdaemon taking the other 50%. If I killall ubuntuone- indicator, the syncdaemon process goes back down to a manageable level.