Containers should expose change-indicator
Bug #961547 reported by
justinsb
This bug affects 1 person
| Affects | Status | Importance | Assigned to | Milestone | |
|---|---|---|---|---|---|
| OpenStack Object Storage (swift) |
Won't Fix
|
Undecided
|
Unassigned | ||
Bug Description
Looking at the code, the containers keep track of information used for synchronization. It looks like the keep a hash, a timestamp, and also a sequence id.
It would be very nice if that was exposed, so that we could synchronize / cache nicely.
This can be "lossy" i.e. false positives are acceptable. But then I could cache the directory listing long-term e.g. in cloudfuse.
Revision history for this message
| Greg Lange (greglange) wrote | #1 |
| Changed in swift: | |
| status: | New → Won't Fix |
To post a comment you must log in.
So, I ran this idea past "top people" and this is the answer I got.
This isn't a bad idea. However, it's very dependent upon the way Swift works now.
If, for example, we implemented container splitting, returning such a value (and/or keeping it up to date), might prove too complicated.
Container splitting would allow the limits in number of objects per container to go away. It would also result in containers having more than a single database. Swift having to return this value after container splitting is done might create more complexity than we want.