best effort snapshot pre-fetch when starting/restarting

[jonboulle]

No description provided.

[xiang90]

Change this to unplanned. We have not found a problem for large snapshot so far. We will keep tracking this issue.

We want to per-fetch snapshot since it 100% has no side effect to raft pipeline. And if a node cannot catch up for some reason, we can ignore it early.

/cc @heyitsanthony

[philips]

Xiang, can you add more description into this? I understand what we are
trying to say with the title but it is unlikely everyone else does.

On Tue, Dec 22, 2015 at 4:14 PM Xiang Li notifications@github.com wrote:

Change this to unplanned. We have not found a problem for large snapshot
so far. We will keep tracking this issue.

We want to per-fetch snapshot since it 100% has no side effect to raft
pipeline. And if a node cannot catch up for some reason, we can ignore it
early.

/cc @heyitsanthony https://github.com/heyitsanthony

—
Reply to this email directly or view it on GitHub
#3715 (comment).

[jpbetz]

Closing this one. Issues with snapshots is being discussed/address with #9161