kvserver: increase snapshot timeout and use envvar #76589
Conversation
|
This change is |
There was a problem hiding this comment.
Reviewable status:
complete! 0 of 0 LGTMs obtained (waiting on @erikgrinaker and @tbg)
-- commits, line 8 at r1:
Drive-by comment: since the 20min timeout was never released, should this release note just be talking about the fact that now this new env var can be used to set a general timeout?
There was a problem hiding this comment.
Reviewable status:
Previously, aayushshah15 (Aayush Shah) wrote…
Drive-by comment: since the 20min timeout was never released, should this release note just be talking about the fact that now this new env var can be used to set a general timeout?
Yeah. I figured I'd have to coordinate with docs anyway to undo the 20 minute release note (since the timeout is no longer 20 minutes), but I can word this independently and figure out how to nullify the old one.
This patch increases the default snapshot transfer timeout from 20 minutes to 1 hour, to be more lenient, and adds an environment variable `COCKROACH_RAFT_SEND_SNAPSHOT_TIMEOUT` as an escape hatch. Release note (bug fix): There is now a 1 hour timeout when sending Raft snapshots, to avoid stalled snapshot transfers preventing Raft log truncation and growing the Raft log very large. This is configurable via the `COCKROACH_RAFT_SEND_SNAPSHOT_TIMEOUT` environment variable.
erikgrinaker
force-pushed
the
send-snapshot-timeout-envvar
branch
from
9e4c3aa to
23c673e
Compare
|
bors r=tbg |
|
Build succeeded: |
This patch increases the default snapshot transfer timeout from 20
minutes to 1 hour, to be more lenient, and adds an environment variable
COCKROACH_RAFT_SEND_SNAPSHOT_TIMEOUTas an escape hatch.Touches #76295.
Release note (bug fix): There is now a 1 hour timeout when sending Raft
snapshots, to avoid stalled snapshot transfers preventing Raft log
truncation and growing the Raft log very large. This is configurable via
the
COCKROACH_RAFT_SEND_SNAPSHOT_TIMEOUTenvironment variable.