Recovery merge #679
Recovery merge #679
Conversation
abudnik
force-pushed
the
recovery_merge
branch
from
911b139
to
f251b74
Compare
| status = result.response.status | ||
| self._update_stats(start_time, index + 1, recovers_in_progress, status) |
There was a problem hiding this comment.
enumerate accepts optional parametr, which sets initial value of index. It's 0 by default.
enumerate(iterator, 1)to remove index + 1
|
Fixed: 20e5376 |
abudnik
force-pushed
the
recovery_merge
branch
from
20e5376
to
8333f37
Compare
|
Hi all, do you have plans on merging this PR into master? |
|
@abudnik Is it save for data? |
|
@agend It successfully passes our recovery tests but I wouldn't recommend you to test it on your production cluster until it is not tested on variety of test dataset or at least you have tested it on your test dataset that are close to your production. PRs #679 and #681 are completed but before merging them we should fully test it to be sure that it is safe. Also we have one open question under discussion which doesn't affect to recovery correctness but defines recovery behaviour in case of error: "On which errors should copy-iterator stop?". |
| self.stats.set_counter('recovers_in_progress', recovers_in_progress) | ||
| if status != -errno.ETIMEDOUT: | ||
| self.stats.counter('recovered_keys', 1 if status == 0 else -1) | ||
| self.ctx.stats.counter('recovered_keys', 1 if status == 0 else -1) |
There was a problem hiding this comment.
I've fixed distinction between _update_stats() code in both PRs (509bd3b). They are the same now.
|
Should I merge this pull request? |
|
I will overlook and actualize it later today. |
|
It depends on #681 which has conflicts with current master. |
abudnik
force-pushed
the
recovery_merge
branch
from
688a8f1
to
80d3482
Compare
…py-iter everywhere
abudnik
force-pushed
the
recovery_merge
branch
from
80d3482
to
d3382e6
Compare
recovery: merge: removed obsolete Recovery code; used server-send, copy-iter everywhere
This PR sends greetings to PR #678