backup: fix retry of fine-grained backup #43252
Conversation
Signed-off-by: hillium <[email protected]>
Signed-off-by: hillium <[email protected]>
Signed-off-by: hillium <[email protected]>
Signed-off-by: hillium <[email protected]>
|
[REVIEW NOTIFICATION] This pull request has been approved by:
To complete the pull request process, please ask the reviewers in the list to review by filling The full list of commands accepted by this bot can be found here. Reviewer can indicate their review by submitting an approval review. |
ti-chi-bot
added
release-note
YuJuncen
changed the title
Signed-off-by: hillium <[email protected]>
|
|
||
| // TotalSleepInMS returns the total sleeped time in ms. | ||
| func (r *RetryWithBackoffer) TotalSleepInMS() int { | ||
| return r.totalBackoff + r.bo.GetTotalSleep() |
There was a problem hiding this comment.
It seems that the r.bo is always forked, and only the forked bos are updated something. So is r.bo.GetTotalSleep always 0?
There was a problem hiding this comment.
Yeah indeed, but the origin logic is this. I'm not sure why it was implemented as this(and have no idea about how to improve that), so I just inherited that.
There was a problem hiding this comment.
Anyway unlocking locks has a bounded backoff time, which is the TTL of the lock. I think omit the retry time it cost won't be really harmful, hopefully...
There was a problem hiding this comment.
The sole difference between the original logic is that resolve lock may fail due to the accumulated sleep time reaches the sleep time limit. I think that would be fine to eliminate that fail case.
ti-chi-bot
added
status/LGT2
|
/merge |
|
This pull request has been accepted and is ready to merge. Commit hash: 886bc2b
|
ti-chi-bot
added
the
status/can-merge
YuJuncen
added
needs-cherry-pick-release-5.3
|
/run-cherry-picker |
|
In response to a cherrypick label: new pull request created to branch |
|
In response to a cherrypick label: new pull request created to branch |
Signed-off-by: ti-chi-bot <[email protected]>
|
In response to a cherrypick label: new pull request created to branch |
Signed-off-by: ti-chi-bot <[email protected]>
|
In response to a cherrypick label: new pull request created to branch |
* cherry-pick 1a94e5d Signed-off-by: hillium <[email protected]> * run gofmt Signed-off-by: hillium <[email protected]> * fix a typo Signed-off-by: hillium <[email protected]> --------- Signed-off-by: hillium <[email protected]>
YuJuncen
added
the
needs-cherry-pick-release-7.1
|
/run-cherry-picker |
|
In response to a cherrypick label: new pull request created to branch |
What problem does this PR solve?
Issue Number: close #43236
Problem Summary:
We are using the
tikv.Backofferto implement retry. Which isn't so customizable hence we cannot fully control the printed error messages and back off duration.What is changed and how it works?
This PR added a new structure
RetryWithBackofferto adapt the retry logic with the TiDB backoffer logic, at the same time it has enhanced the error message.This PR will also increase the max retry time of fine-grained backup. The reason is simple: given it begins the fine-grained backup, there must be some problems in the cluster. We need to be more patient.
Check List
Tests
Release note
Please refer to Release Notes Language Style Guide to write a quality release note.