iamqq23ue commented on pull request #3382:
URL: https://github.com/apache/rocketmq/pull/3382#issuecomment-975005267
I think anyone who sets SYNC_FLUSH expects that SEND_OK will only be
returned when all data is written to the disk.
In financial scenarios, any data loss is unacceptable. If the user does not
need the absolute security of
the data, ASYNC_FLUSH will be set.
------------------ 原始邮件 ------------------
发件人:
"apache/rocketmq"
***@***.***>;
发送时间: 2021年11月22日(星期一) 上午9:49
***@***.***>;
***@***.******@***.***>;
主题: Re: [apache/rocketmq] FIX ISSUE#2706 the problem of returning
SEND_OK after flush failed,The original pull request was #2707. (#3382)
@Jason918 commented on this pull request.
In store/src/main/java/org/apache/rocketmq/store/MappedFile.java:
> @@ -288,6 +289,7 @@ public int flush(final int flushLeastPages) {
} } catch (Throwable e) {
log.error("Error occurred when force data to disk.", e); +
this.flushError = true;
It's ok for some users that data is written on the disk eventually. Changing
this default behavior have too much compact on the compatibility.
IMHO, maybe adding an optional config to enable this fast failure feature
on disk error?
—
You are receiving this because you authored the thread.
Reply to this email directly, view it on GitHub, or unsubscribe.
Triage notifications on the go with GitHub Mobile for iOS or Android.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
