shuaiqig commented on PR #6236: URL: https://github.com/apache/hadoop/pull/6236#issuecomment-1801256224
> > it uploads a fsimage, which may be tens of gigabytes, which will make the disk where the ANN metadata is stored very busy > > This makes sense to me. I don't know if the case @shuaiqig reported is same as mentioned here. If true, I don't think it could be prevented only avoid do checkpoint and roll editlog at same time from Standby, because sometimes Active could trigger roll editlog by itself. Another side, if not roll editlog in time, it will involve more load for Standby when replay it because there will overstock much more transaction to process. Right? IMO, we should improve the local storage performance or limit the throughput when upload fsimage such as decrease `dfs.image.transfer.chunksize`. FYI. @Hexiaoqiao Yes, you are right. As I mentioned in jira, this PR is not the best way to solve this problem, it is just the fastest and most convenient way that I can think of temporarily. And thank you for your review. -- 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] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
