tag:blogger.com,1999:blog-9149523927864751087.post5267522888031381674..comments2024-03-26T09:43:01.052-07:00Comments on Small Datum: Compaction stalls: something to make better in RocksDBMark Callaghanhttp://www.blogger.com/profile/09590445221922043181noreply@blogger.comBlogger2125tag:blogger.com,1999:blog-9149523927864751087.post-21369613319289058122017-02-03T01:28:08.158-08:002017-02-03T01:28:08.158-08:00I have results for reads concurrent with insert be...I have results for reads concurrent with insert benchmark writes. I have yet to publish it. The summary is that MyRocks does much better than InnoDB. The problem for InnoDB is that dirty page write-back falls behind so a query might need to read a page and to get a free-page from the buffer pool a dirty page must be written back. So the read is sometimes stalled by the writes.Mark Callaghanhttps://www.blogger.com/profile/09590445221922043181noreply@blogger.comtag:blogger.com,1999:blog-9149523927864751087.post-10995847780451608372017-02-02T02:25:25.064-08:002017-02-02T02:25:25.064-08:00Great post, I love to see charts presented in such...Great post, I love to see charts presented in such a way as seeing the bands of expected iops is useful. From the look of things for my rocks, there are two distinct areas where consistency suffers. I agree that the behavior looks like stalls from compaction. I imagine that 99th (and other) percentile latency spikes around the same time.<br /><br />As a db for things that do a lot of analytics or churn through larger workloads that are write heavy via a lot of workers, this is ok. I am not sure if the read latency and QPS is affected while a compaction is happening. That would be an interesting question, as it is more representative of the workload of a large job kicking off while production user facing reads are happening.<br /><br />I am somewhat impressed with the burst performance for 5.7 : with memory available 3m inserts can happen in 100s. As per your previous post, of course, in a longer running job my rocks does things very consistently. Alex Malafeevnoreply@blogger.com