Some complex SQL statements in our current system take a long time to execute. Will the official consider making a parallel query function? Of course, we have also made a parallel query function for percona-server-8.0.36 based on some open source patches from Huawei, but we are not professional innodb developers, so we cannot develop some new functions for parallel query. We will fork and create a pq branch to submit the code later. It would be great if the official can provide help for such functions.
1 Like
We consider any patches that the community provides. If you’d like to open a https://jira.percona.com/ ticket with the relevant code patches from Huawei, we will certainly consider it.
Hello, matthewb, I submitted the parallel query feature to a new branch. The address is GitHub - k1n9met/percona-server at 8.0.36-28-pq. Just compile it normally. The usage method is the same. The introduction has been explained when submitting, but this patch has modified a lot of the original files. I guess it is unlikely to be merged into the official branch, but it can be used as a small function (I personally guess that the plug-in does not count, after all, compared with the official version There are too many files to change, and we don’t have the ability to transform it into a plug-in). If possible, the official can develop some new features for this patch or interested developers can also fork and develop it.
Hello @Crimson, that’s amazing work! I have absolutely shared this repo directly in our mysql engineering channel. Still no guarantees of if/when, but this certainly helps! I’ve created a feature request for tracking purposes: [PS-9249] - Percona JIRA
Hey @Crimson, is there any testing for this patch? One of the first comments I saw in the thread was “it misses the test suite”. Do you have examples for where it passes and where it fails, or are there any from Huaui?
Hey, @matthewb , I have submitted the mtr test, but I am not sure whether all the tests can pass, but I think more than 99% of the tests can pass. In addition, in the details of the submission, I will introduce the parameters and its value range and The variables are also described in detail, and pictures of the implementation principles and key process functions are also put into the project. If there is anything else missing about parallel queries, please let me know.
I am glad that you have created a feature request. At the same time, Thank you also for sharing in the mysql engineer channel. I think the mysql official may not adopt this request (after all, they are still business-oriented and may conflict with their commercial products), but I think the percona official may adopt it. Because percona server provides many enterprise version features that native mysql does not have, this is why we modify it based on the percona server branch. In any case, we welcome all developers to develop new features or use them at any time.