Thomas Steur

@DevDavido We're not planning to offer such a refactored plugin where we track the data ourselves currently. That's because for many/most users this would result in doubling the number of...

In case it helps, it seems to mostly happen with high traffic instances. It could be though that it's unrelated to high traffic and maybe they use certain features or...

@sgiehl I'm seeing in one of them around visitorType and there are some nested calls which I can see in the full backtrace. fyi we're seeing this error around 20+...

@sgiehl what would be the next steps to resolve this issue and make progress? If the distinct was the problem, I would assume then it be maybe more reproducible but...

I had a quick search for "when does mysql distinct not return distinct values" and "mysql distinct vs group by" as we can pretty much rule out the table exists...

Be good to test it if the log * makes the archiving fail if using cron archiving * if the error could show up in the UI when this happens...

@JasonMortonNZ @samjf could we log the query mentioned in https://github.com/matomo-org/matomo/compare/m18797?expand=1#diff-4140e491c64a9dcbf541f11e1db846adc3c27a4000810494ba8688e5feeee776R332 using our cloud file logger whenever the error occurs? Because the temp table insert eventually works at some point when...

We've been logging the queries that produced this error and then executed the `select` part. It's not clear why they would produce any duplicate visitIDs. Often they don't even have...

Verified that the table is empty when this error happens, then the `insert` is executed and the duplicate record error appears.

FYI I've also applied below patch but the error seems to still happen meaning it doesn't look like the read uncommitted is causing this. ``` diff --git a/core/DataAccess/LogAggregator.php b/core/DataAccess/LogAggregator.php index...