Comment 12 for bug 1797816

Ok, if the txn was broken in 2.2.6 that would explain why we're seeing the
large document in 2.3, and the fixes that we have around not letting the
queue grow too large wasn't in 2.2.

On Tue, Oct 16, 2018 at 2:40 PM John A Meinel <email address hidden>
wrote:

> I should make a few other comments
>
> 1) New versions of juju have a patch that should prevent ever getting a
> 200k transaction queue. (if when adding a transaction the queue would be
> >1000, we reject the txn as invalid.) That should prevent the txn queue
> from growing large, which was what blocked us from automatically pruning
> the table, and keeping the size under control.
>
> 2) We are currently looking at plans for how we can change the apply
> step so that it can note invalid transactions and abort them. That also
> helps get into this situation in the first place.
>
> 3) I don't think that we will add something to mgopurge explicitly for
> this use case, as we should mostly be able to avoid it in the future.
>
> --
> You received this bug notification because you are a bug assignee.
> Matching subscriptions: juju bugs
> https://bugs.launchpad.net/bugs/1797816
>
> Title:
> [2.3.8] jujud exhausts resources
>
> To manage notifications about this bug go to:
> https://bugs.launchpad.net/juju/+bug/1797816/+subscriptions
>