You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@asterixdb.apache.org by Wail Alkowaileet <wa...@gmail.com> on 2018/12/09 05:21:05 UTC
LSM Filter thread safety
Dev,
Is the in-memory LSMCompnentFilter (min and max) tuples thread safe? I
could not notice any locking mechanism that guarantees they are thread safe?
--
*Regards,*
Wail Alkowaileet
Re: LSM Filter thread safety
Posted by Wail Alkowaileet <wa...@gmail.com>.
Actually, it was just by looking at the code. After Investigating, I can
confirm it's not thread safe. I filed an issue at ASTERIXDB-2493
<https://issues.apache.org/jira/browse/ASTERIXDB-2493>.
On Sun, Dec 9, 2018 at 10:31 AM Ian Maxon <im...@uci.edu> wrote:
> If my memory serves correctly the locking is done outside the filter
> modification, by virtue of the other locks around component modifications.
> So there shouldn’t be any within the filters themselves. Did you notice
> unusual behavior?
>
> > On Dec 8, 2018, at 21:21, Wail Alkowaileet <wa...@gmail.com> wrote:
> >
> > Dev,
> >
> > Is the in-memory LSMCompnentFilter (min and max) tuples thread safe? I
> > could not notice any locking mechanism that guarantees they are thread
> safe?
> >
> > --
> >
> > *Regards,*
> > Wail Alkowaileet
>
--
*Regards,*
Wail Alkowaileet
Re: LSM Filter thread safety
Posted by Ian Maxon <im...@uci.edu>.
If my memory serves correctly the locking is done outside the filter modification, by virtue of the other locks around component modifications. So there shouldn’t be any within the filters themselves. Did you notice unusual behavior?
> On Dec 8, 2018, at 21:21, Wail Alkowaileet <wa...@gmail.com> wrote:
>
> Dev,
>
> Is the in-memory LSMCompnentFilter (min and max) tuples thread safe? I
> could not notice any locking mechanism that guarantees they are thread safe?
>
> --
>
> *Regards,*
> Wail Alkowaileet