Using the internals in ThreadAwareSecurityManager has caused countless
problems, and needs to be fixed once and for all -

There are 2 different patches up for review in
https://issues.apache.org/jira/browse/CASSANDRA-13396 - would be nice if
one could be selected, and hopefully whichever is chosen can be a final
workaround for upgrading safely as well.




On Tue, Feb 13, 2018 at 9:41 AM, Jacques-Henri Berthemet <
jacques-henri.berthe...@genesys.com> wrote:

> Hi,
>
> I suppose upgrading Logback breaks Cassandra because some classes are used
> directly like in StorageService, ThreadAwareSecurityManager and
> StorageServiceMBean.
> This was a problem in my case as we're embedding Cassandra for our
> functional tests, I had to stub it as it was conflicting with log4j2
> configuration.
>
> Ideally Cassandra should only use pure SLF4J so that logging can be easily
> upgraded or changed.
> --
> Jacques-Henri Berthemet
>
> -----Original Message-----
> From: Ariel Weisberg [mailto:ar...@weisberg.ws]
> Sent: Tuesday, February 13, 2018 6:28 PM
> To: dev@cassandra.apache.org
> Subject: Re: CASSANDRA-14183 review request -> logback upgrade to fix CVE
>
> Hi,
>
> So our options are:
>
> 1. Ignore it.
> Most people aren't using this functionality.
> Most people aren't and shouldn't be exposing the logging port to untrusted
> networks But everyone loses at defense in depth (or is it breadth) if they
> use this functionality and someone might expose the port
>
> 2. Remove the offending classes from the 1.1.10 jar My crazy idea, break
> it, but only for the people using the vulnerable functionality. Possibly no
> one, but probably someone. Maybe they can upgrade it manually for their
> usage?
> This also has an issue when working with maven.
>
> 3. Upgrade it
> Definitely going to break some apps according to Michael Shuler. Happened
> when he tried it.
>
> Certainly we can upgrade in trunk? While we are at it come up to the
> latest version.
>
> Ariel
>
> On Tue, Feb 13, 2018, at 12:03 PM, Ariel Weisberg wrote:
> > Hi,
> >
> > I don't think the fix is in 1.1.11 looking at the diff between 1.1.11
> > and 1.2.0
> > https://github.com/qos-ch/logback/compare/v_1.1.11...v_1.2.0
> > .com
> >
> > I looked at 1.1.11 and 1.1.10 and didn't see it there either.
> >
> > When you say stuff broke do you mean stuff not in the dtests or utests?
> >
> > Ariel
> >
> > On Tue, Feb 13, 2018, at 11:57 AM, Michael Shuler wrote:
> > > I tried a logback 1.2.x jar update a number of months ago to fix the
> > > broken log rotation (try setting rotation to a large number - you'll
> > > find you only get I think it was 10 files, regardless of setting).
> > >
> > > Like we've found updating other jars in the past, this seemingly
> > > "simple" update broke a number of application components, so we
> > > rolled it back and worked out another log rotation method.
> > >
> > > Looking at the logback changelog, I cannot tell if version 1.1.11 is
> > > fixed for this, or if that might be less breakage? There are a
> > > pretty significant number of API-looking changes from 1.1.3 to
> > > 1.2.3, so I do not wish to break other user's applications, as I have
> experienced.
> > >
> > > I do not think this should block the current releases, unless
> > > someone wants to do some significant testing and user outreach for
> > > tentatively breaking their applications.
> > >
> > > --
> > > Michael
> > >
> > > On 02/13/2018 10:48 AM, Jason Brown wrote:
> > > > Ariel,
> > > >
> > > > If this is a legit CVE, then we would want to patch all the
> > > > current versions we support - which is 2.1 and higher.
> > > >
> > > > Also, is this worth stopping the current open vote for this patch?
> > > > (Not in a place to look at the patch and affects to impacted
> branches right now).
> > > >
> > > > Jason
> > > >
> > > > On Tue, Feb 13, 2018 at 08:43 Ariel Weisberg <ar...@weisberg.ws>
> wrote:
> > > >
> > > >> Hi,
> > > >>
> > > >> Seems like users could conceivably be using the vulnerable
> > > >> component. Also seems like like we need potentially need to do this
> as far back as 2.1?
> > > >>
> > > >> Anyone else have an opinion before I commit this? What version to
> > > >> start from?
> > > >>
> > > >> Ariel
> > > >>
> > > >> On Tue, Feb 13, 2018, at 5:59 AM, Thiago Veronezi wrote:
> > > >>> Hi dev team,
> > > >>>
> > > >>> Sorry to keep bothering you.
> > > >>>
> > > >>> This is just a friendly reminder that I would like to contribute
> > > >>> to this project starting with a fix for CASSANDRA-14183
> > > >>> <https://issues.apache.org/jira/browse/CASSANDRA-14183>.
> > > >>>
> > > >>> []s,
> > > >>> Thiago.
> > > >>>
> > > >>>
> > > >>>
> > > >>> On Tue, Jan 30, 2018 at 8:05 AM, Thiago Veronezi
> > > >>> <thi...@veronezi.org>
> > > >>> wrote:
> > > >>>
> > > >>>> Hi dev team,
> > > >>>>
> > > >>>> Can one of you guys take a look on this jira ticket?
> > > >>>> https://issues.apache.org/jira/browse/CASSANDRA-14183
> > > >>>> =issues.apache.org
> > > >>>>
> > > >>>> It has an a patch available for a known security issue with one
> > > >>>> of the dependencies. It has only with trivial code changes. It
> > > >>>> should be straightforward to review it. Any feedback is very
> welcome.
> > > >>>>
> > > >>>> Thanks,
> > > >>>> Thiago
> > > >>>>
> > > >>
> > > >> -----------------------------------------------------------------
> > > >> ---- To unsubscribe, e-mail: dev-unsubscr...@cassandra.apache.org
> > > >> For additional commands, e-mail: dev-h...@cassandra.apache.org
> > > >>
> > > >>
> > > >
> > >
> > >
> > > --------------------------------------------------------------------
> > > - To unsubscribe, e-mail: dev-unsubscr...@cassandra.apache.org
> > > For additional commands, e-mail: dev-h...@cassandra.apache.org
> > >
> >
> > ---------------------------------------------------------------------
> > To unsubscribe, e-mail: dev-unsubscr...@cassandra.apache.org
> > For additional commands, e-mail: dev-h...@cassandra.apache.org
> >
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: dev-unsubscr...@cassandra.apache.org
> For additional commands, e-mail: dev-h...@cassandra.apache.org
>
>

Reply via email to