Mailing List Archive

The MDC Patterns used by solr are for the collection, shard, replica, core
and node names, and a potential trace id. All of those are restricted to
alphanumeric, no special characters like $ or { needed for the injection.
And trying to access a collection that didn’t exist Returns 404 without
logging.

Upgrading is always going to be more complete, but I think we’re still ok
for now, at least until the next iteration of this attack surfaces.



On Tue, Dec 14, 2021 at 3:37 PM solr <fredrik@rodland.no> wrote:

> Only setting -Dlog4j2.formatMsgNoLookups=true might not be enough to
> mitigate the log4j vulnerability.
>
> See https://github.com/kmindi/log4shell-vulnerable-app
> “So even with LOG4J_FORMAT_MSG_NO_LOOKUPS true version 2.14.1 of log4j is
> vulnerable when using ThreadContextMap in PatternLayout.”
>
> ThreadContext.put(key, value) is used under the hood by MDC. I’m not sure
> wether any user-input is actually stored in MDC in SOLR.
>
>
> Probably this should be updated:
> https://solr.apache.org/security.html#apache-solr-affected-by-apache-log4j-cve-2021-44228
>
> And maybe consider releasing patch releases for other versions than 8.11
> as well which includes log4j 2.16.0?
>
>
>
> Regards,
>
>
> Fredrik
>
>
> --
> Fredrik Rødland Cell: +47 99 21 98 17
> Maisen Pedersens vei 1 Twitter: @fredrikr
> NO-1363 Høvik, NORWAY flickr: http://www.flickr.com/fmmr/
> http://rodland.no about.me http://about.me/fmr
>
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: dev-unsubscribe@lucene.apache.org
> For additional commands, e-mail: dev-help@lucene.apache.org
>
>

Ok.

But FTR - apache/log4j has discredited just setting the system property as a mitigation measure, so I still think the SOLR security-page should be changed to not list this as a valid mitigation:

https://logging.apache.org/log4j/2.x/security.html
"Older (discredited) mitigation measures

This page previously mentioned other mitigation measures, but we discovered that these measures only limit exposure while leaving some attack vectors open.

Other insufficient mitigation measures are: setting system property log4j2.formatMsgNoLookups or environment variable LOG4J_FORMAT_MSG_NO_LOOKUPS to true for releases >= 2.10, or modifying the logging configuration to disable message lookups with %m{nolookups}, %msg{nolookups} or %message{nolookups} for releases >= 2.7 and <= 2.14.1.
“

Regards,


Fredrik


--
Fredrik Rødland Cell: +47 99 21 98 17
Maisen Pedersens vei 1 Twitter: @fredrikr
NO-1363 Høvik, NORWAY flickr: http://www.flickr.com/fmmr/
http://rodland.no about.me http://about.me/fmr



> On 14 Dec 2021, at 23:44, Mike Drob <mdrob@mdrob.com> wrote:
>
> The MDC Patterns used by solr are for the collection, shard, replica, core and node names, and a potential trace id. All of those are restricted to alphanumeric, no special characters like $ or { needed for the injection. And trying to access a collection that didn’t exist Returns 404 without logging.
>
> Upgrading is always going to be more complete, but I think we’re still ok for now, at least until the next iteration of this attack surfaces.
>
>
>
> On Tue, Dec 14, 2021 at 3:37 PM solr <fredrik@rodland.no> wrote:
> Only setting -Dlog4j2.formatMsgNoLookups=true might not be enough to mitigate the log4j vulnerability.
>
> See https://github.com/kmindi/log4shell-vulnerable-app
> “So even with LOG4J_FORMAT_MSG_NO_LOOKUPS true version 2.14.1 of log4j is vulnerable when using ThreadContextMap in PatternLayout.”
>
> ThreadContext.put(key, value) is used under the hood by MDC. I’m not sure wether any user-input is actually stored in MDC in SOLR.
>
>
> Probably this should be updated: https://solr.apache.org/security.html#apache-solr-affected-by-apache-log4j-cve-2021-44228
>
> And maybe consider releasing patch releases for other versions than 8.11 as well which includes log4j 2.16.0?
>
>
>
> Regards,
>
>
> Fredrik
>
>
> --
> Fredrik Rødland Cell: +47 99 21 98 17
> Maisen Pedersens vei 1 Twitter: @fredrikr
> NO-1363 Høvik, NORWAY flickr: http://www.flickr.com/fmmr/
> http://rodland.no about.me http://about.me/fmr
>
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: dev-unsubscribe@lucene.apache.org
> For additional commands, e-mail: dev-help@lucene.apache.org
>


---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@lucene.apache.org
For additional commands, e-mail: dev-help@lucene.apache.org

Hi,

Solr does use MDC (the %X pattern), but the values are not user generated and all come from config files and are enforced to comply to certain formats (e.g., no $ possible). Shard, replica, collection names are sanitized.

In short all fine, no need to change the mitigation instructions. There is also no need to update log4j in older versions of Solr.

Uwe

Am 14. Dezember 2021 21:10:26 UTC schrieb solr <fredrik@rodland.no>:
>Only setting -Dlog4j2.formatMsgNoLookups=true might not be enough to mitigate the log4j vulnerability.
>
>See https://github.com/kmindi/log4shell-vulnerable-app
>“So even with LOG4J_FORMAT_MSG_NO_LOOKUPS true version 2.14.1 of log4j is vulnerable when using ThreadContextMap in PatternLayout.”
>
>ThreadContext.put(key, value) is used under the hood by MDC. I’m not sure wether any user-input is actually stored in MDC in SOLR.
>
>
>Probably this should be updated: https://solr.apache.org/security.html#apache-solr-affected-by-apache-log4j-cve-2021-44228
>
>And maybe consider releasing patch releases for other versions than 8.11 as well which includes log4j 2.16.0?
>
>
>
>Regards,
>
>
>Fredrik
>
>
>--
>Fredrik Rødland Cell: +47 99 21 98 17
>Maisen Pedersens vei 1 Twitter: @fredrikr
>NO-1363 Høvik, NORWAY flickr: http://www.flickr.com/fmmr/
>http://rodland.no about.me http://about.me/fmr
>
>
>
>---------------------------------------------------------------------
>To unsubscribe, e-mail: dev-unsubscribe@lucene.apache.org
>For additional commands, e-mail: dev-help@lucene.apache.org
>

--
Uwe Schindler
Achterdiek 19, 28357 Bremen
https://www.thetaphi.de

It is still a valid mitigation.

Mike Drobban I explained it. MDC is the other attack vector and that's not an issue with Solr.

Please accept this, just because the documentation of log4j changes, there's no additional risk. We may update the mitigation to mention that in Solr's case the system property is fine.

Uwe

Am 14. Dezember 2021 22:52:29 UTC schrieb solr <fredrik@rodland.no>:
>Ok.
>
>But FTR - apache/log4j has discredited just setting the system property as a mitigation measure, so I still think the SOLR security-page should be changed to not list this as a valid mitigation:
>
>https://logging.apache.org/log4j/2.x/security.html
>"Older (discredited) mitigation measures
>
>This page previously mentioned other mitigation measures, but we discovered that these measures only limit exposure while leaving some attack vectors open.
>
>Other insufficient mitigation measures are: setting system property log4j2.formatMsgNoLookups or environment variable LOG4J_FORMAT_MSG_NO_LOOKUPS to true for releases >= 2.10, or modifying the logging configuration to disable message lookups with %m{nolookups}, %msg{nolookups} or %message{nolookups} for releases >= 2.7 and <= 2.14.1.
>“
>
>Regards,
>
>
>Fredrik
>
>
>--
>Fredrik Rødland Cell: +47 99 21 98 17
>Maisen Pedersens vei 1 Twitter: @fredrikr
>NO-1363 Høvik, NORWAY flickr: http://www.flickr.com/fmmr/
>http://rodland.no about.me http://about.me/fmr
>
>
>
>> On 14 Dec 2021, at 23:44, Mike Drob <mdrob@mdrob.com> wrote:
>>
>> The MDC Patterns used by solr are for the collection, shard, replica, core and node names, and a potential trace id. All of those are restricted to alphanumeric, no special characters like $ or { needed for the injection. And trying to access a collection that didn’t exist Returns 404 without logging.
>>
>> Upgrading is always going to be more complete, but I think we’re still ok for now, at least until the next iteration of this attack surfaces.
>>
>>
>>
>> On Tue, Dec 14, 2021 at 3:37 PM solr <fredrik@rodland.no> wrote:
>> Only setting -Dlog4j2.formatMsgNoLookups=true might not be enough to mitigate the log4j vulnerability.
>>
>> See https://github.com/kmindi/log4shell-vulnerable-app
>> “So even with LOG4J_FORMAT_MSG_NO_LOOKUPS true version 2.14.1 of log4j is vulnerable when using ThreadContextMap in PatternLayout.”
>>
>> ThreadContext.put(key, value) is used under the hood by MDC. I’m not sure wether any user-input is actually stored in MDC in SOLR.
>>
>>
>> Probably this should be updated: https://solr.apache.org/security.html#apache-solr-affected-by-apache-log4j-cve-2021-44228
>>
>> And maybe consider releasing patch releases for other versions than 8.11 as well which includes log4j 2.16.0?
>>
>>
>>
>> Regards,
>>
>>
>> Fredrik
>>
>>
>> --
>> Fredrik Rødland Cell: +47 99 21 98 17
>> Maisen Pedersens vei 1 Twitter: @fredrikr
>> NO-1363 Høvik, NORWAY flickr: http://www.flickr.com/fmmr/
>> http://rodland.no about.me http://about.me/fmr
>>
>>
>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: dev-unsubscribe@lucene.apache.org
>> For additional commands, e-mail: dev-help@lucene.apache.org
>>
>
>
>---------------------------------------------------------------------
>To unsubscribe, e-mail: dev-unsubscribe@lucene.apache.org
>For additional commands, e-mail: dev-help@lucene.apache.org
>

--
Uwe Schindler
Achterdiek 19, 28357 Bremen
https://www.thetaphi.de

The other attack vectors are also not possible with Solr:

- Logger.printf("%s", userInput) is not used
- custom message factory is not used

Uwe

Am 14. Dezember 2021 22:59:26 UTC schrieb Uwe Schindler <uwe@thetaphi.de>:
>It is still a valid mitigation.
>
>Mike Drobban I explained it. MDC is the other attack vector and that's not an issue with Solr.
>
>Please accept this, just because the documentation of log4j changes, there's no additional risk. We may update the mitigation to mention that in Solr's case the system property is fine.
>
>Uwe
>
>Am 14. Dezember 2021 22:52:29 UTC schrieb solr <fredrik@rodland.no>:
>>Ok.
>>
>>But FTR - apache/log4j has discredited just setting the system property as a mitigation measure, so I still think the SOLR security-page should be changed to not list this as a valid mitigation:
>>
>>https://logging.apache.org/log4j/2.x/security.html
>>"Older (discredited) mitigation measures
>>
>>This page previously mentioned other mitigation measures, but we discovered that these measures only limit exposure while leaving some attack vectors open.
>>
>>Other insufficient mitigation measures are: setting system property log4j2.formatMsgNoLookups or environment variable LOG4J_FORMAT_MSG_NO_LOOKUPS to true for releases >= 2.10, or modifying the logging configuration to disable message lookups with %m{nolookups}, %msg{nolookups} or %message{nolookups} for releases >= 2.7 and <= 2.14.1.
>>“
>>
>>Regards,
>>
>>
>>Fredrik
>>
>>
>>--
>>Fredrik Rødland Cell: +47 99 21 98 17
>>Maisen Pedersens vei 1 Twitter: @fredrikr
>>NO-1363 Høvik, NORWAY flickr: http://www.flickr.com/fmmr/
>>http://rodland.no about.me http://about.me/fmr
>>
>>
>>
>>> On 14 Dec 2021, at 23:44, Mike Drob <mdrob@mdrob.com> wrote:
>>>
>>> The MDC Patterns used by solr are for the collection, shard, replica, core and node names, and a potential trace id. All of those are restricted to alphanumeric, no special characters like $ or { needed for the injection. And trying to access a collection that didn’t exist Returns 404 without logging.
>>>
>>> Upgrading is always going to be more complete, but I think we’re still ok for now, at least until the next iteration of this attack surfaces.
>>>
>>>
>>>
>>> On Tue, Dec 14, 2021 at 3:37 PM solr <fredrik@rodland.no> wrote:
>>> Only setting -Dlog4j2.formatMsgNoLookups=true might not be enough to mitigate the log4j vulnerability.
>>>
>>> See https://github.com/kmindi/log4shell-vulnerable-app
>>> “So even with LOG4J_FORMAT_MSG_NO_LOOKUPS true version 2.14.1 of log4j is vulnerable when using ThreadContextMap in PatternLayout.”
>>>
>>> ThreadContext.put(key, value) is used under the hood by MDC. I’m not sure wether any user-input is actually stored in MDC in SOLR.
>>>
>>>
>>> Probably this should be updated: https://solr.apache.org/security.html#apache-solr-affected-by-apache-log4j-cve-2021-44228
>>>
>>> And maybe consider releasing patch releases for other versions than 8.11 as well which includes log4j 2.16.0?
>>>
>>>
>>>
>>> Regards,
>>>
>>>
>>> Fredrik
>>>
>>>
>>> --
>>> Fredrik Rødland Cell: +47 99 21 98 17
>>> Maisen Pedersens vei 1 Twitter: @fredrikr
>>> NO-1363 Høvik, NORWAY flickr: http://www.flickr.com/fmmr/
>>> http://rodland.no about.me http://about.me/fmr
>>>
>>>
>>>
>>> ---------------------------------------------------------------------
>>> To unsubscribe, e-mail: dev-unsubscribe@lucene.apache.org
>>> For additional commands, e-mail: dev-help@lucene.apache.org
>>>
>>
>>
>>---------------------------------------------------------------------
>>To unsubscribe, e-mail: dev-unsubscribe@lucene.apache.org
>>For additional commands, e-mail: dev-help@lucene.apache.org
>>
>
>--
>Uwe Schindler
>Achterdiek 19, 28357 Bremen
>https://www.thetaphi.de
--
Uwe Schindler
Achterdiek 19, 28357 Bremen
https://www.thetaphi.de

Perhaps we could tweak it to say that the system property fix is sufficient
*for Solr* (i.e. not imply that it is a valid work around for all cases)

On Tue, Dec 14, 2021 at 6:20 PM Uwe Schindler <uwe@thetaphi.de> wrote:

> The other attack vectors are also not possible with Solr:
>
> - Logger.printf("%s", userInput) is not used
> - custom message factory is not used
>
> Uwe
>
> Am 14. Dezember 2021 22:59:26 UTC schrieb Uwe Schindler <uwe@thetaphi.de>:
>>
>> It is still a valid mitigation.
>>
>> Mike Drobban I explained it. MDC is the other attack vector and that's
>> not an issue with Solr.
>>
>> Please accept this, just because the documentation of log4j changes,
>> there's no additional risk. We may update the mitigation to mention that in
>> Solr's case the system property is fine.
>>
>> Uwe
>>
>> Am 14. Dezember 2021 22:52:29 UTC schrieb solr <fredrik@rodland.no>:
>>>
>>> Ok.
>>>
>>> But FTR - apache/log4j has discredited just setting the system property as a mitigation measure, so I still think the SOLR security-page should be changed to not list this as a valid mitigation:
>>>
>>> https://logging.apache.org/log4j/2.x/security.html
>>> "Older (discredited) mitigation measures
>>>
>>> This page previously mentioned other mitigation measures, but we discovered that these measures only limit exposure while leaving some attack vectors open.
>>>
>>> Other insufficient mitigation measures are: setting system property log4j2.formatMsgNoLookups or environment variable LOG4J_FORMAT_MSG_NO_LOOKUPS to true for releases >= 2.10, or modifying the logging configuration to disable message lookups with %m{nolookups}, %msg{nolookups} or %message{nolookups} for releases >= 2.7 and <= 2.14.1.
>>> “
>>>
>>> Regards,
>>>
>>>
>>> Fredrik
>>>
>>>
>>> --
>>> Fredrik Rødland Cell: +47 99 21 98 17
>>> Maisen Pedersens vei 1 Twitter: @fredrikr
>>> NO-1363 Høvik, NORWAY flickr: http://www.flickr.com/fmmr/
>>> http://rodland.no about.me http://about.me/fmr
>>>
>>>
>>>
>>> On 14 Dec 2021, at 23:44, Mike Drob <mdrob@mdrob.com> wrote:
>>>>
>>>> The MDC Patterns used by solr are for the collection, shard, replica, core and node names, and a potential trace id. All of those are restricted to alphanumeric, no special characters like $ or { needed for the injection. And trying to access a collection that didn’t exist Returns 404 without logging.
>>>>
>>>> Upgrading is always going to be more complete, but I think we’re still ok for now, at least until the next iteration of this attack surfaces.
>>>>
>>>>
>>>>
>>>> On Tue, Dec 14, 2021 at 3:37 PM solr <fredrik@rodland.no> wrote:
>>>> Only setting -Dlog4j2.formatMsgNoLookups=true might not be enough to mitigate the log4j vulnerability.
>>>>
>>>> See https://github.com/kmindi/log4shell-vulnerable-app
>>>> “So even with LOG4J_FORMAT_MSG_NO_LOOKUPS true version 2.14.1 of log4j is vulnerable when using ThreadContextMap in PatternLayout.”
>>>>
>>>> ThreadContext.put(key, value) is used under the hood by MDC. I’m not sure wether any user-input is actually stored in MDC in SOLR.
>>>>
>>>>
>>>> Probably this should be updated: https://solr.apache.org/security.html#apache-solr-affected-by-apache-log4j-cve-2021-44228
>>>>
>>>> And maybe consider releasing patch releases for other versions than 8.11 as well which includes log4j 2.16.0?
>>>>
>>>>
>>>>
>>>> Regards,
>>>>
>>>>
>>>> Fredrik
>>>>
>>>>
>>>> --
>>>> Fredrik Rødland Cell: +47 99 21 98 17
>>>> Maisen Pedersens vei 1 Twitter: @fredrikr
>>>> NO-1363 Høvik, NORWAY flickr: http://www.flickr.com/fmmr/
>>>> http://rodland.no about.me http://about.me/fmr
>>>> ------------------------------
>>>> To unsubscribe, e-mail: dev-unsubscribe@lucene.apache.org
>>>> For additional commands, e-mail: dev-help@lucene.apache.org
>>>>
>>>> ------------------------------
>>> To unsubscribe, e-mail: dev-unsubscribe@lucene.apache.org
>>> For additional commands, e-mail: dev-help@lucene.apache.org
>>>
>>> --
>> Uwe Schindler
>> Achterdiek 19, 28357 Bremen
>> https://www.thetaphi.de
>>
> --
> Uwe Schindler
> Achterdiek 19, 28357 Bremen
> https://www.thetaphi.de
>


--
http://www.needhamsoftware.com (work)
http://www.the111shift.com (play)

We should add this to the webpage. Another one asked on the security mailing list.



Uwe



-----

Uwe Schindler

Achterdiek 19, D-28357 Bremen

https://www.thetaphi.de

eMail: uwe@thetaphi.de



From: Gus Heck <gus.heck@gmail.com>
Sent: Wednesday, December 15, 2021 12:39 AM
To: dev <dev@lucene.apache.org>
Subject: Re: Log4j < 2.15.0 may still be vulnerable even if -Dlog4j2.formatMsgNoLookups=true is set



Perhaps we could tweak it to say that the system property fix is sufficient *for Solr* (i.e. not imply that it is a valid work around for all cases)



On Tue, Dec 14, 2021 at 6:20 PM Uwe Schindler <uwe@thetaphi.de <mailto:uwe@thetaphi.de> > wrote:

The other attack vectors are also not possible with Solr:

- Logger.printf("%s", userInput) is not used
- custom message factory is not used

Uwe

Am 14. Dezember 2021 22:59:26 UTC schrieb Uwe Schindler <uwe@thetaphi.de <mailto:uwe@thetaphi.de> >:

It is still a valid mitigation.

Mike Drobban I explained it. MDC is the other attack vector and that's not an issue with Solr.

Please accept this, just because the documentation of log4j changes, there's no additional risk. We may update the mitigation to mention that in Solr's case the system property is fine.

Uwe

Am 14. Dezember 2021 22:52:29 UTC schrieb solr <fredrik@rodland.no <mailto:fredrik@rodland.no> >:

Ok.

But FTR - apache/log4j has discredited just setting the system property as a mitigation measure, so I still think the SOLR security-page should be changed to not list this as a valid mitigation:

https://logging.apache.org/log4j/2.x/security.html
"Older (discredited) mitigation measures

This page previously mentioned other mitigation measures, but we discovered that these measures only limit exposure while leaving some attack vectors open.

Other insufficient mitigation measures are: setting system property log4j2.formatMsgNoLookups or environment variable LOG4J_FORMAT_MSG_NO_LOOKUPS to true for releases >= 2.10, or modifying the logging configuration to disable message lookups with %m{nolookups}, %msg{nolookups} or %message{nolookups} for releases >= 2.7 and <= 2.14.1.
“

Regards,


Fredrik


--
Fredrik Rødland Cell: +47 99 21 98 17
Maisen Pedersens vei 1 Twitter: @fredrikr
NO-1363 Høvik, NORWAY flickr: http://www.flickr.com/fmmr/
http://rodland.no about.me <http://about.me> http://about.me/fmr




On 14 Dec 2021, at 23:44, Mike Drob <mdrob@mdrob.com <mailto:mdrob@mdrob.com> > wrote:

The MDC Patterns used by solr are for the collection, shard, replica, core and node names, and a potential trace id. All of those are restricted to alphanumeric, no special characters like $ or { needed for the injection. And trying to access a collection that didn’t exist Returns 404 without logging.

Upgrading is always going to be more complete, but I think we’re still ok for now, at least until the next iteration of this attack surfaces.



On Tue, Dec 14, 2021 at 3:37 PM solr <fredrik@rodland.no <mailto:fredrik@rodland.no> > wrote:
Only setting -Dlog4j2.formatMsgNoLookups=true might not be enough to mitigate the log4j vulnerability.

See https://github.com/kmindi/log4shell-vulnerable-app
“So even with LOG4J_FORMAT_MSG_NO_LOOKUPS true version 2.14.1 of log4j is vulnerable when using ThreadContextMap in PatternLayout.”

ThreadContext.put(key, value) is used under the hood by MDC. I’m not sure wether any user-input is actually stored in MDC in SOLR.


Probably this should be updated: https://solr.apache.org/security.html#apache-solr-affected-by-apache-log4j-cve-2021-44228

And maybe consider releasing patch releases for other versions than 8.11 as well which includes log4j 2.16.0?



Regards,


Fredrik


--
Fredrik Rødland Cell: +47 99 21 98 17
Maisen Pedersens vei 1 Twitter: @fredrikr
NO-1363 Høvik, NORWAY flickr: http://www.flickr.com/fmmr/
http://rodland.no about.me <http://about.me> http://about.me/fmr


_____

To unsubscribe, e-mail: dev-unsubscribe@lucene.apache.org <mailto:dev-unsubscribe@lucene.apache.org>
For additional commands, e-mail: dev-help@lucene.apache.org <mailto:dev-help@lucene.apache.org>



_____

To unsubscribe, e-mail: dev-unsubscribe@lucene.apache.org <mailto:dev-unsubscribe@lucene.apache.org>
For additional commands, e-mail: dev-help@lucene.apache.org <mailto:dev-help@lucene.apache.org>

--
Uwe Schindler
Achterdiek 19, 28357 Bremen
https://www.thetaphi.de

--
Uwe Schindler
Achterdiek 19, 28357 Bremen
https://www.thetaphi.de






--

http://www.needhamsoftware.com (work)

http://www.the111shift.com (play)

Hi all, I prepared a PR about the followup CVE-2021-45046: https://github.com/apache/solr-site/pull/59



Please verify and make suggestion. I will merge this into main/production later.



Uwe



-----

Uwe Schindler

Achterdiek 19, D-28357 Bremen

https://www.thetaphi.de

eMail: uwe@thetaphi.de



From: Uwe Schindler <uwe@thetaphi.de>
Sent: Wednesday, December 15, 2021 3:31 PM
To: 'dev@lucene.apache.org' <dev@lucene.apache.org>
Subject: RE: Log4j < 2.15.0 may still be vulnerable even if -Dlog4j2.formatMsgNoLookups=true is set



We should add this to the webpage. Another one asked on the security mailing list.



Uwe



-----

Uwe Schindler

Achterdiek 19, D-28357 Bremen

https://www.thetaphi.de

eMail: uwe@thetaphi.de <mailto:uwe@thetaphi.de>



From: Gus Heck <gus.heck@gmail.com <mailto:gus.heck@gmail.com> >
Sent: Wednesday, December 15, 2021 12:39 AM
To: dev <dev@lucene.apache.org <mailto:dev@lucene.apache.org> >
Subject: Re: Log4j < 2.15.0 may still be vulnerable even if -Dlog4j2.formatMsgNoLookups=true is set



Perhaps we could tweak it to say that the system property fix is sufficient *for Solr* (i.e. not imply that it is a valid work around for all cases)



On Tue, Dec 14, 2021 at 6:20 PM Uwe Schindler <uwe@thetaphi.de <mailto:uwe@thetaphi.de> > wrote:

The other attack vectors are also not possible with Solr:

- Logger.printf("%s", userInput) is not used
- custom message factory is not used

Uwe

Am 14. Dezember 2021 22:59:26 UTC schrieb Uwe Schindler <uwe@thetaphi.de <mailto:uwe@thetaphi.de> >:

It is still a valid mitigation.

Mike Drobban I explained it. MDC is the other attack vector and that's not an issue with Solr.

Please accept this, just because the documentation of log4j changes, there's no additional risk. We may update the mitigation to mention that in Solr's case the system property is fine.

Uwe

Am 14. Dezember 2021 22:52:29 UTC schrieb solr <fredrik@rodland.no <mailto:fredrik@rodland.no> >:

Ok.

But FTR - apache/log4j has discredited just setting the system property as a mitigation measure, so I still think the SOLR security-page should be changed to not list this as a valid mitigation:

https://logging.apache.org/log4j/2.x/security.html
"Older (discredited) mitigation measures

This page previously mentioned other mitigation measures, but we discovered that these measures only limit exposure while leaving some attack vectors open.

Other insufficient mitigation measures are: setting system property log4j2.formatMsgNoLookups or environment variable LOG4J_FORMAT_MSG_NO_LOOKUPS to true for releases >= 2.10, or modifying the logging configuration to disable message lookups with %m{nolookups}, %msg{nolookups} or %message{nolookups} for releases >= 2.7 and <= 2.14.1.
“

Regards,


Fredrik


--
Fredrik Rødland Cell: +47 99 21 98 17
Maisen Pedersens vei 1 Twitter: @fredrikr
NO-1363 Høvik, NORWAY flickr: http://www.flickr.com/fmmr/
http://rodland.no about.me <http://about.me> http://about.me/fmr



On 14 Dec 2021, at 23:44, Mike Drob <mdrob@mdrob.com <mailto:mdrob@mdrob.com> > wrote:

The MDC Patterns used by solr are for the collection, shard, replica, core and node names, and a potential trace id. All of those are restricted to alphanumeric, no special characters like $ or { needed for the injection. And trying to access a collection that didn’t exist Returns 404 without logging.

Upgrading is always going to be more complete, but I think we’re still ok for now, at least until the next iteration of this attack surfaces.



On Tue, Dec 14, 2021 at 3:37 PM solr <fredrik@rodland.no <mailto:fredrik@rodland.no> > wrote:
Only setting -Dlog4j2.formatMsgNoLookups=true might not be enough to mitigate the log4j vulnerability.

See https://github.com/kmindi/log4shell-vulnerable-app
“So even with LOG4J_FORMAT_MSG_NO_LOOKUPS true version 2.14.1 of log4j is vulnerable when using ThreadContextMap in PatternLayout.”

ThreadContext.put(key, value) is used under the hood by MDC. I’m not sure wether any user-input is actually stored in MDC in SOLR.


Probably this should be updated: https://solr.apache.org/security.html#apache-solr-affected-by-apache-log4j-cve-2021-44228

And maybe consider releasing patch releases for other versions than 8.11 as well which includes log4j 2.16.0?



Regards,


Fredrik


--
Fredrik Rødland Cell: +47 99 21 98 17
Maisen Pedersens vei 1 Twitter: @fredrikr
NO-1363 Høvik, NORWAY flickr: http://www.flickr.com/fmmr/
http://rodland.no about.me <http://about.me> http://about.me/fmr


_____

To unsubscribe, e-mail: dev-unsubscribe@lucene.apache.org <mailto:dev-unsubscribe@lucene.apache.org>
For additional commands, e-mail: dev-help@lucene.apache.org <mailto:dev-help@lucene.apache.org>



_____

To unsubscribe, e-mail: dev-unsubscribe@lucene.apache.org <mailto:dev-unsubscribe@lucene.apache.org>
For additional commands, e-mail: dev-help@lucene.apache.org <mailto:dev-help@lucene.apache.org>

--
Uwe Schindler
Achterdiek 19, 28357 Bremen
https://www.thetaphi.de

--
Uwe Schindler
Achterdiek 19, 28357 Bremen
https://www.thetaphi.de






--

http://www.needhamsoftware.com (work)

http://www.the111shift.com (play)

We make edits to the log4j advisory almost daily, see https://github.com/apache/solr-site/commits/e10a6a9fe0eed8dcba3ad1a076c8208e014e76ff/content/solr/security/2021-12-10-cve-2021-44228.md
I wonder if we should include a "Revision history" paragraph in the advisory for transparency?

Jan

> 15. des. 2021 kl. 19:09 skrev Uwe Schindler <uwe@thetaphi.de>:
>
> Hi all, I prepared a PR about the followup CVE-2021-45046: https://github.com/apache/solr-site/pull/59 <https://github.com/apache/solr-site/pull/59>
>
> Please verify and make suggestion. I will merge this into main/production later.
>
> Uwe
>
> -----
> Uwe Schindler
> Achterdiek 19, D-28357 Bremen
> https://www.thetaphi.de <https://www.thetaphi.de/>
> eMail: uwe@thetaphi.de <mailto:uwe@thetaphi.de>
>
> From: Uwe Schindler <uwe@thetaphi.de <mailto:uwe@thetaphi.de>>
> Sent: Wednesday, December 15, 2021 3:31 PM
> To: 'dev@lucene.apache.org <mailto:dev@lucene.apache.org>' <dev@lucene.apache.org <mailto:dev@lucene.apache.org>>
> Subject: RE: Log4j < 2.15.0 may still be vulnerable even if -Dlog4j2.formatMsgNoLookups=true is set
>
> We should add this to the webpage. Another one asked on the security mailing list.
>
> Uwe
>
> -----
> Uwe Schindler
> Achterdiek 19, D-28357 Bremen
> https://www.thetaphi.de <https://www.thetaphi.de/>
> eMail: uwe@thetaphi.de <mailto:uwe@thetaphi.de>
>
> From: Gus Heck <gus.heck@gmail.com <mailto:gus.heck@gmail.com>>
> Sent: Wednesday, December 15, 2021 12:39 AM
> To: dev <dev@lucene.apache.org <mailto:dev@lucene.apache.org>>
> Subject: Re: Log4j < 2.15.0 may still be vulnerable even if -Dlog4j2.formatMsgNoLookups=true is set
>
> Perhaps we could tweak it to say that the system property fix is sufficient *for Solr* (i.e. not imply that it is a valid work around for all cases)
>
> On Tue, Dec 14, 2021 at 6:20 PM Uwe Schindler <uwe@thetaphi.de <mailto:uwe@thetaphi.de>> wrote:
>> The other attack vectors are also not possible with Solr:
>>
>> - Logger.printf("%s", userInput) is not used
>> - custom message factory is not used
>>
>> Uwe
>>
>> Am 14. Dezember 2021 22:59:26 UTC schrieb Uwe Schindler <uwe@thetaphi.de <mailto:uwe@thetaphi.de>>:
>>> It is still a valid mitigation.
>>>
>>> Mike Drobban I explained it. MDC is the other attack vector and that's not an issue with Solr.
>>>
>>> Please accept this, just because the documentation of log4j changes, there's no additional risk. We may update the mitigation to mention that in Solr's case the system property is fine.
>>>
>>> Uwe
>>>
>>> Am 14. Dezember 2021 22:52:29 UTC schrieb solr <fredrik@rodland.no <mailto:fredrik@rodland.no>>:
>>>> Ok.
>>>>
>>>> But FTR - apache/log4j has discredited just setting the system property as a mitigation measure, so I still think the SOLR security-page should be changed to not list this as a valid mitigation:
>>>>
>>>> https://logging.apache.org/log4j/2.x/security.html <https://logging.apache.org/log4j/2.x/security.html>
>>>> "Older (discredited) mitigation measures
>>>>
>>>> This page previously mentioned other mitigation measures, but we discovered that these measures only limit exposure while leaving some attack vectors open.
>>>>
>>>> Other insufficient mitigation measures are: setting system property log4j2.formatMsgNoLookups or environment variable LOG4J_FORMAT_MSG_NO_LOOKUPS to true for releases >= 2.10, or modifying the logging configuration to disable message lookups with %m{nolookups}, %msg{nolookups} or %message{nolookups} for releases >= 2.7 and <= 2.14.1.
>>>> “
>>>>
>>>> Regards,
>>>>
>>>>
>>>> Fredrik
>>>>
>>>>
>>>> --
>>>> Fredrik Rødland Cell: +47 99 21 98 17
>>>> Maisen Pedersens vei 1 Twitter: @fredrikr
>>>> NO-1363 Høvik, NORWAY flickr: http://www.flickr.com/fmmr/ <http://www.flickr.com/fmmr/>
>>>> http://rodland.no <http://rodland.no/> about.me <http://about.me/> http://about.me/fmr <http://about.me/fmr>
>>>>
>>>>> On 14 Dec 2021, at 23:44, Mike Drob <mdrob@mdrob.com <mailto:mdrob@mdrob.com>> wrote:
>>>>>
>>>>> The MDC Patterns used by solr are for the collection, shard, replica, core and node names, and a potential trace id. All of those are restricted to alphanumeric, no special characters like $ or { needed for the injection. And trying to access a collection that didn’t exist Returns 404 without logging.
>>>>>
>>>>> Upgrading is always going to be more complete, but I think we’re still ok for now, at least until the next iteration of this attack surfaces.
>>>>>
>>>>>
>>>>>
>>>>> On Tue, Dec 14, 2021 at 3:37 PM solr <fredrik@rodland.no <mailto:fredrik@rodland.no>> wrote:
>>>>> Only setting -Dlog4j2.formatMsgNoLookups=true might not be enough to mitigate the log4j vulnerability.
>>>>>
>>>>> See https://github.com/kmindi/log4shell-vulnerable-app <https://github.com/kmindi/log4shell-vulnerable-app>
>>>>> “So even with LOG4J_FORMAT_MSG_NO_LOOKUPS true version 2.14.1 of log4j is vulnerable when using ThreadContextMap in PatternLayout.”
>>>>>
>>>>> ThreadContext.put(key, value) is used under the hood by MDC. I’m not sure wether any user-input is actually stored in MDC in SOLR.
>>>>>
>>>>>
>>>>> Probably this should be updated: https://solr.apache.org/security.html#apache-solr-affected-by-apache-log4j-cve-2021-44228 <https://solr.apache.org/security.html#apache-solr-affected-by-apache-log4j-cve-2021-44228>
>>>>>
>>>>> And maybe consider releasing patch releases for other versions than 8.11 as well which includes log4j 2.16.0?
>>>>>
>>>>>
>>>>>
>>>>> Regards,
>>>>>
>>>>>
>>>>> Fredrik
>>>>>
>>>>>
>>>>> --
>>>>> Fredrik Rødland Cell: +47 99 21 98 17
>>>>> Maisen Pedersens vei 1 Twitter: @fredrikr
>>>>> NO-1363 Høvik, NORWAY flickr: http://www.flickr.com/fmmr/ <http://www.flickr.com/fmmr/>
>>>>> http://rodland.no <http://rodland.no/> about.me <http://about.me/> http://about.me/fmr <http://about.me/fmr>
>>>>> To unsubscribe, e-mail: dev-unsubscribe@lucene.apache.org <mailto:dev-unsubscribe@lucene.apache.org>
>>>>> For additional commands, e-mail: dev-help@lucene.apache.org <mailto:dev-help@lucene.apache.org>
>>>> To unsubscribe, e-mail: dev-unsubscribe@lucene.apache.org <mailto:dev-unsubscribe@lucene.apache.org>
>>>> For additional commands, e-mail: dev-help@lucene.apache.org <mailto:dev-help@lucene.apache.org>
>>> --
>>> Uwe Schindler
>>> Achterdiek 19, 28357 Bremen
>>> https://www.thetaphi.de <https://www.thetaphi.de/>
>> --
>> Uwe Schindler
>> Achterdiek 19, 28357 Bremen
>> https://www.thetaphi.de <https://www.thetaphi.de/>
>
>
> --
> http://www.needhamsoftware.com <http://www.needhamsoftware.com/> (work)
> http://www.the111shift.com <http://www.the111shift.com/> (play)

I think perhaps in the shock of such a deep and surprising vulnerability
with such high visibility, we've begun to break with how we normally handle
CVE's that don't apply to our usage of the library. Previously, they just
got added to the list of known false positives
<https://cwiki.apache.org/confluence/display/SOLR/SolrSecurity#SolrSecurity-SolrandVulnerabilityScanningTools>.
Normally we wouldn't even mention them on the security news page, but
because of the high visibility we should simply have a line mentioning that
these two CVE's are on our false positives page and explain details there.
The wiki would provide revision history automatically.

On Sat, Dec 18, 2021 at 11:25 AM Jan Høydahl <jan.asf@cominvent.com> wrote:

> We make edits to the log4j advisory almost daily, see
> https://github.com/apache/solr-site/commits/e10a6a9fe0eed8dcba3ad1a076c8208e014e76ff/content/solr/security/2021-12-10-cve-2021-44228.md
> I wonder if we should include a "Revision history" paragraph in the
> advisory for transparency?
>
> Jan
>
> 15. des. 2021 kl. 19:09 skrev Uwe Schindler <uwe@thetaphi.de>:
>
> Hi all, I prepared a PR about the followup CVE-2021-45046:
> https://github.com/apache/solr-site/pull/59
>
> Please verify and make suggestion. I will merge this into main/production
> later.
>
> Uwe
>
> -----
> Uwe Schindler
> Achterdiek 19, D-28357 Bremen
> https://www.thetaphi.de
> eMail: uwe@thetaphi.de
>
> *From:* Uwe Schindler <uwe@thetaphi.de>
> *Sent:* Wednesday, December 15, 2021 3:31 PM
> *To:* 'dev@lucene.apache.org' <dev@lucene.apache.org>
> *Subject:* RE: Log4j < 2.15.0 may still be vulnerable even if
> -Dlog4j2.formatMsgNoLookups=true is set
>
> We should add this to the webpage. Another one asked on the security
> mailing list.
>
> Uwe
>
> -----
> Uwe Schindler
> Achterdiek 19, D-28357 Bremen
> https://www.thetaphi.de
> eMail: uwe@thetaphi.de
>
> *From:* Gus Heck <gus.heck@gmail.com>
> *Sent:* Wednesday, December 15, 2021 12:39 AM
> *To:* dev <dev@lucene.apache.org>
> *Subject:* Re: Log4j < 2.15.0 may still be vulnerable even if
> -Dlog4j2.formatMsgNoLookups=true is set
>
> Perhaps we could tweak it to say that the system property fix is
> sufficient *for Solr* (i.e. not imply that it is a valid work around for
> all cases)
>
> On Tue, Dec 14, 2021 at 6:20 PM Uwe Schindler <uwe@thetaphi.de> wrote:
>
> The other attack vectors are also not possible with Solr:
>
> - Logger.printf("%s", userInput) is not used
> - custom message factory is not used
>
> Uwe
> Am 14. Dezember 2021 22:59:26 UTC schrieb Uwe Schindler <uwe@thetaphi.de>:
>
> It is still a valid mitigation.
>
> Mike Drobban I explained it. MDC is the other attack vector and that's not
> an issue with Solr.
>
> Please accept this, just because the documentation of log4j changes,
> there's no additional risk. We may update the mitigation to mention that in
> Solr's case the system property is fine.
>
> Uwe
> Am 14. Dezember 2021 22:52:29 UTC schrieb solr <fredrik@rodland.no>:
>
> Ok.
>
> But FTR - apache/log4j has discredited just setting the system property as a mitigation measure, so I still think the SOLR security-page should be changed to not list this as a valid mitigation:
>
> https://logging.apache.org/log4j/2.x/security.html
> "Older (discredited) mitigation measures
>
> This page previously mentioned other mitigation measures, but we discovered that these measures only limit exposure while leaving some attack vectors open.
>
> Other insufficient mitigation measures are: setting system property log4j2.formatMsgNoLookups or environment variable LOG4J_FORMAT_MSG_NO_LOOKUPS to true for releases >= 2.10, or modifying the logging configuration to disable message lookups with %m{nolookups}, %msg{nolookups} or %message{nolookups} for releases >= 2.7 and <= 2.14.1.
> “
>
> Regards,
>
>
> Fredrik
>
>
> --
> Fredrik Rødland Cell: +47 99 21 98 17
> Maisen Pedersens vei 1 Twitter: @fredrikr
> NO-1363 Høvik, NORWAY flickr: http://www.flickr.com/fmmr/
> http://rodland.no about.me http://about.me/fmr
>
> On 14 Dec 2021, at 23:44, Mike Drob <mdrob@mdrob.com> wrote:
>
> The MDC Patterns used by solr are for the collection, shard, replica, core and node names, and a potential trace id. All of those are restricted to alphanumeric, no special characters like $ or { needed for the injection. And trying to access a collection that didn’t exist Returns 404 without logging.
>
> Upgrading is always going to be more complete, but I think we’re still ok for now, at least until the next iteration of this attack surfaces.
>
>
>
> On Tue, Dec 14, 2021 at 3:37 PM solr <fredrik@rodland.no> wrote:
> Only setting -Dlog4j2.formatMsgNoLookups=true might not be enough to mitigate the log4j vulnerability.
>
> See https://github.com/kmindi/log4shell-vulnerable-app
> “So even with LOG4J_FORMAT_MSG_NO_LOOKUPS true version 2.14.1 of log4j is vulnerable when using ThreadContextMap in PatternLayout.”
>
> ThreadContext.put(key, value) is used under the hood by MDC. I’m not sure wether any user-input is actually stored in MDC in SOLR.
>
>
> Probably this should be updated: https://solr.apache.org/security.html#apache-solr-affected-by-apache-log4j-cve-2021-44228
>
> And maybe consider releasing patch releases for other versions than 8.11 as well which includes log4j 2.16.0?
>
>
>
> Regards,
>
>
> Fredrik
>
>
> --
> Fredrik Rødland Cell: +47 99 21 98 17
> Maisen Pedersens vei 1 Twitter: @fredrikr
> NO-1363 Høvik, NORWAY flickr: http://www.flickr.com/fmmr/
> http://rodland.no about.me http://about.me/fmr
>
> ------------------------------
>
> To unsubscribe, e-mail: dev-unsubscribe@lucene.apache.org
> For additional commands, e-mail: dev-help@lucene.apache.org
>
> ------------------------------
>
> To unsubscribe, e-mail: dev-unsubscribe@lucene.apache.org
> For additional commands, e-mail: dev-help@lucene.apache.org
>
> --
> Uwe Schindler
> Achterdiek 19, 28357 Bremen
> https://www.thetaphi.de
>
> --
> Uwe Schindler
> Achterdiek 19, 28357 Bremen
> https://www.thetaphi.de
>
>
>
> --
> http://www.needhamsoftware.com (work)
> http://www.the111shift.com (play)
>
>
>

--
http://www.needhamsoftware.com (work)
http://www.the111shift.com (play)

Thinking about it some more, maybe the problem with my suggestion is
the table on that page is organized by the library version and, if
unmitigated, the version of the library is still a problem. Maybe another
way to be clearer about it and avoid rewriting things that people have
already read would be to add independent entries to the security news page
for the newer CVE's

On Sat, Dec 18, 2021 at 12:20 PM Gus Heck <gus.heck@gmail.com> wrote:

> I think perhaps in the shock of such a deep and surprising vulnerability
> with such high visibility, we've begun to break with how we normally handle
> CVE's that don't apply to our usage of the library. Previously, they just
> got added to the list of known false positives
> <https://cwiki.apache.org/confluence/display/SOLR/SolrSecurity#SolrSecurity-SolrandVulnerabilityScanningTools>.
> Normally we wouldn't even mention them on the security news page, but
> because of the high visibility we should simply have a line mentioning that
> these two CVE's are on our false positives page and explain details there.
> The wiki would provide revision history automatically.
>
> On Sat, Dec 18, 2021 at 11:25 AM Jan Høydahl <jan.asf@cominvent.com>
> wrote:
>
>> We make edits to the log4j advisory almost daily, see
>> https://github.com/apache/solr-site/commits/e10a6a9fe0eed8dcba3ad1a076c8208e014e76ff/content/solr/security/2021-12-10-cve-2021-44228.md
>> I wonder if we should include a "Revision history" paragraph in the
>> advisory for transparency?
>>
>> Jan
>>
>> 15. des. 2021 kl. 19:09 skrev Uwe Schindler <uwe@thetaphi.de>:
>>
>> Hi all, I prepared a PR about the followup CVE-2021-45046:
>> https://github.com/apache/solr-site/pull/59
>>
>> Please verify and make suggestion. I will merge this into main/production
>> later.
>>
>> Uwe
>>
>> -----
>> Uwe Schindler
>> Achterdiek 19, D-28357 Bremen
>> https://www.thetaphi.de
>> eMail: uwe@thetaphi.de
>>
>> *From:* Uwe Schindler <uwe@thetaphi.de>
>> *Sent:* Wednesday, December 15, 2021 3:31 PM
>> *To:* 'dev@lucene.apache.org' <dev@lucene.apache.org>
>> *Subject:* RE: Log4j < 2.15.0 may still be vulnerable even if
>> -Dlog4j2.formatMsgNoLookups=true is set
>>
>> We should add this to the webpage. Another one asked on the security
>> mailing list.
>>
>> Uwe
>>
>> -----
>> Uwe Schindler
>> Achterdiek 19, D-28357 Bremen
>> https://www.thetaphi.de
>> eMail: uwe@thetaphi.de
>>
>> *From:* Gus Heck <gus.heck@gmail.com>
>> *Sent:* Wednesday, December 15, 2021 12:39 AM
>> *To:* dev <dev@lucene.apache.org>
>> *Subject:* Re: Log4j < 2.15.0 may still be vulnerable even if
>> -Dlog4j2.formatMsgNoLookups=true is set
>>
>> Perhaps we could tweak it to say that the system property fix is
>> sufficient *for Solr* (i.e. not imply that it is a valid work around for
>> all cases)
>>
>> On Tue, Dec 14, 2021 at 6:20 PM Uwe Schindler <uwe@thetaphi.de> wrote:
>>
>> The other attack vectors are also not possible with Solr:
>>
>> - Logger.printf("%s", userInput) is not used
>> - custom message factory is not used
>>
>> Uwe
>> Am 14. Dezember 2021 22:59:26 UTC schrieb Uwe Schindler <uwe@thetaphi.de
>> >:
>>
>> It is still a valid mitigation.
>>
>> Mike Drobban I explained it. MDC is the other attack vector and that's
>> not an issue with Solr.
>>
>> Please accept this, just because the documentation of log4j changes,
>> there's no additional risk. We may update the mitigation to mention that in
>> Solr's case the system property is fine.
>>
>> Uwe
>> Am 14. Dezember 2021 22:52:29 UTC schrieb solr <fredrik@rodland.no>:
>>
>> Ok.
>>
>> But FTR - apache/log4j has discredited just setting the system property as a mitigation measure, so I still think the SOLR security-page should be changed to not list this as a valid mitigation:
>>
>> https://logging.apache.org/log4j/2.x/security.html
>> "Older (discredited) mitigation measures
>>
>> This page previously mentioned other mitigation measures, but we discovered that these measures only limit exposure while leaving some attack vectors open.
>>
>> Other insufficient mitigation measures are: setting system property log4j2.formatMsgNoLookups or environment variable LOG4J_FORMAT_MSG_NO_LOOKUPS to true for releases >= 2.10, or modifying the logging configuration to disable message lookups with %m{nolookups}, %msg{nolookups} or %message{nolookups} for releases >= 2.7 and <= 2.14.1.
>> “
>>
>> Regards,
>>
>>
>> Fredrik
>>
>>
>> --
>> Fredrik Rødland Cell: +47 99 21 98 17
>> Maisen Pedersens vei 1 Twitter: @fredrikr
>> NO-1363 Høvik, NORWAY flickr: http://www.flickr.com/fmmr/
>> http://rodland.no about.me http://about.me/fmr
>>
>> On 14 Dec 2021, at 23:44, Mike Drob <mdrob@mdrob.com> wrote:
>>
>> The MDC Patterns used by solr are for the collection, shard, replica, core and node names, and a potential trace id. All of those are restricted to alphanumeric, no special characters like $ or { needed for the injection. And trying to access a collection that didn’t exist Returns 404 without logging.
>>
>> Upgrading is always going to be more complete, but I think we’re still ok for now, at least until the next iteration of this attack surfaces.
>>
>>
>>
>> On Tue, Dec 14, 2021 at 3:37 PM solr <fredrik@rodland.no> wrote:
>> Only setting -Dlog4j2.formatMsgNoLookups=true might not be enough to mitigate the log4j vulnerability.
>>
>> See https://github.com/kmindi/log4shell-vulnerable-app
>> “So even with LOG4J_FORMAT_MSG_NO_LOOKUPS true version 2.14.1 of log4j is vulnerable when using ThreadContextMap in PatternLayout.”
>>
>> ThreadContext.put(key, value) is used under the hood by MDC. I’m not sure wether any user-input is actually stored in MDC in SOLR.
>>
>>
>> Probably this should be updated: https://solr.apache.org/security.html#apache-solr-affected-by-apache-log4j-cve-2021-44228
>>
>> And maybe consider releasing patch releases for other versions than 8.11 as well which includes log4j 2.16.0?
>>
>>
>>
>> Regards,
>>
>>
>> Fredrik
>>
>>
>> --
>> Fredrik Rødland Cell: +47 99 21 98 17
>> Maisen Pedersens vei 1 Twitter: @fredrikr
>> NO-1363 Høvik, NORWAY flickr: http://www.flickr.com/fmmr/
>> http://rodland.no about.me http://about.me/fmr
>>
>> ------------------------------
>>
>> To unsubscribe, e-mail: dev-unsubscribe@lucene.apache.org
>> For additional commands, e-mail: dev-help@lucene.apache.org
>>
>> ------------------------------
>>
>> To unsubscribe, e-mail: dev-unsubscribe@lucene.apache.org
>> For additional commands, e-mail: dev-help@lucene.apache.org
>>
>> --
>> Uwe Schindler
>> Achterdiek 19, 28357 Bremen
>> https://www.thetaphi.de
>>
>> --
>> Uwe Schindler
>> Achterdiek 19, 28357 Bremen
>> https://www.thetaphi.de
>>
>>
>>
>> --
>> http://www.needhamsoftware.com (work)
>> http://www.the111shift.com (play)
>>
>>
>>
>
> --
> http://www.needhamsoftware.com (work)
> http://www.the111shift.com (play)
>


--
http://www.needhamsoftware.com (work)
http://www.the111shift.com (play)

I like the idea of using our Wiki more as you describe. Not so much
*new* news entries because I think search-ability of these CVEs is fine to
an existing entry.

~ David Smiley
Apache Lucene/Solr Search Developer
http://www.linkedin.com/in/davidwsmiley


On Sat, Dec 18, 2021 at 4:39 PM Gus Heck <gus.heck@gmail.com> wrote:

> Thinking about it some more, maybe the problem with my suggestion is
> the table on that page is organized by the library version and, if
> unmitigated, the version of the library is still a problem. Maybe another
> way to be clearer about it and avoid rewriting things that people have
> already read would be to add independent entries to the security news page
> for the newer CVE's
>
> On Sat, Dec 18, 2021 at 12:20 PM Gus Heck <gus.heck@gmail.com> wrote:
>
>> I think perhaps in the shock of such a deep and surprising vulnerability
>> with such high visibility, we've begun to break with how we normally handle
>> CVE's that don't apply to our usage of the library. Previously, they just
>> got added to the list of known false positives
>> <https://cwiki.apache.org/confluence/display/SOLR/SolrSecurity#SolrSecurity-SolrandVulnerabilityScanningTools>.
>> Normally we wouldn't even mention them on the security news page, but
>> because of the high visibility we should simply have a line mentioning that
>> these two CVE's are on our false positives page and explain details there.
>> The wiki would provide revision history automatically.
>>
>> On Sat, Dec 18, 2021 at 11:25 AM Jan Høydahl <jan.asf@cominvent.com>
>> wrote:
>>
>>> We make edits to the log4j advisory almost daily, see
>>> https://github.com/apache/solr-site/commits/e10a6a9fe0eed8dcba3ad1a076c8208e014e76ff/content/solr/security/2021-12-10-cve-2021-44228.md
>>> I wonder if we should include a "Revision history" paragraph in the
>>> advisory for transparency?
>>>
>>> Jan
>>>
>>> 15. des. 2021 kl. 19:09 skrev Uwe Schindler <uwe@thetaphi.de>:
>>>
>>> Hi all, I prepared a PR about the followup CVE-2021-45046:
>>> https://github.com/apache/solr-site/pull/59
>>>
>>> Please verify and make suggestion. I will merge this into
>>> main/production later.
>>>
>>> Uwe
>>>
>>> -----
>>> Uwe Schindler
>>> Achterdiek 19, D-28357 Bremen
>>> https://www.thetaphi.de
>>> eMail: uwe@thetaphi.de
>>>
>>> *From:* Uwe Schindler <uwe@thetaphi.de>
>>> *Sent:* Wednesday, December 15, 2021 3:31 PM
>>> *To:* 'dev@lucene.apache.org' <dev@lucene.apache.org>
>>> *Subject:* RE: Log4j < 2.15.0 may still be vulnerable even if
>>> -Dlog4j2.formatMsgNoLookups=true is set
>>>
>>> We should add this to the webpage. Another one asked on the security
>>> mailing list.
>>>
>>> Uwe
>>>
>>> -----
>>> Uwe Schindler
>>> Achterdiek 19, D-28357 Bremen
>>> https://www.thetaphi.de
>>> eMail: uwe@thetaphi.de
>>>
>>> *From:* Gus Heck <gus.heck@gmail.com>
>>> *Sent:* Wednesday, December 15, 2021 12:39 AM
>>> *To:* dev <dev@lucene.apache.org>
>>> *Subject:* Re: Log4j < 2.15.0 may still be vulnerable even if
>>> -Dlog4j2.formatMsgNoLookups=true is set
>>>
>>> Perhaps we could tweak it to say that the system property fix is
>>> sufficient *for Solr* (i.e. not imply that it is a valid work around for
>>> all cases)
>>>
>>> On Tue, Dec 14, 2021 at 6:20 PM Uwe Schindler <uwe@thetaphi.de> wrote:
>>>
>>> The other attack vectors are also not possible with Solr:
>>>
>>> - Logger.printf("%s", userInput) is not used
>>> - custom message factory is not used
>>>
>>> Uwe
>>> Am 14. Dezember 2021 22:59:26 UTC schrieb Uwe Schindler <uwe@thetaphi.de
>>> >:
>>>
>>> It is still a valid mitigation.
>>>
>>> Mike Drobban I explained it. MDC is the other attack vector and that's
>>> not an issue with Solr.
>>>
>>> Please accept this, just because the documentation of log4j changes,
>>> there's no additional risk. We may update the mitigation to mention that in
>>> Solr's case the system property is fine.
>>>
>>> Uwe
>>> Am 14. Dezember 2021 22:52:29 UTC schrieb solr <fredrik@rodland.no>:
>>>
>>> Ok.
>>>
>>> But FTR - apache/log4j has discredited just setting the system property as a mitigation measure, so I still think the SOLR security-page should be changed to not list this as a valid mitigation:
>>>
>>> https://logging.apache.org/log4j/2.x/security.html
>>> "Older (discredited) mitigation measures
>>>
>>> This page previously mentioned other mitigation measures, but we discovered that these measures only limit exposure while leaving some attack vectors open.
>>>
>>> Other insufficient mitigation measures are: setting system property log4j2.formatMsgNoLookups or environment variable LOG4J_FORMAT_MSG_NO_LOOKUPS to true for releases >= 2.10, or modifying the logging configuration to disable message lookups with %m{nolookups}, %msg{nolookups} or %message{nolookups} for releases >= 2.7 and <= 2.14.1.
>>> “
>>>
>>> Regards,
>>>
>>>
>>> Fredrik
>>>
>>>
>>> --
>>> Fredrik Rødland Cell: +47 99 21 98 17
>>> Maisen Pedersens vei 1 Twitter: @fredrikr
>>> NO-1363 Høvik, NORWAY flickr: http://www.flickr.com/fmmr/
>>> http://rodland.no about.me http://about.me/fmr
>>>
>>> On 14 Dec 2021, at 23:44, Mike Drob <mdrob@mdrob.com> wrote:
>>>
>>> The MDC Patterns used by solr are for the collection, shard, replica, core and node names, and a potential trace id. All of those are restricted to alphanumeric, no special characters like $ or { needed for the injection. And trying to access a collection that didn’t exist Returns 404 without logging.
>>>
>>> Upgrading is always going to be more complete, but I think we’re still ok for now, at least until the next iteration of this attack surfaces.
>>>
>>>
>>>
>>> On Tue, Dec 14, 2021 at 3:37 PM solr <fredrik@rodland.no> wrote:
>>> Only setting -Dlog4j2.formatMsgNoLookups=true might not be enough to mitigate the log4j vulnerability.
>>>
>>> See https://github.com/kmindi/log4shell-vulnerable-app
>>> “So even with LOG4J_FORMAT_MSG_NO_LOOKUPS true version 2.14.1 of log4j is vulnerable when using ThreadContextMap in PatternLayout.”
>>>
>>> ThreadContext.put(key, value) is used under the hood by MDC. I’m not sure wether any user-input is actually stored in MDC in SOLR.
>>>
>>>
>>> Probably this should be updated: https://solr.apache.org/security.html#apache-solr-affected-by-apache-log4j-cve-2021-44228
>>>
>>> And maybe consider releasing patch releases for other versions than 8.11 as well which includes log4j 2.16.0?
>>>
>>>
>>>
>>> Regards,
>>>
>>>
>>> Fredrik
>>>
>>>
>>> --
>>> Fredrik Rødland Cell: +47 99 21 98 17
>>> Maisen Pedersens vei 1 Twitter: @fredrikr
>>> NO-1363 Høvik, NORWAY flickr: http://www.flickr.com/fmmr/
>>> http://rodland.no about.me http://about.me/fmr
>>>
>>> ------------------------------
>>>
>>> To unsubscribe, e-mail: dev-unsubscribe@lucene.apache.org
>>> For additional commands, e-mail: dev-help@lucene.apache.org
>>>
>>> ------------------------------
>>>
>>> To unsubscribe, e-mail: dev-unsubscribe@lucene.apache.org
>>> For additional commands, e-mail: dev-help@lucene.apache.org
>>>
>>> --
>>> Uwe Schindler
>>> Achterdiek 19, 28357 Bremen
>>> https://www.thetaphi.de
>>>
>>> --
>>> Uwe Schindler
>>> Achterdiek 19, 28357 Bremen
>>> https://www.thetaphi.de
>>>
>>>
>>>
>>> --
>>> http://www.needhamsoftware.com (work)
>>> http://www.the111shift.com (play)
>>>
>>>
>>>
>>
>> --
>> http://www.needhamsoftware.com (work)
>> http://www.the111shift.com (play)
>>
>
>
> --
> http://www.needhamsoftware.com (work)
> http://www.the111shift.com (play)
>