Algorithm roll-over, unexpected content in dsset-file
Mark Elkins
mje at posix.co.za
Mon Aug 12 17:44:27 UTC 2019
Hi Ondřej,
Thanks for the info. I was happy to run an Algorithm roll-over the way I
did previously (years ago) and was unaware of the combined signing key
issue.
Regardless, decided to create a new Algo-13 KSK when ever I create the
first Algo-13 ZSK for any particular domain - which removes the problem.
Algorithm roll-overs are way more fun than a normal KSK roll-over :-)
Was trying to keep all my KSK roll-overs as spread out around the year
as possible. I suspect what I want to do is keep ZSK's using the old
algo until its time to create a new KSK (with Algo-13) then immediately
roll-over a ZSK (created with Algo-13)... so the KSK's timetable does
not change.
Might do a presentation on this one day.
On 2019/08/11 19:32, Ondřej Surý wrote:
> When there’s no DNSSEC KEY with SEP bit set, the Algo-13 “chain” has
> single entry point, e.g. Single-Type Signing Scheme (Combined Signing
> Key), see https://tools.ietf.org/html/rfc6781#section-3.1 and
> https://tools.ietf.org/html/rfc8499#section-10
>
> I don’t know exactly know what are you trying achieve, but adding only
> ZSK with new algorithm serves no purpose.
>
> Ondřej
> --
> Ondřej Surý — ISC
>
>> On 11 Aug 2019, at 12:59, Mark Elkins <mje at posix.co.za> wrote:
>>
>> Hi, Running BIND 9.14.4 on Gentoo.
>>
>> I've been running BIND and DNSSEC for a long time. Years ago - I
>> changed from Algorithm 5 to 8 and am now changing from 8 to 13.
>>
>> My ZSK's have a lifetime of 34 days and my KSK a lifetime of 370
>> days. I've chosen to create a new ZSK every 17 days and KSK every 185
>> days - so usually have two of each per domain. I've 1532 domains in
>> my system, 74 are NSEC and 54 are NSEC3 - the rest are unsigned.
>> Currently - resigning happens randomly over the year - which keeps
>> loads (and changes) scattered over the year. Some parents - I manage
>> via EPP, others have various other methods of updating - so having
>> the best part of six month to do a key-rollover has its advantages.
>>
>> I've just started allowing the system to add Algo-13 Keys... and this
>> is what I see:
>>
>> 1 - Created a test zone file: contains enough to be legal.
>>
>> 2 - Created some old keys, one ZSK and one KSK - Signed the domain
>> and looked inside the "dsset-" file....
>> # vi db.foo.com
>> # dnssec-keygen -a RSASHA256 foo.com
>> # dnssec-keygen -a RSASHA256 -f KSK foo.com
>> # cat Kfoo.com.*.key >> db.foo.com
>> # dnssec-signzone -a -o foo.com db.foo.com
>> # cat dsset-foo.com.
>> foo.com. IN DS 62014 8 1 6CFC9114C7D0954195C01B488C3B8D6E1BD0F220
>> foo.com. IN DS 62014 8 2
>> 74112FE6AE6E9DFB5A00EB3C26397E92CAA27038FA176A9B4B67D9FE 4E0154FE
>> (perfect!)
>>
>> 3 - created one additional ZSK - Algo 13, add that to the zone and
>> resigned.
>>
>> # dnssec-keygen -a ECDSAP256SHA256 foo.com
>> # cat Kfoo.com.+013+34805.key >> db.foo.com
>>
>> # dnssec-signzone -a -o foo.com db.foo.com
>> Verifying the zone using the following algorithms: RSASHA256.
>> Missing self-signed KSK for algorithm ECDSAP256SHA256
>> The zone is not fully signed for the following algorithms:
>> ECDSAP256SHA256.
>> DNSSEC completeness test failed.
>> Zone verification failed (failure)
>>
>> I don't mind the failure - after all there is no complete Algo-13
>> chain - but the existing Algo-8 chain is complete. Interestingly, the
>> signed zone contains no mention of the Algo-13 ZSK. No problem.
>>
>> However - WTF is a ZSK doing in the dsset file??? Is this a bug?
>>
>> # cat dsset-foo.com.
>> foo.com. IN DS 34805 13 1 A040C7BC6C03778B20AD4A3AEE3BC0648CE8AC23
>> foo.com. IN DS 34805 13 2
>> 340DC924A523CAD49E5C22357BBFECC07A2944D7723AD8F44DCE5332 FACD3AB2
>> foo.com. IN DS 62014 8 1 6CFC9114C7D0954195C01B488C3B8D6E1BD0F220
>> foo.com. IN DS 62014 8 2
>> 74112FE6AE6E9DFB5A00EB3C26397E92CAA27038FA176A9B4B67D9FE 4E0154FE
>>
>> 4 - If I then add an algo-13 KSK, add it to the Zone and run signzone
>> - everything looks good again...
>>
>> (New KSK) foo.com. IN DNSKEY 257 3 13
>> pD6yMHdPCKmLj0hY8oiLA1F2Vdb8a1LgpzU8XiO2c6eIh16n1Dv2Vogc
>> 9WylTUjOLTZlhRXPDmAcIUTyBUuC2Q==
>> # dnssec-signzone -a -o foo.com db.foo.com
>> Verifying the zone using the following algorithms: RSASHA256
>> ECDSAP256SHA256.
>> Zone fully signed:
>> Algorithm: RSASHA256: KSKs: 1 active, 0 stand-by, 0 revoked
>> ZSKs: 1 active, 0 stand-by, 0 revoked
>> Algorithm: ECDSAP256SHA256: KSKs: 1 active, 0 stand-by, 0 revoked
>> ZSKs: 1 active, 0 stand-by, 0 revoked
>> db.foo.com.signed
>> # cat dsset-foo.com.
>> foo.com. IN DS 23294 13 1 AADE5856B8EB3BB946DAAD2804A735816B3B3132
>> foo.com. IN DS 23294 13 2
>> 7FD5F02A2E52A1C6E3DEBD54C05A9A974CE88FA78548E3644B36FA2D 27E0F910
>> foo.com. IN DS 62014 8 1 6CFC9114C7D0954195C01B488C3B8D6E1BD0F220
>> foo.com. IN DS 62014 8 2
>> 74112FE6AE6E9DFB5A00EB3C26397E92CAA27038FA176A9B4B67D9FE 4E0154FE
>>
>> I really don't like the bad ZSK DS record in dsset because I
>> currently use that info to build CDS records in to my Database -
>> which is used to verify when the DS record is seen in the parent.
>> Anyway - I think this is a bug in the code of dnssec-signzone.
>>
>> I'm trying to wait as long as possible before forcibly creating
>> Algo-13 KSK's in the hope that they will be created according to the
>> existing schedule. If the Second Algo-13 ZSK is to be created in
>> under 3 days - then I'll replace the oldest KSK with an Algo-13 KSK.
>>
>> --
>> Mark James ELKINS - Posix Systems - (South) Africa
>> mje at posix.co.za Tel: +27.128070590 Cell: +27.826010496
>> For fast, reliable, low cost Internet in ZA: https://ftth.posix.co.za
>>
>> _______________________________________________
>> Please visit https://lists.isc.org/mailman/listinfo/bind-users to
>> unsubscribe from this list
>>
>> bind-users mailing list
>> bind-users at lists.isc.org
>> https://lists.isc.org/mailman/listinfo/bind-users
--
Mark James ELKINS - Posix Systems - (South) Africa
mje at posix.co.za Tel: +27.128070590 Cell: +27.826010496
For fast, reliable, low cost Internet in ZA: https://ftth.posix.co.za
More information about the bind-users
mailing list