Re: Generic records and DKIM
Matthew Dempsky <[email protected]>
| Newsgroups | gmane.network.djbdns |
|---|---|
| Message-ID | <[email protected]> |
On Wed, Mar 18, 2009 at 1:29 PM, Matthew Dempsky <[email protected]> wrote: > The problem I was referring to isn't a tinydns limitation, it's a DNS > limitation. A TXT record's data is split into chunks at most 255 > bytes long. A base64 encoded 1536 or 2048 bit key will be larger than > 255 bytes, so it'll necessarily span multiple chunks. If clients > don't support concatenating chunks, then they can't support these > larger keys. I see this example in RFC 4871: brisbane IN TXT ("v=DKIM1; p=MIGfMA0GCSqGSIb3DQEBAQUAA4GNADCBiQ" "KBgQDwIRP/UC3SBsEmGqZ9ZJW3/DkMoGeLnQg1fWn7/zYt" "IxN2SnFCjxOCKG9v3b4jYfcTNh5ijSsq631uBItLa7od+v" "/RtdC2UzJ1lWT947qR+Rcac2gbto/NMqJ0fzfVjH4OuKhi" "tdY9tf6mcwGjaNBcWToIMmPSPDdQPNUYckcQ2QIDAQAB") I'm not positive, but I understand this to be a TXT record split across 5 chunks. dig(1)'s output is consistent with this, but I don't have a BIND server to test this against to confirm. It seems like judging from that, there shouldn't be any problems with tinydns-data splitting your TXT records into 127 byte chunks. If it's still not working with the s/127/255/ change, I suspect your problem is somewhere else (e.g., publishing the wrong public key, formatting the DKIM record wrong, etc.).