SIP phones no longer connected after 3.2.1 (Build: 5794EX) Saturday 1/21/18
RFC3261:
Quote from: LarryDP on January 25, 2018, 08:32:03 pm
Obi, what's the fix?
OBi staff (at least in the general case) does not read/respond on this forum (yes, there are exceptions, but do not count on it).
If your device is still under a warranty or a support contract (or you are willing to pay the $10 fee to obtain a new contract) open a case to get some sort of formal response.
Or you can try to approaches reported to be successful earlier in this thread (although, as with all else, your results may vary).
RFC3261:
FWIW (in case someone has an open ticket with OBi, and wants to send some addtional info), I took a look at the traffic on the wire (well, actually the logs).
It appears that the new firmware is responding to the first unauthenticated registration request with the expect 401 (not authenticated), asking for authentication with an interesting WWW-Authenticate string of the form:
WWW-Authenticate: Digest algorithm=MD5,nonce="37b7b4ad3BCEF40E",opaque="f66fc17e4fff9789",realm="pnn.obihai.com",nonce="5E7F538C",response="5E7F538C5E7F538C5E7F538C5E7F538C"
The old firmware replies with:
WWW-Authenticate: Digest algorithm=MD5,nonce="7fdddb54F4FF4579",opaque="5aafd8f577bcdc0d",realm="pnn.obihai.com"
Note that the new firmware is sending two nonces (which it really should not), along with a response value (which should not exist) which is just the second nonce value repeated 4 times. It would appear my SIP phone is replying using the second nonce value (which is arguably correct (use the last value, based on Postel's law))), but I am guessing the OBi is calculating the values with the first.
Anyway, I see a bug.
TheLliez:
Great detective work looking at the WWW-Authenticate.
BTW, how did you guys downgrade? I went back to 3.1.1 (Build: 5774) but that did not resurrect my phone.
Did you have to reset to factory the 202 before the update?
LarryDP:
Reverted to Build 5757 available at fw and problem solved. This doesn't say much for Obi QA and customer care that this has gone on for a week. Polycom needs to fix or I'll go elsewhere.
RFC3261:
Quote from: LarryDP on January 29, 2018, 08:54:14 am
This doesn't say much for Obi QA and customer care that this has gone on for a week.
Well, has anyone actually opened a support case with a supported (under contract) OBi? I have not (I don't have any OBi's under contract). You can't expect fixes if there are no open support tickets (and OBi seems to have been rather consistent recently in closing tickets with no further process if you do not have a valid support contract; even if they agree there is a bug (unknown), you don't prioritize such fixes if no one that is paying for support is asking for it).
As far as QA, from long experience, I know you that it is not atypical to tend to test in QA things that are normal/expected cases, and a few you think could be issues, and then tend to add things only later that you failed to test in the past and had caused issues with new releases. It is possible configuring an OBi with SIP phones was not in the test harness (it just worked). Perhaps they will add it to future QA testing.
And (to play devils advocate), if they do have a QA process, you can't expect a new release before the fixes have gone through QA. A week might be quick for their QA process (in the telco world an agile development methodology has not been typical).
Personally (again, since I have a fairly trival workaround) I am willing to wait for a few weeks with a possible new firmware update to see what happens before escalating.
Navigation
[0] Message Index
[#] Next page
[*] Previous page