SPA 504 after upgrade to 7.4.4 Phone does not provision - even with the simplest config files

[FlyingDoctor_2]

Syslog message: LOCAL0.INFO: SPA504G 00:25:84:d8:c2:03 -- Successful resync http://secprov2.xxx.com/sipura/Configurator.aspx?type=504G&mac=002584d8c203&mode=update

contents of config file

<flat-profile>
<SIP_User_Agent_Name ua="na" >D5FRIG-$VERSION</SIP_User_Agent_Name>
</flat-profile>

17    12:52:43.271073    192.168.1.69    88.215.56.187    Syslog    LOCAL0.INFO: SPA504G 00:25:84:d8:c2:03 -- Resync failed: corrupt file\n

[Steven Howes]

Hi,

I have found another undocumented change relating to HTTP. The Useragent has changed, see below:

"Cisco/SPA508G-7.4.4 (002584D8XXXX)(CBT1323XXXX)"

"Linksys/SPA508G-7.4.3 (002584D8XXXX)(CBT1323XXXX)"

This series of undocumented changes is making something which should be a simple firmware upgrade into a fault finding task. Surely Cisco must have some sort of version tracking and release control so changes like this are documented? Something like a useragent change could quite easily mess something up if a provisioning system relies on useragent to determine phone model.

S

[icastrov]

Hi Steve

Can you please provide me the complete traces you've just pasted below?

Thanks!

-nacho

[Steven Howes]

Define 'trace'. I can give you the full line from the log, but doesn't contain anything more than the time and the IP of my server and the useragent.. If you want a packet trace of the entire firmware download I can provide it... Although it would be much quicker to get one phone on 7.4.3 and 7.4.4 and just do a download. It's simple enough to see.

S

[icastrov]

I'd need a packet or network trace.

Thanks!

-nacho

[Steven Howes]

Ok I'll get one tomorrow, although it makes little or no sense. Clearly what has happened is the HTTP useragent has been updated to 'Cisco' to try and lose the old 'Linksys' brand. The 'broken bit' is your updating of the changelogs, not the code. I'm not entirely sure that's been understood though.

S

[icastrov]

It's being perfectly understood, issue here is that this change has been around since 7.3.5 so I did just want to confirm that you're seeing this in 7.4.3. I'll give it a try tomorrow in my lab so don't worry if you had not the network traces handy.

Thanks!

-nacho

[Steven Howes]

I can get a dump of the HTTP request tomorrow morning (depends who's tomorrow is first ;) damned time zones, 13 or so hours for me ). Worth noting this is a 508 not a 504. I'll check the logs for the 504's at the same time tomorrow to see if they are consistent (I'd assume so given the firmware file looks to be identical). Could be the change was made in the older version but never actually worked until the HTTP stuff was touched in the last release (breaking chunked encoding handling at the same time)?.

S

[Roland_Pope]

I have the issue with a 502G and the workaround of adding the 'Content-Length' header seems to have fixed the problem for me too. The phone was on 7.4.3a and provisioning fine, but this issue showed up after the 7.4.4 upgrade. Odd that I have a SPA504G which seems to be going fine on another system.

Roland

[info]

"Cisco/SPA508G-7.4.4 (002584D8XXXX)(CBT1323XXXX)"

"Linksys/SPA508G-7.4.3  (002584D8XXXX)(CBT1323XXXX)"

is not the only thing that changed

before it was like this, so if your provision system extracts the serial number from it, it could run into problems when looking at first and last brackets only (as our system did)

someone at cisco thought it would be handy to add MAC address to the useragent string..

"Cisco/SPA508G-7.4.4 (CBT1323XXXX)"

"Linksys/SPA508G-7.4.3  (CBT1323XXXX)"