Had a quick search for possible similar questions and not found anything that really matches, we have a service where by home users using their own pcs/laptops can take up to 3 attempts before getting in ok. So these pcs can have anything on them, kids games, itunes (aware of the bonjour issue) etc. Also users are not restricted on what browser or OS they use, so some are using ie8 and windows 7 which i know is not supported on 6.3r3. But we are being told this happens for all uses across the board regardless of whether they are using supported or non supported set ups. Users get the log on page, enter their credentials and hit enter, at our end we see nothing at all on the juniper or authentication server, not even a failed attempt. we can ask users to clear their browser history, cache and cookies and 9 times out of 10 they get in ok, so its as if the browser is bringing up a cached copy of the log on page, but why when we set the browser to search for the page on line every time, does its still behave like this, any suggestions would be greatfully recieved.
I don't recall hearing about this issue in 6.3r3. I would suggest testing out 7.0 in your lab though. 7.0 is now the reccomended release for IVE OS. 6.3 is no longer supported by engineering (it has reached EoE (end of engineering).
IVE software download: http://www.juniper.net/techpubs/software/ive/#sw
IVE OS software end of life chart: http://www.juniper.net/support/eol/ive.html
we are in the process of getting the sign off on the upgrade but the client still wants an answer as to why this is happening, some users state that it started last november, we have changed nothing in the setup up our end , so we have been wondering if some kind of microsoft update in the browser or os may have caused it.
It could be due to an update to the OS. What I would try is testing this with a vanilla install of XP (I use vmware for this) in order to see if this works before updating the OS. 6.3 was tested by our QA department with only the OSes/service packs that are listed in the 6.3 supported platforms document.
If you'd like more analysis you could also open a JTAC case. JTAC can look into the issue further, but engineering would not be able to create a fix for the unsupported code. If the issue was seen on a supported release of code we would file a problem report (PR) for the issue for a fix.
thanks for the suggestion, totally appreciatte its end of life, but restricted by the clients wishes regarding the upgrade, we are doing reboots of the devices every 48 hours due to the known memory leak issue on this version of software, just wondering what know side affects the memory leak issue has, i know it restarts the web services if we dont reboot on a regular basis, but could this log on issue have anything to do with the memory leak? or has anyone else come acros a similar issue nd found a fix, upgrade could be another 2 months away.