Windows Thread, Loss of internet in Technical; We are pulling our hair out on this one!
We use 2 proxies in the school - a filtered one, ...
11th June 2008, 11:12 AM #1
Loss of internet
We are pulling our hair out on this one!
We use 2 proxies in the school - a filtered one, and an unfiltered one, both of which are SEGFL (RM) supplied.
The filtered one is the one staff and students use, the unfiltered one is for us techies, naturally.
Randomly, without any warning, the Internet on the filtered proxy is lost. This lasts only a few minutes at most, and doesn't always seem to affect all machines equally. That is, you will find that some machines can still access certain sites, whereas others can't, but then there are instances where all machines are equally affected and no-one can access the internet at all. The latter is the most common issue however, so all machines are affected.
This happens across both domains we have here, our curriculum one and our admin one, so it doesn't matter which domain the user is logged into, the exact same problem occurs.
We use Opera on our admin machines as our filtered browser, so we can replicate the same access as the filtered users get; Firefox is set as our unfiltered browser. WE only use IE7 if we have to...
All access to the internet is via the same router for all machines, so we don't think it's the router as when the filtered proxy goes down, the unfiltered proxy still works through the same machine, i.e. we can still get internet access using Firefox, but not Opera.
Convinced therefore we are that the problem lies with SEGFL.
We are being asked by them to provide SQUID errors, but of course, without any access to the internet proxy, we aren't going to be able to get these errors.
Does anyone know of a FREE network diagnostic tool we could run that would generate some sort of clue as to what is going on, and that doesn't utilise IE's connection settings?
11th June 2008, 11:34 AM #2
So your filters aren't physically at your school, they're at your grid HQ? Is there any sort of user authentication going on?
11th June 2008, 11:41 AM #3
I'm using both the filtered and unfiltered segfl proxies here, and apart from when the grid dies completely I've had no issues with them. What's the router? It may not be able to handle the connections, although it should be a 2611xm or something like that which should be fine.
How do you connect to the grid? We are connected via LES to our LEA so sometimes it can be issues there instead.
11th June 2008, 01:24 PM #4
No authentication going on. We pretty well know the issue is beyond our network, we just have to find a way to show it is.
Originally Posted by bizzel
11th June 2008, 01:30 PM #5
RM has a bank of servers that share the load.
Ours routinely drop out and come back online after a while. If its regular, then when it goes wrong again type in a random url that you know does not exist and on the cache error page note which sever is at the bottom, it will read like this
Generated Wed, 11 Jun 2008 11:27:04 GMT by proxy-02.dorset.ifl.net (squid/2.6.STABLE12)
Phone up RM and tell them that particular server is giving you grief.
11th June 2008, 01:42 PM #6
That's just it, we don't get squid errors, just a browser message that the proxy is unreachable.
Originally Posted by ICT_GUY
11th June 2008, 01:43 PM #7
If you can access some websites and not others, this does indicate an external DNS issue.
11th June 2008, 01:49 PM #8
I'm being told by the guvnor that it's not DNS.
Originally Posted by Michael
11th June 2008, 01:56 PM #9
We're on SEGfL and the proxy service is such utter sh*t that we put in a smoothwall box and send everyone out through that.
11th June 2008, 01:56 PM #10
Well if the proxies aren't located at your site and the filtering is load balanced, it could still be a combination of DNS or a load issue their end.
Can you ping the proxy when internet connectivity is lost? And under proxy settings (within IE), enter the IP of the proxy instead. If the problem still happens it will be a load issue for certain.
We had similar problems here in Birmingham (not with RM) and it was a load issue. There was a pattern that at certain times (especially lunchtime) connectivity would go down then come back online for no reason.
11th June 2008, 02:04 PM #11
What sort of cost was involved in implementing the smoothwall box?
Originally Posted by CyberNerd
11th June 2008, 02:11 PM #12
about £1400 per year IIRC + the cost of a box.
Originally Posted by Tx2online
Thanks to CyberNerd from:
11th June 2008, 02:13 PM #13
Will try this next time it goes down. Thanks.
Originally Posted by Michael
12th June 2008, 12:25 PM #14
Similar problem here with EMBC during their recent transition - where (I think) they had a pair of outwardly facing routers - one of which had been updated to the latest firmware and the other hadn't - so depending which one your session got allocated to and depending on overall loading you got to the internet...or not...
By dgsmith in forum Windows
Last Post: 14th March 2008, 09:24 PM
By badboyvtec in forum Wireless Networks
Last Post: 20th February 2008, 02:02 PM
By GrumbleDook in forum General Chat
Last Post: 8th February 2008, 02:44 PM
By DaveP in forum MIS Systems
Last Post: 3rd May 2007, 05:23 PM
By steelrazor in forum Wireless Networks
Last Post: 6th July 2006, 02:36 PM
Users Browsing this Thread
There are currently 1 users browsing this thread. (0 members and 1 guests)