Tuesday, September 25, 2012

Westell 2200 Modem bridge mode problem - DSLextreme

After hours of trying every possible combination on my Westell 2200 modem from DSLextreme to configure for transparent bridge mode, the solution is to make sure you disable the "Private LAN" DHCP server under Configuration, Private LAN Configuration, Private LAN DHCP server, UNCHECK. 

The magic combination is under VC Configuration, VC1 (VPI 0, VCI 35) set protocol to Bridge, and mode to Bridge under Bridge Settings, then disable Private LAN DHCP. This is specific to DSLextreme and will not work with providers that use PPPOE at the modem or router. 

Most modems will disable any DHCP server when configured for Bridged mode, but the Westell 2200 does not, requiring you to manual disable it.  The more annoying problem is that DSLextreme tier 1 support will tell you "We ship all of our modems in bridge mode", which is technically true, but this modem is in "Routed Bridge" mode which does not allow your router to get the public IP and handle NAT for you. After a couple of chat sessions with tier 1 support I had to call and get it escalated to tier 2 to finally talk with a tech that understood the problem.

Monday, October 4, 2010

Cisco VPN Client and Network connection issues

If you are having network connection issues and you have Cisco VPN client installed, try stopping the Cisco Systems, Inc. VPN Services. This service seems to conflict with certain LAN protocols.

Specifically, the ongoing issue I have experienced is with EMC Retrospect backup client and the Cisco VPN service. The workaround is to simply stop the Cisco VPN service, allow the backup to complete, then restart Cisco VPN. I believe it has to do with the Piton network protocol that Retrospect client uses because I don't have any other issues except for Retrospect backups.
Here are the specifics:
  • OS Windows XP Pro SP3
  • Retrospect Client version 7.0.107 (more recent versions have the same issue)
  • Cisco VPN Client version (more recent version have the same issue)

Wednesday, September 8, 2010

ESET NOD32 and blocked HTTP requests

Users complaining they do not have internet access and cannot POP their email from server. Turns out network connections are fine (can ping, telnet, ftp etc) but all HTTP and SMTP requests were blocked. First thought, there is a nasty virus or malware lurking around the network. But no... turns out NOD32 (our virus and spyware protection software) was the culprit itself. In order to provide it's "Web access protection" and "Email client protection" it creates a local proxy to filter all requests. Well, somehow (yet to be determined) the proxy was hung or crashed in some way that blocked all Http and SMTP requests. For most machines, this was fixed with a simple restart of the computer, others, NOD32 had to be reinstalled.
While I dont yet have a firm solution to the problem or a determination as to what triggered it, I do know that at the time of the NOD32 issues, our central NOD32 server was very slow and nearly non-responsive. Why that would cause clients to hand, idk, but at this point, it is the only indication of a possible trigger.

I will follow up with any new solutions. Please post here if you have more info. Thanks.

We are currently running ESET NOD32 on Windows XP sp3 clients.

Thursday, January 22, 2009

Audio CD playback problem on Sony DVD RW drives

My Dell Optiplex 620 has two oem installed Sony DVD RW drives. Everything works fine with the exception of playing audio CDs. One of the drives plays music very choppy so I set out to fix the issue. Here is the setup:
IDE 0 is a SONY DVD RW DRU-180A (plays OK... "good drive")
IDE 1 is a SONY DVD RW DRU-835A (problem drive)

These are the only IDE devices in the machine. The hard drives are SATA. In the original factory configuration, both drives were set to cable select and only the IDE 0 drive would play audio properly, but IDE 1 would be choppy. So I tried a number of configurations as follows:
  1. I tried setting the jumper to force IDE 0 as master and IDE 1 as secondary... no change.
  2. I tried swapping the jumpers to reverse the order (DRU-835A now IDE 0 master) and now the "good drive, bad drive" problem is reversed... the DRU-835A plays ok.
  3. Only connect one drive at a time with jumper on cable select and at the end cable connector. In this config, either drive would play properly (as IDE 0).
So, I sumized that these drives do not like to be secodary drives, at least not for audio playback purposes. I have done disk copies and many burns without issue on either drive.

  1. I checked firmware versions on both drives and the DRU-180A is on the most recent firmware, and the DRU-835A is assumed to be on the latest firmware as Sony has not post a firware update on their site for this drive.
  2. Audio playback was tried with both Windows Media 11 and Nero Media Player
I don't have a solution here except to simply always playback on the IDE 0 drive.

Helpful comments always welcome :)

Monday, April 7, 2008

Remote Desktop Connection has Expired (Pre-Release Version)

Today, many Remote Desktop users were struck with the following message:

"The Pre-Release Version of Remote Desktop Connection has expired. To download the full version of Remote Desktop Connection, Go to Windows update or contact your system administrator"

This is a result of installing the recently released XP SP3 RC2 from Microsoft. Apparently they packaged the new Remote Desktop Connection in SP3 with an expiration date (assuming 4/7/2008). Thanks so much for the warning Microsoft! This came on as a big Monday morning surprise to many IT pros as well as home server users (WHS).

The fix?
I have seen many posts where people are uninstalling SP3, which works however rather extreme.

As a temporary workaround, you can simply copy the previous version of the following two files into your windows System32 folder. You can copy these files from another pre-SP3 XP machine or simply search your local hard drive and use a previous version which will be in a previous $NtServicePackUninstall$ folder or in the DLLCACHE folder:

Then execute the mstsc.exe (which is the same exe name as the expired version). This effectively rolls back your Remote Desktop version to the previous version.

Wednesday, November 7, 2007

Firefox - Have they Tamed the Animal?

I have written a couple of times about Firefox 2.0+ issues mostly regarding memory usage and memory leaks. Recently Mozilla released which is intended to primarily be a stability (not security this time) release. Well, I am pleased to say that after about a week of use, the memory leak problem seems to have been resolved.
Thank You Mozilla!

In addition, I have determined that one of the extentions I was using was a big memory hog. The extension was ANSWERS 2.2.27
It would use nearly 100mb of ram upon Firefox startup. I expect that they were using some sort of memory resident database of common words. Well, it was not worth the 100mb to me, so goodby Answers 2.2.27.

Now without Answers and with the resolved Firefox memory leaks, I am happy to report that Firefox is back to the "lean mean browsing machine" that I once loved.

Monday, September 17, 2007

Microsoft Forces Stealth Updates Even without Automatic Updates On

Here is another one from the "Microsoft is smarter than you are" file.

Microsoft was recently exposed for pushing out updates to your computer even if you don't have Automatic Updates turned on. Their position is that they need to continue to update the Windows Update Service components in order to notify you of updates and "maintain the quality of the service".

This presents a huge trust issue for many, especially IT support personnel, myself included. The reasons that Microsoft has announced on this issue are somewhat valid, but still cause a trust issue. If they are doing this unilaterally with some components, what keeps them from updating other components that they choose to update at will? The fact that Microsoft tried to do this behind the scenes is quite alarming.

Partnering with BIG BROTHER

There are many that are totally outraged by this Microsoft trust issue. My take on it is that you knew what you were getting into when purchasing and choosing to run Microsoft products, so deal with it. If you don't trust Microsoft to write your OS and write patches for your OS, then go play with someone else's OS. Granted, there are not many choices that most corporate average desktop users can deal with and with that, many IT support personnel are stuck with Microsoft desktops. If Microsoft screws this up and some compromised code gets pushed out to your desktops, yes, it will hurt and screw you up for a few days, but think about the pain you would endure to attempt to convert those desktop users to Ubuntu, or any other Unix like environment. Ultimately, Microsoft would be at fault for leaving the back door open.

For more detail on the MS stealth updates:



Wednesday, August 15, 2007

Does Qwest still do Cold-Potato Routing?

This is somewhat of a rhetorical question, but I do question it. Qwest has always (AFAIK) advertised and bragged about this feature with their DIA services.

First, some may be wondering, what the heck is Cold-Potato Routing? The best explanation may be that it is the opposite of Hot-Potato routing in which an ISP will hand off it packets to a peer or downstream ISP as soon as possible to get it off of their system. So, in Cold-Potato routing, the well intentioned ISP will hold onto the traffic and haul it across a geographic region as far as possible before handing it over to the downstream provider. This provides a best end-to-end quality of service for the ISP customers, assuming the ISP has the backbone to handle it.

The reason I question weather Qwest is still providing Cold-potato routing is that if I trace routes from my Los Angeles based POP (Qwest provided) to my co-lo in Chicago, IL the traffic gets handed off in Los Angeles to Level 3. net. Level3 then hauls it across the country to it's Chicago facility. Now, I know that Qwest has a big OC-192 backbone direct from their Burbank, CA TeraPOP direct to Chicago, IL TeraPOP. So, what the heck? Can anyone shine some light on this issue?

Firefox memory leak revisited

I recently wrote about Firefox 2 becoming a pig, meaning that it consumes huge amounts of memory and slows down. This of course is a memory leak as expected. During the course of research of the problem I have discovered a couple of things that are short term work arounds.
  1. Restart Firefox every day. It consumes memory over time (NSS!) so by restarting it, the memory will be released
  2. uninstall any add-ons (extensions) that you don't use.
  3. Uninstall extra themes
  4. Run Firefox in Safe Mode
  5. Keep unused browser tabs closed
I found this list of known problematic extensions that you should check out to make sure you are not running them:

There are claims that the extensions (especially the problematic ones) are causing the memory leaks. However, I only run about 4 lightweight extensions that I keep updated and only the default Firefox theme and I still have the memory leak problem. In fact, today Firefox on my desktop is comsuming over 163mb after restarting it about 18 hours ago. Guess it's about time for another restart.

Come On Mozilla Foundation! Get this fixed or I fear that many will resort back to IE. Oh the guilt...

Thursday, August 9, 2007

Check Your Network Switch Speed Negotiation

In dealing with a recent bandwidth issue, we discovered that it was a speed negotiation issue between a Cisco 3845 router and a Cisco PIX Firewall. You would think that these two interfaces would do a better job of speed negotiation. There were no errors recorded on either side but we experienced horrible throughput... like 500kpbs. The issue was resolved by setting both sides to AUTO and reset the interfaces.

I have had many issues in the past with Cisco interfaces to other manufacturer interfaces (Netgear and Linksys to name a couple). They don't do auto negotiate well, so I have learned to fix the Cisco interface to the max speed of the other side. But, in the case of Cisco to Cisco, it seems that AUTO is the best choice... at least in this case.

The moral to the story is that when troubleshooting bandwidth issues, alway check your speed setting settings on interfaces... even when no errors are recorded.

Wednesday, July 25, 2007

Rotate Your Spare UPS Batteries

If you keep spare UPS batteries on the shelf, make sure you rotate them into your running UPS unit every month or so to keep them fresh and charged.
I am mostly speaking of data center level user replaceable UPS batteries like a APC Symmetra or similar. Keeping spare batteries is a good idea for replacement of a failed battery and for extra runtime when the lights do go out. In the case of extra runtime needed, you can keep extra batteries on the shelf and swap them in if your runtime gets low due to a power outage and you don't have a generator on standby. Just make sure you rotate them into the unit every month or so. The approximate shelf life of a Symmetra battery is 6 months. After that, they are considered dead or at least reduced efficiency.

Another UPS battery tip is to date stamp the battery when you buy it. This will help control the shelf life as well as overall life as the life expectancy of one of these is usually about 2 years.