August, 2008

SQL backup jobs failing

 I don’t work with SQL jobs too much, but here is an error I just got, but I was troubleshooting an issue with SQL backup jobs failing, here was the error in event log.

tcpsvcs (2640) The backup has been stopped because it was halted by the client or the connection with the client failed.
Though the SQL server agent looked fine, I stopped ans started it anyways and that fixed the issue.

Watchguard MUVPN tunnel dropping after three minutes

Here’s a tip to developers if you don’t want to be universally hated by me. If you want to create your VPN software so that it drop’s the tunnel after 3 minutes if it doesn’t receive keep alive packets on UDP ports 500 and 4500 thats fine. However please have the software create exceptions in widows firewall or alert me or document it somewhere on your site so that I don’t have to waste an entire morning trying to figure out why it’s not working. ARGH!

Getting non-supported printers to work with Citrix Print Manager

As a general rule I quite like the Citrix Print manager, overall it does it’s job quite well and I think it supports a diverse group of printers. What I don’t like is when some people buy the crappiest, cheapest printers imaginable and then get angry at me when they don’t work with Citrix, particularly when they have had this issue before and I sent them the approved printers list and this printer isn’t on it. So for those times when “Buy a better printer” isn’t an appropriate response, here is the best solution I have.

Set the printer up as a network printer (buy the absolute cheapest usb print server if you have to), and then create a startup script to map LPT1 to the network share.

net use lpt1: \\servername\printername

Create a new printer using the HP LJ 1100 driver and assign it to print to LPT1. Rename the printer to something the user will associate with to his printer.

It’s quick and dirty but it works.

Broken hub causing Windows 2003 SBS server to freeze

So I am called in to fix a problem with a server being offline. I get onsite to find that the server is frozen and the last message was a delayed write failure. Give it a reboot and it blue screens. Boot into safe mode, works fine, I can get into windows, only suspicious event log errors are the delayed write failures. Boot into Safe Mode with Networking causes it to freeze again. Boot into normal mode with network cable unplugged and everything comes up ok. Plug in the network cable and try pinging, pings are intermitant, some work, some not, response is in the 1000’s of milliseconds. Try from another machine, same issue. Server freezes again. It’s about this time I realize the server is likely just a red herring (a crappy designed red herring) and start troubleshooting the network issue. Sometimes pings work fine, sometimes slow, sometimes not at all. Reboot all the main switches, same issue. So I start asking questions around. Turns out all the issues started around the time that someone dropped the small d-link hub by there desk. Bingo. Turn that off problems cease. Server comes back on, no more freezing. Now this was an interesting problem but it illustrates a rant. What the hell is wrong with server 2003? All the XP machines were fine, none of them froze up. I am not sure if it’s a server 2003 issue or an SBS issue (I wouldn’t be shocked, since SBS sucks) but regardless this is just bad.

“The specified domain either does not exist or could not be contacted” on a DC

I got this error today when trying to log into the only DC at one of our client sites. Needless to say this is a bad sign when a DC can’t see it’s own domain. After a few checks I quickly found that none of the machines could log into the domain at all. Didn’t take me long to figure out that DNS must be hooped. To get it working I logged in locally on another server, opened up services.msc, connected to the DC by IP address and stopped and started the DNS server service, and just like that everything came back. Now here is the rant, DNS is so tightly integrated with AD that you can get these catch-22 scenario’s where you can’t log into a server because DNS isn’t working but you need to log in to fix DNS. While one can easily restart a service from another machine, any other in depth troubleshooting is difficult especially since in this situation it would have been a 45 minute drive to get to the physical box.

Citrix Print Manager service hanging

I have been seeing this issue alot lately. Citrix Print manager service stops working properly so you go to restart the service and the service hangs. You end up having to kill the process CPSVC.EXE, then restart the spooler service and then start the Citrix Print Manager service. It’s really obnoxious but at least it’s a relatively quick fix to get it working. If this happens alot it can be resolved permanently by Citrix hot fix PSE400W2K3022.

Troubleshooting wireless networks

Wireless networks are getting more and more common and with good reason, they are damn useful, particularly when you get a lot of guests or vendors who visit your office. However they can be an absolute pain in the butt to troubleshoot. Part of this is because in a 9 times of 10 the person installing it buys the cheapest linksys or d-link wireless router they can find and installs it with the standard settings so you end up have 15 networks in proximity all with the same name and all on the same channel. It also  seems that some people  don’t see the correlation between the 2.4 ghz router they bought and all those cordless phones in their office that say 2.4 ghz on them. Plus why don’t some people realize that security could be important. To quote one of my co-workers “God bless the people who don’t realize they should put some form of security on their wireless networks because otherwise I would have to pay for internet at home”, great for your neighbours, bad for your network.

So there is a bunch of stuff that can go wrong but how do you make it right? Well I like to use a little tool called Net Stumbler. A great tool which shows you all the networks in proximity, their names, which channel they are on, their signal strength, whether or not they are secure and what type encryption they are using. Take it from me, it is a god send when troubleshooting wireless network issues.

 http://www.softpedia.com/get/Network-Too…

KVM issue with Proliant DL380 server

This was an interesting issue I recent run into. We had just put in some new servers at a location and we needed to reboot them to get them to work with the KVM(Star tech). This had happend before with this KVM and the servers just needed to be rebooted with the KVM focus on them. I scheduled some downtime and went through the reboot of each one. Each one came back up fine except for one, a HP proliant DL380 server, the keyboard and mouse still didn’t work after the reboot.  Went through another reboot and everything seemed fine during the boot, keyboard was working fine (I was able to get into BIOS and I was able to cycle num lock) but as soon as it got to the Windows splash screen the keyboard stopped working (unable to cycle num lock).  I tried a different KVM port and different KVM cable and had the same issue

After a little searching I found this post on the HP forums

 http://forums11.itrc.hp.com/service/foru…

Apparently It has to do with the keyboard port switching operating modes during bootup, the Star Tech KVM just couldn’t handle it. I will probably recommend a switch to an Avocent IP KVM, I have worked with them in the past and never had any issues plus it’s pretty awesome to be able to log into the web interface and manage all the servers remotely from the KVM.

Access denied when trying to access MDB files

For the people who don’t think that integrating Internet Explorer into the Window’s OS was a bad idea. here is a problem I recently faced. This was a really weird issue. I was working with a client who was trying to access an .mdb file on the network and was getting an access denied error. The weird part was that she had NTFS access to the folder. So I started by trying the standard stuff, I made sure share permissions were set to Full Control for everyone, pushed down folder NTFS permissions to all the files, even made her the owner of the file, nothing worked.  Then I found that she could access any of the files in the folder alright except for MDB files. It took a bit of doing I got it working. I actually had to add the UNC path to the list of Internet Explorer trusted sites. This resolved the issue but it begs the question why this was an issue at all? Thanks Microsoft.

Mcafee issues with Goldmine

Ok seriously now, Mcafee is the suckiest bunch of suck that ever sucked. I don’t form this opinion over one instance but a career worth of experience. Everytime I start to think that maybe just maybe Mcafee is now ok they go ahead and prove me wrong. Take my experience with it today for example. I am working with a client who uses Goldmine for their email client. They are able to receive email but are unable to send it. They get the error Connection refused: System error 10061.  A quick look in the Application event log shows this:

Blocked by port blocking rule (Anti-virus Standard Protection:Prevent mass mailing worms from sending mail).

You actually have to disable the Mcafee worm protection to get this to work.  So now sure Goldmine works but if and when this client actually gets a real worm there is no protection against it!  What is the point of buying software if you have to disable it’s functionality to get it to work?