Monday, July 12, 2010

Linux frustrations!!!

What a weekend, I decided that I should get my IPv6 network up and properly running (I was missing the proper firewall rules so I couldn't leave the IPv6 up). After a great deal of frustration with ip6tables I finally managed to get a mostly workable firewall (icmpv6 packets are dropped). It really needs some more work and I'll work on that a little later on. I'll post all my related OpenWRT/IPv6 information on my Home IPv6 page. At the moment I've got the related information (IPv6, firewall and RADVD information) all over the place and I need to 'write' it all down. While I was at it I decided to consolidate my servers (actually the whole point of all these servers). I've been having a little bit of trouble with my main server (write errors). To make matters worse the 'new' server was behaving odd. At first the machine was geting stalls and I thought it might be Centos (I hate SELINUX and I was getting audit errors). I then tried Ubuntu but ran into the same stalls. At the same time I installed Ubuntu server on a laptop with no LCD screen to handle some server work. Well when the laptop rebooted the network wouldn't come up. I found that it was missing the auto in the interfaces file. I found out about it accidentally while researching something else. Now understand that I come from a SYSV, Red Hat background and I'm no way familiar with where everything is under Ubuntu. This makes it very frustrating to manage a server without some GUI (and I dislike GUIs). So I now have 6 computers (1 a desktop) in various states of disarray. But I'm just beginning to make a little bit of progress and hope to have at least one fully working server up and running tonight. After which I'll put all my necessary services in my OpenWRT gateway and the network server. Then I can turn off all the extra computers.

To add to my frustration I spent the weekend also working on my fall graduate class schedule (and necessary prep work) for NJIT. This might not have been so difficult except that I've had problems with my email and the various files were on the stalling computer. I've managed to get the information backed up to a NSLU and I was able to get the information from the files. Still the work that goes into a new school is difficult and I still need to do a few more things.


At 7/12/2010 10:16 PM, Blogger Jon Smirl said...

Check out those are 802.15.4 (Zigbee) devices, but they run IPv6 instead of Zigbee. I have five of them running an IPv6 mesh (6lowpan/RPL) on my desk right now. You can get cheap devices here: You could incorporate one into your IPv6 router.

At 7/12/2010 10:43 PM, Blogger Neil Cherry said...

Oooh! Neato!

In all seriousness those sound very cool. I'll try to see if I can get my hands on some of those. I've been interested in 6lowpan. Having something working right in front of me would be very nice.

At 7/12/2010 10:50 PM, Blogger Neil Cherry said...

As a side note, I've got Ubunto 10.04 LTS on both of my new servers. I tried to get Centos loaded on the second but it fail to load properly on my USB flash stick. I guess I'll be learning the Ubuntu way (as opposed to the Red Hat way). I don't have that much time to fight with installations.

And the stalls were caused by my 'nice' WDC 1.5TB drive. Seems there's something wrong in the first 500G of the drive. I'm running tests on it now to see if I can isolate the problem. The last 1TB seems fine but I'm a little nervous about trusting such a drive.

Once I get a stable server I'll move Misterhouse over to it and then I can clean up the older system (Fedora FC6) it has a drive with a couple of troublesome sections.

At 7/13/2010 8:58 AM, Blogger Jon Smirl said...

The econotag is the best one to start with. Multiple ones will work in the same box. You can also flash them and use phone chargers to spread them around the house to test meshing.

The econotag has integrated JTAG. You don't need a JTAG pod, the JTAG pod is built into the stick. Just plug the stick in and OpenOCD will work on it. That will let you single step your code, set breakpoints, etc.

Check out We have been working on light dimmers using the mc13224 but they aren't ready yet.

At 7/13/2010 9:29 AM, Blogger Neil Cherry said...

Jon, excellent information! Thanks :-)

At 7/13/2010 11:45 AM, Blogger harald said...

Don't know if it's affecting you, but I've tracked down issues with a pair of WDC 1.5TB drives to power. The drives were throwing many many transient read errors and bus resets until I upgraded the power supply in my old server - now there are no errors being reported, even if I hammer on the drives.

At 7/13/2010 12:18 PM, Blogger Neil Cherry said...

Harald, thanks for the warning. I'll check the power. At the moment I've localized the problem to the first 512G and I've run the bad block check on the drive. I've also let SMART at it to see if it notes anything. I'd hate to lose such a large drive but we'll see.


Post a Comment

<< Home