Archive for the ‘Tech’ Category


TiVo cablecard woes

17 August 2008 – 13:47 CDT

Ever since shortly after getting cable cards for my S3, it “loses” channels seemingly randomly. It was a pretty rare thing at first, once a month or so a channel would just be black. Not all channels, and usually not more than one. The problem has gotten progressively worse over the last 10 months or so – happening more and more often. I’ve noticed that changing the channel seems to cause the behavior. That is, switch to a channel and it will show the programming for about 1 – 1.5 seconds and then go black. This is happening every day or so now. A couple of days ago I was at home and the TiVo switched both tuners to record shows and both stopped working – just a black screen.

I’ve changed from 2 single-stream cards to 2 multi-stream cards as per TiVo support’s direction, but it didn’t help. Within the last week or so, I’ve had the box itself replaced, but that hasn’t helped either. I can’t get a straight answer from TiVo as to where the problem is. Some say it is an issue with the cablecards not being properly authorized, others say it is a software issue that is my fault because I have 9.x on the box. (Too bad it wasn’t my choice, and the replacement box has 8.0, with the same problems). Still another says it is a problem with Scientific Atlanta cablecards. I’m trying to figure out if Timewarner has Motorola cable cards, but I’m seriously doubting it. One suggestion from TiVo was to just power down and reboot when the channel stops working. That is annoying – there goes 10 minutes of the show I was trying to watch while waiting for the TiVo to reboot. This is also not useful if I’m not home — because this happens both when I change the channel manually and when the TiVo does so to record something.

While on the phone with TiVo the other night, one of the “missing” channels came back on its own. From black screen to clear picture with no prompting. Other times, it will go for the entire length of a show it should be recording and not actually record anything – because the TiVo doesn’t think it has anything to record.

Except – the “signal strength” meter for the channel looks normal – 95-97%. Really, really frustrated that such an expensive piece of equipment has such an incomprehensible problem.

Ubuntu thought the disk was a sound card?

28 June 2008 – 14:35 CDT

Had some problems with my Ubuntu server at home which I wasn’t able to resolve – something with the BIOS settings getting lost when the power went out rearranging the disks in such a way that the box wouldn’t boot. There was more to it than that, but after a few hours of diagnostics and getting nowhere I just decided to reinstall the OS. I wanted to upgrade from 6.x to 8.04 anyways.

Everything seemed to be fine, until I powered down the box last night to put it behind a UPS. I first noticed the problem when I logged in remotely and it couldn’t cd into my /home directory, which is on a separate physical disk than the OS. First off, I’m not entirely sure why, but Ubuntu is treating the IDE disk as a scsi device. I don’t think this is a problem per-se, but seems strange.

tinman kernel: scsi 2:0:0:0: Direct-Access     ATA      ST3320620A       3.AA PQ: 0 ANSI: 5
tinman kernel: sd 2:0:0:0: [sdb] 625142448 512-byte hardware sectors (320073 MB)
tinman kernel: sd 2:0:0:0: [sdb] Write Protect is off
tinman kernel: sd 2:0:0:0: [sdb] Write cache: enabled, read cache: enabled, doesn't
support DPO or FUA

The problem is the next line of the log

Jun 27 21:48:31 tinman kernel: [   39.495985]
sdb:<6>input: PC Speaker as /devices/platform/pcspkr/input/input4

So not only is the PATA drive a SCSI device, but now it is a sound device as well? Being detected as a sound device, I believe is what caused the filesystem to appear corrupt, but there were no complaints when the disk was mounted

tinman kernel: [   83.035617] EXT3 FS on sdb2, internal journal
tinman kernel: [   83.035621] EXT3-fs: mounted filesystem with ordered data mode.

The complaints started later trying to read the disk

tinman kernel: [  157.659996] attempt to access beyond end of device
tinman kernel: [  157.660003] sdb2: rw=32, want=37486616, limit=530082
tinman kernel: [  157.664594] attempt to access beyond end of device

Because it is sdb, and wasn’t able to set me up in /home anyways when I logged in, I was able to unmount the disk. I ran e2fsck which said it was clean, so I remounted the partition and everything went normally. It seems fine now, but it is a bit puzzling why the OS decided that the disk looked like a sound device and half-treated it as such? Starting to wonder if there isn’t something wrong hardware wise, either with the disk itself or with the controller.

Talking to your IT admins

21 June 2008 – 16:02 CDT

Looking for thoughts/ideas on how to talk to an IT admin. Started a job a few weeks ago where basically everything outbound except http and https are blocked. This means that ssh tunneling does not work. The traffic is packet inspected by the firewall and the http proxy requires authentication, so just moving ssh to port 443 doesn’t work either. The web traffic is filtered, so many things are blocked including gmail.

I’ve looked into solutions like corkscrew but it looks like it is going to take me a combination of ssh-over-https-proxies to get through it, because some of the tools only support Basic auth and the ISA server only accepts NTLM, Kerberos, and something else. It would be much easier to get to my box at home with its “library of files and tools” if they would just open up port 22.

I’m looking for anyone with ideas on how to talk to the IT admin staff about this. I’ve emailed them several times, and am not getting any response at all. I even included my MAC addresses and suggested they just unblock those. I’ve talked to my supervisor, and so far no luck – they mostly just don’t know what to do about it and the answers provided by the IT team range from the absurd to just dumb. Unfortunately, this is the same IT staff who:

– don’t know that terminal services is running on one of the windows 2003 servers I need access to (or insist that it isn’t running on the system at all)
– apparently when their company bought our company, dismantled the VPN because it was “insecure”
– set up a remote terminal services system exposed to the internet with the entire thing locked down (only one app is available, and the start menu is useless) as a solution for VPN/remote access.
– block gmail because it “has viruses”
– refuse to give the software developers, including those writing drivers, admin rights to their windows box

I don’t know who is responsible for these guys or who made up these “policies” but it seems like they just do whatever they want. My impression is that this team (who work out of the parent company’s office) is led by a guy who only cares that giving local admin rights to anyone would supposedly cause him to have to do more work to fix broken systems. Obviously that means that he is actively interfering with the business process of the org, but since no one seems to know what to do about it I’m throwing it out there to the three readers of this journal 🙂

How would you talk to your IT administrators about opening up port 22? Unblocking gmail? Putting the VPN back up? The only way the developers have admin rights on their own computers is the local VPs have domain admin rights to log in and let us reconfigure our own boxes – but this is not something we discuss with or even talk to the IT people about, which I don’t think is right, but we don’t seem to have much choice because they’re basically uncooperative. The local folks can’t modify the network or add new services like VPN though.

So how would you talk to windows sysadmins and convince them that they’re being unreasonable?

OS X Default keys

25 February 2008 – 12:53 CDT

Some default key bindings in OS X are, honestly, annoying. Especially the home/end keys. Of the “big three” desktop OSes, Mac seems to be the odd man out. Most windows/linux applications treat home/end the same way. In the OS X terminal application, the home key sends you to the top of the scroll buffer. The end key sends you to the end of the window’s scroll buffer. Not the expected behavior. Home means “home” – the beginning of the line. End means “end” – the end of the line. Pgup and pgdn are supposed to serve to move in large chunks around the buffer. I can’t count the number of times I’ve hit the end key expecting to get to the end of the line (like while I’m typing this post) and instead get the end of the buffer. Highly annoying to have my work interrupted by this constantly. Instead of me trying to learn how to use yet another interface/computer, I’m making the computer learn this time.

Here is how to remap the home/end keys in the terminal application

Terminal>Window Settings>Keyboard

end:

\033[4~

home:

\033[1~

the 033 part can be obtained by ^[ (aka ctrl+[ )

After doing that, you need to modify (or create) your .inputrc file to contain the following

# Be 8 bit clean.
set input-meta on
set output-meta on
set convert-meta off

# allow the use of the Home/End keys
"\e[1~": beginning-of-line
"\e[4~": end-of-line

# allow the use of the Delete/Insert keys
"\e[3~": delete-char
"\e[2~": quoted-insert

Restart the terminal and you’ll be good. Firefox, however still acts whacked out. That one is more complicated.