NOTE TO THE SLEEPY: This post is about Debian Planet (founded 2000), not Planet Debian (founded 2004).

Sorry about the needless verbosity of the previous post. I’ll try keep this one shorter because I’d like some feedback on it. robster (who founded it) and I are essentially the only active staff of Debian Planet, and neither of us are able dedicate much time to it because we’re both kept busy with all manner of university-related stuff, both academic and not. It’s been suggested by some people within Debian that DWN and especially Planet Debian render it defunct and nobody reads it anyway. However, robster and I don’t really share this opinion – Planet Debian contains opinions, hacks, experiences, recipes and stuff that can be very interesting if you know the people involved, but really doesn’t necessarily perform the function of making Debian’s development more translucent to the outside world if you’re not on the mailing lists (although certain relevant blog posts can be highlighted, as DWN does). DWN seems to be aimed at a fairly technical audience who is generally aware of the projects’ internals but wants to keep up with day-to-day issues – it frequently includes stuff that’s not hugely relevant to the big picture unless you’re inside the project.

As far as we know there aren’t any other sites that are aimed at demystifying Debian for users, and with the advent of Ubuntu, depending on your viewpoint we stand to lose or gain (indirectly) many more users too, some of whom may take (or retain) an interest in Debian development (and we can cover Ubuntu too). The site does something in the region of 30 GB a month, a lot of this down to RSS feeds, so I’m sure articles that are posted actually do get read by quite a few people. Posts to the site itself asking if we should continue, and in what form, have prompted a surprising amount of positive feedback, and have been enough to dispel any ideas of closing down the site, although we’ve been quite lax at following up the resultant offers of assistance thus far (an ESR style list subscription of anyone who shows interest is exceedingly tempting).

What we’d like to do now is make it easier for other people to contribute (our editorial control has been an iron grasp thus far, with most articles being written or mostly rewritten by a very select few), and find some more people to help us out with keeping up with goings-on in the project, and we’ve also potentially got to find new hosting over the next few weeks. Are we misguided? Should we bother securing new hosting for the site and keep it going, or is it as needless as people say? Does anyone who’s reading this agree with our goals and want to help us out?

I’ve just been round at robster’s room helping him replace the AMD Sempron 3000+ which he had in his new Biostar 210V (similar to my 200V) and the BIOS apparently didn’t recognise. We put in an SB Live 7.1 because on-board VIA 82xx sound is crap (even if the planets are in alignment, DXS is in the right randomly numbered mode, and you manage to get ALSA and friends to do the appropriate resampling to 48kHz, it still sounds crap as a consequence – although do let me know if you know how to make it sound good), a 60GB HDD I had lying around while he’s waiting for CCL to replace or give him credit for his 2nd broken pair of HGST Deskstars, and swapped the Sempron for an AMD Athlon XP 2800+.

Unfortunately it was recognised by the BIOS as a 1250MHz processor or something… so I looked at the motherboard manual and found some jumpers to set the front-side bus speed. Firstly we were amazed it had jumpers (can’t BIOSes autoconfigure this stuff nowdays?), and secondly how the default setting was for a 100MHz FSB speed. The consequence was that the DDR400 RAM was running at DDR266 and the processor was running massively slowly. Setting the FSB to 166MHz made the processor and RAM run at the correct speeds, indicating that probably the Sempron would’ve worked anyway. Oh well, the Athlon XP is a better chip, and given the hassle with the actually broken hardware, CCL definitely won’t refund the Sempron if it’s not broken.

We set about installing Debian, really enjoying the partition manager in sarge d-i rc3, setting up LVM on degraded RAID arrays and such like, although it’s a shame that it falls back to crappy old lilo when installing on XFS root partitions. If sarge had a kernel where XFS wasn’t completely f🤬ed, unlike 2.6.8, then the xfs_freeze/unfreeze hack in grub-install would work. I just switched to the shell and installed 2.6.11 and grub manually before the reboot. I get the feeling joeyh really wouldn’t appreciate any kernel change right now though. Anyway, I digress…

The first boot was a disaster, the kernel was oopsing all over the place in all sorts of random locations, and usually ended up panicing some way into the boot. My assumption was that 2.6.11 was broken somehow, so we booted the install CD again to change it, and the kernel on that was hosed too. Uh oh! I was worried that I’d fried the processor with static whilst installing it, but it’s easier to eliminate the possibility that the RAM wasn’t broken. I removed one DIMM and it started working… but when I tried that DIMM on it’s own, the system worked too. I tried all permutations of the two DIMMs in the two slots, and discovered that when they were both in, the system screwed up, but when one or other was in either slot, the system worked fine. Apparently the CCL value RAM didn’t actually work when it actually got clocked at DDR400, so it needs to go back for replacement or credit too. Interestingly, memtest86 didn’t find any problem with any configuration within any sensible amount of time, and for me in the past it’s either found problems very quickly, or never found them at all.

A few lessons learnt (or re-learnt):

• Double check for for relevant jumpers on your motherboard if CPUs run at the wrong speed. Apparently some computers on sale now still have them!
• Don’t buy cheap RAM, it’s a false economy, life’s too short…
• If you’re planning to set up a RAID1 array in a new system, buy two different brands of disk or they’ll be from the same batch, experience the same wear, and fail in the same way at the same time. This happened to me with my system about a year ago, when both 3-month old HGST drives failed within 24 hours of each other (I now have a Seagate I bought at the time to recover my data onto, use one of the HGST replacements, and sold the other), and robster’s on his 2nd pair of broken HGST drives from the same batches…

Some people seem to be lucky with hardware and it works for them, or fails in obvious or non-critical ways whilst within warranty, and some people seem to be very unlucky and everything goes wrong – horribly, insidiously, and without any suitable recourse. In general I seem to be one of the former people, and robster, my brother and most of my friends are the latter people, embarrassingly including quite a few people I’ve helped to build their own computers. Although saying that, lightning will probably now strike all of the hardware that I own… I’ll keep you posted.