yet another major server failure

urgh. this early morning (night) my barebone (o2) once again failed (i am not kidding, this is *real* though i’ve recently made some fun of “system failures” on this blog). looks like i’ve roasted my second hard disk (this time it’s a toshiba). that’s a gau. dang. conclusion: trying new things such as using a passively cooled, tiny, totally silent, environmental-friendly low-power server is inherently risky. atm, i would not recommend it (lex barebone) to anybody.

so i had to quickly replace my just installed 40 gb hdd (hitachi former ibm) of my vaio notebook with the previous 6 gb one (luckily i haven’t erased its contents yet). replaced the damaged 40 gb hdd of o2 with the equally sized one taken from my notebook. restored the latest backup using dd, sfdisk and rsync (went “relatively” smooth as from the last super-gau i’ve written down some walk-through instructions) and here we are again.

unfortunately, the last full backup dated from june 13 (need to lower backup intervals in cron as the specified sync points are often missed). this means, that i’ve lost all data since then (mail, blog, website, data repository, settings, updates etc.). among them some very important e-mail messages (please resend messages since june 13, thanks!). it’s a pity for the blog too (i particularly miss the second feedster entry and the interesting comments about it). i hope there’s something in google’s cache as i’ve recently just re-allowed google to cache /news ;)

nevertheless i will try to “restore” missing mails and blog entries, but chances to succeed are very small. the bios doesn’t even recognize the damaged/burnt hdd anymore. pretty bad thus..

as a first emergency measure, the barebone now runs cover-less (“cabrio” ;) to prevent overheating (i suspect the nightly running updatedb overheated the hdd). i will draw the necessary conclusions from the whole mess..

an exec shield for o2

i’ve just applied ingo molnar’s exec shield patch to a current vanilla kernel rc:

The exec-shield feature provides protection against stack, buffer or
function pointer overflows, and against other types of exploits that rely
on overwriting data structures and/or putting code into those structures.
The patch also makes it harder to pass in and execute the so-called
‘shell-code’ of exploits. The patch works transparently, ie. no
application recompilation is necessary.

works fine. i haven’t noticed any significant performance decrease so far.

downtime of o2

heh.. today at 08:32 cet lovely o2 crashed due to overheating. I nearly burnt my hands when taking it out of its “glasshouse”. nothing serious though.

today is a hot day indeed ;)

Currently evaluating b2++ and WordPress

I’ve just installed WordPress and b2++. WordPress works, but with a few errors still. B2++ looks (code-wise) much more advanced, but so far it does not render its index page ;) Here are some screenshots:

wordpress_01.png
wordpress_02.png
b2++.png

Although I really would have liked to test the new Smarty templating engine of b2++ I have to postpone it till later. WordPress’ features are nice, but not overwhelmingly nice. It offers a better editor than MT and it doesn’t require site rebuilding upon changes.

I guess I will evaluate bloxsom soon. Until then, I won’t migrate anything.