I recently realized that I no longer needed a "full-scale" notebook, and that since netbooks were so inexpensive, I could actually make money selling my Inspiron 6400 and buying a Dell Mini 10v to replace it. So I did that. Much like the Inspiron it replaced, I did pimp out the 10v a bit over baseline - upgraded to Dell's 16GB solid state drive, upgraded the wireless, and upgraded the battery (from 3-cell to 6-cell).
I was never a big fan of the netbook (or, formerly, "subnotebook") form factor - I thought that it sacrificed too much usability for the sake of "being tiny and cute" - but given that I no longer really needed a desktop replacement to go on long trips with me, and given the much lower cost, the netbook was suddenly attractive. For occasional, relatively brief use doing network testing, it's hard to beat.
OK, OK, so we were paid in merchandise and it's just for this chick's Etsy store... but hey, there was a professional photographer shooting it.
This one is by far my favorite... Jane kicking it Roaring Twenties style with a crazy giant-flower-adorned lace knit headband. I suspect we are going to be seeing that expression A LOT. In Janis' words, "that's the Daddy look."
Seriously, why isn't this getting more news coverage? Let me recap:
We have not declared war on Pakistan
We are using military hardware to attack targets within Pakistan's borders
We are attacking civilian targets within Pakistan's borders
WE HAVEN'T EVEN DECLARED WAR!
Why the hell isn't this raising red flags all over the country? I mean, for crying out loud, have you ever watched a Western? The guys who ride by the ranch firing their guns into it through the walls at midnight are not the heroes of the piece.
I'm really not coming at this from the perspective of a fainting dove: I know the place is a goddamn mess, and there's a lot of Taliban in there. But this just isn't the right way to go about... whatever the hell it is we're trying to accomplish. We should either be getting openly and officially invited in by the Pakistani government, declaring war and doing it our way, or butting the hell out. "Just sorta sending in drones to attack civilian targets with HE missiles", with or without the "tacit" support of the Pakistani government - and with increasingly strident protests from Pakistani civilians - is a direct slap in the face of everything this country is SUPPOSED to be for.
Update to the prior ZFS post: I did some more benchmarks. The newer set was performed with one 5MB burst of data written to a random location on the target drive each second, which is probably a better model for most real-world conditions.
I've been benchmarking Sun's ZFS filesystem alongside more conventional ones lately, and as a result, somebody asked me to include numbers from Sun's actual operating system running ZFS. So I figured hey, why not.
I felt like that little girl from Jurassic Park. First, I sit in front of the computer, and stare at the monitor with a relieved grin. "Hey, this is Unix - I know this!"
Then the velociraptors attacked.
I suspect there must have been driver problems with the particular hardware I was using, because holy SHIT was OpenSolaris (2009-06) ever broken. First, any attempt to read from /home/export/myname severely impacted performance of the whole system - making it damn near unusable - until the read finished. Um, what? Then I discovered that if I had a terminal window open, and I fired up nano (a text editor) and then maximized the console window... suddenly, arrow keys started producing ASCII codes on the screen instead of... you know... arrowing. And once a terminal window actually crashed while I was using nano. Then I tried to install the Data::Random module from Perl's CPAN, and... well, the less said about that the better.
I could also go on about dd not understanding the "m" or "g" suffixes to blocksize arguments - despite the manpage saying that it does - or the lack of throughput report at the end of a dd run - despite the manpage saying there should be - or the lack of normally pretty-universal utilities like pv (pipe viewer) in the package repositories... hell, I think you get the picture. There may be some Very Good Reasons to go Solaris instead of BSD or Linux in the Enterprise world that I'm just not aware of, but from a smallbiz ISV's perspective, it felt like HP-UX: sorta like what you're used to, only painfully obsolescent and quirky and treacherous, with a lot of things you expect to be available inexplicably missing.
Again, I have to believe a lot of my problems must have been endemic to the hardware I was using. (Athlon64 3500+, ASUS motherboard with nvidia chipset, 2GB DDR2 SDRAM, if anybody is wondering.) But honestly, from my perspective as an ISV catering to small business... that doesn't help much. I don't want to have to carefully pick and choose from a list of exotic crap to feed my OS; I need to by and large expect it to just fucking work with commodity off-the-shelf parts.
I'll put up new numbers tomorrow, including results from FreeBSD 8.0-CURRENT (which includes ZFS v13, as opposed to the ZFSv6 in FreeBSD 7.2-RELEASE) as well as the OpenSolaris results. But I'll save you the tension, if it's the Solaris numbers you want: they sucked. (Which may largely have had to do with the fact that OpenSolaris seems bound and determined to run X Windows, whereas FreeBSD didn't - all I really know is, from the "install it and see how it works" perspective, it blew.)
A comment on the Ars Technica Linux Kung Fu forum a couple of weeks ago got me curious - a user there said that as far as he knew, RAIDZ was not supposed to be a performance configuration, with RAIDZ performance not much better, on average, than that of any single disk in the RAIDZ.
I just happened to have a RAID storage server in the shop that was due for a complete wipe anyway, so I decided to take the opportunity to do some benchmarking. Somewhat to my surprise, ZFS turned out to be quite a good performer - despite its advanced data-protection features, it was the fastest filesystem tested for single-process reads, with or without RAIDZ. RAIDZ did quite well too; on multiple concurrent reads it is significantly slower than RAID5/ext3, but still manages to nearly double single-drive performance across the board.