"Get ... the COMFY CHAIR!!!"

[unixronin]

"Nobody expects the Spanish Inquisition!"  But it's not due to them that I have a comfortable office chair again, it's thanks to Costco.  My feet now lightly touch the floor instead of my knees being raised off the edge of the chair, I'm looking straight at my monitor instead of just up enough at it to give me a stiff neck, and my forearms are parallel to the desktop instead of inclined across the edge.  Plus, of course, now I can just swivel around between babylon5's and vorlon's keyboard, instead of having to physically pick up the [shaky, loose at the joints] chair and scoot it around.

I also acquired a nice pot-bellied cast-iron teapot, finished rebuilding my balky disk array as a plain stripe again (having had performance issues with Solaris' software RAID5), have a full backup running, wrote some Perl code, learned some odd quirks about file copying to and from Windows¹, got checks deposited in the bank, and ... hmm.  What else?  I have a feeling there's something else... Oh, yes, I have my paperwork for New Hampshire walking-disability plates for when I go to see my new doctor on Thursday now that we're insured again.

What I did NOT manage was to attend either of the planned rider gatherings this week in memorial of Andy Bell, a local rider killed in Connecticut last weekend by a Volvo-driving woman who turned left directly in front of him.  I wasn't expecting to make the ride to the Cape today (apart from anything else, I don't know yet if I can manage a ride of that distance with my foot in the shape it is), but I'd hoped to make the gathering at MotoMarket in Acton Thursday night.  As it was, though, the way timing worked out on various, it just didn't happen.  I didn't personally know Andy, but all the same.  I feel a bit like I've failed an obligation.

---

[1]  The first thing, not a big surprise, is that copying files from a Samba server to local disk on a Windows machine is several times faster using cp -av in a cygwin Xterm than by dragging on the desktop.  The second is that the threading of cygwin and win32 produces some very visually disconcerting results, with events that are expected to be strictly serial actually often occurring in parallel².  The third was that although copying from Samba TO Windows was about three times faster using cp -av than using the Windows GUI, when copying BACK from Windows to the Samba server, GUI click-and-drag was about five times faster than cp -av.

[2]  The command was 'for f in * ; do if [ ! -d /cygdrive/c/minbar/$f ] ; then cp -av $f /cygdrive/c/minbar/ ; fi ; done'.  You'd normally expect each directory to be copied in turn, wouldn't you?  Guess again.  At times this had as many as five cp processes running at once, copying five directories in parallel.

Comments

[robbat2.livejournal.com]

IIRC from messing with cygwin in the past, it's got a weird behavior of going into to the background, instead of staying in the foreground.

[unixronin.livejournal.com]

Yeah, it's apparently all to do with thread behavior. Results are supposedly guaranteed to become available in the specified order of process execution, but there's no guarantee of sequentiality of process execution. (I say supposedly because, well, if I'd been doing something that was dependent on the order of the displayed output....)

[plutosonium.livejournal.com]

I have just the opposite chair problem. It's hard for me to find chairs low enough that my feet rest on the ground instead of dangling.

[unixronin.livejournal.com]

Hey, we can make you a great deal on a set of four chairs. :)

They don't swivel, and they need new seat-cushion foam, and they really need refinishing, and the frames need some tightening .... well, actually, probably not worth the trouble to move them from NH to MD. But I still love the fabric. It reminds me of one of my conspicuous-consumption antique haori.

[databeast.livejournal.com]

Last night I was very drunk.

I dont remember much

but I do remember drunkenly bitching in camaraderie with someone over how shitty the wait states and complete lack of DMA blitting usage was in Solaris 9's software Raid5....

and yes, explorer is a shitty copy system. I usually just remote-mount my windows machines and do that actual copying from the linux side of things, seems to operate much faster and use far less memory to use the windows machine as a server in the equation.

aprenjlty that progress bar notify window is a huge bottleneck in win32 or something equally retarded.

[unixronin.livejournal.com]

and yes, explorer is a shitty copy system. I usually just remote-mount my windows machines and do that actual copying from the linux side of things, seems to operate much faster and use far less memory to use the windows machine as a server in the equation.

Well, what I was actually doing here was backing up the contents of the RAID5 on minbar (Sun U30) to vorlon (Athlon64/PCIE/SATA/Win2KPro), which happened to have 80GB free on its SATA disk, the only place I had room to store the data while I re-initialized the array. I couldn't back it up to tape because the machine wouldn't handle the sustained RAID5 read traffic for long enough without falling down and the array dropping offline, but it could handle a copy to the Win2K machine.

The real surprise wasn't that a cygwin cp was faster copying from minbar to vorlon ... it was that the GUI was so much faster when copying back the other direction. Doing that via cygwin cp was PAINFULLY slow.