Never been one to let the carrier drop

I think one of my major failings as a person is that I spend entirely too much time thinking about the past and the future, and not nearly enough time thinking about the present.

I don’t really understand how I got to be this way, but I think it’s a source of a lot of unhappiness for me. Among other things, a lot of my thoughts of the future are in the format “when I ….. then I’ll be happy”, where … can be replaced with get out of debt, figure out how to make N work correctly, etc. And, of course, I get whatever it is accomplished, and then I come up with some new value for …

As far as looking towards the past, I think of things and people long gone – in some cases, dead – and miss them, and feel sad. I think a little bit of this is normal and even positive, but I think I take it entirely too far.

The past is gone, and I should look towards tomorrow not in a planning sense, or in the sense of being worried or fearful about what could come, but in a sense of being excited about what the future could bring.

I sound like chicken soup for the soul or something.. yeesh…

root@peterbilt:/# tw_cli
//peterbilt> info c4

Unit  UnitType  Status         %Cmpl  Stripe  Size(GB)  Cache  AVerify  IgnECC
——————————————————————————
u0    RAID-50   OK             –      256K    1676.32   ON     OFF      OFF

Port   Status           Unit   Size        Blocks        Serial
—————————————————————
p0     OK               u0     298.09 GB   625142448     9QF1ZFXB
p1     OK               u0     298.09 GB   625142448     9QF1ZCFP
p2     OK               u0     279.46 GB   586072368     4NF1V1WY
p3     OK               u0     279.46 GB   586072368     4NF1ZDHK
p4     OK               u0     279.46 GB   586072368     3NF1HAD5
p5     OK               u0     279.46 GB   586072368     4NF1V2M2
p6     OK               u0     279.46 GB   586072368     4NF1RWDA
p7     OK               u0     279.46 GB   586072368     4NF1V2JD

//peterbilt>

So, peterbilt’s RAID finally formatted out 100% good.

We’ve learned many things from this adventure:

1) PCI-X and PCIe are NOT the same connector

2) When partitioning RAID devices, aligning to a 1 megabyte sector boundery helps (use 32768s in parted)

3) parted is much easier to use than fdisk for certain things

4) msdos disklabels are no good for volumes bigger than 2T. Use gpt disklabel instead

5) onboard RAID controllers on motherboards (i.e. ‘fakeraid’) and ubuntu do not get along very well

6) The RocketRaid products are not very good either although they at least work albiet very slooowly

7)  Do not ignore warnings about RAID arrays lest they become more serious warnings about RAID arrays

8) mke2fs has -T largefile for files that contain mostly huge files. Saves time, makes less inodes

9) -E stride=N – N is your RAID card’s stripe / 4. Does help. I knew this before, but was good to be reminded.

10) parted needs better error checking on msdos filesystem for sector counts that wrap

11) raid 50 is *incredibly* forgiving.

12) On thermaltake cases, DO NOT LEAVE THE SIDE PANELS OFF! They keep drives cool to the touch when the panels are on*.. or will burn you when they are off. I figure i hastened at least two drives to their graves. I feel vaugely bad about doing a RMA on them…  but why doesn’t SMART *report* that they’re overheating? Good equipment shuts itself down and lives on to fight another day, IMHO..

13)  Ultimately, the last ditch backup strategy that worked was to go buy a couple of 1T external disks and manually weed the directories out into <1T chunks, then pass some –exclude= lines into tar. Many many higher tech things were tried first, all failed. [But someday soon I will have one heck of a backup server. Yes, backups have now progressed in my life to where they warrent their very own server]

14) Certain unnamed people were extrordinarily helpful in retaining my sanity, and were pillars of strength..

15) I need to get some kind of psychological help – when the thought of losing a terabyte and a third of data is this traumatic.. I mean, literally, I haven’t slept well since this thing started.

In the course of a long life, a man must be willing to abandon his baggage many times. (probably misquoted) –Lazarus Long

I can’t even abandon my data..

* = in peterbilt’s configuration, which is eight SATA disks in 5.25″ removable caddies with individual fans

Does anyone have any suggestions for tools to aggregate a bunch of RSS pages together to form something like LJ’s friends page, but spanning several different blog sites?

By the way – and I know most of you already know how to do this and wouldn’t need my help – but, if any of my friends want help migrating their LJ to WordPress – including free hosting if necessary, or hints as to how to set up your own server – let me know.

(No, I still have not forgiven LJ for their ToS)

So, I was pondering operating systems recently..

Of the machines currently connected to the net at our house (note that many of these are turned off – recently, in order to save power, I decided to start using wake-on-LAN and managed powerstrips..) here is a breakdown by operating system

1) Ubuntu: the.edge, peterbilt, qm

2) WinXP: vm, ivy, jam, media

3) Win2003: p2p

4) Win2K: gamer

5) AIX: aixer

6) BSD: beastie

7) RaQ linux: ns2

8) OSX: music

9) RedHat: gateway

Unix-based: 8

Windows: 6

Also, we have two ‘guest’ computers:

nonmundane.org: bsd

kanvas-staging: ubuntu

And, two in the process of being built:

Ubuntu: underdog, halcyon

Other computers that aren’t currently unpacked but could be:

*) Sparc 5

*) DEC Multia (2x)

*) AlphaStation LX433

*) Amiga2000 (x2), 4000 (x1) – I actually even found ethernet cards for them!

So, as most of you know, I got a sDSL 1.5 megabit symetrical line. What you may not know is – it’s not enough..

So, today, theoretically, AT&T is supposed to come deliver a 3.0/512k asymetrical line which I am going to configure such that all of our web browsing goes through it, leaving the 1.5 megabit circuit for the server segment.

Ah, if only FIOS came here.

But I can’t resist, so I’m going to mention.

In my mid-teens, I helped a friend pick out a computer – a 286 – and then dumped the contents of my hard disk onto it so I could install stacker. (Big mistake, but that’s another story). The process included moving 40 megabytes from one disk to another, and then back to the first. We did it over a RS232 cable.  It took about 4 hours each way.

My recent backup and restore was interesting in that the backup averaged 40 megabytes a second, and the restore a little over 80.

Yes, that’s right. It’s not enough that we can carry around a mini-SD card the size of your pinkie that stores 20 times the data of my first – full height, 5.25, 10 lb – hard disk – not only that, we can move the entire contents of my first hard disk in less than a second.

Just think how the people 20 years behind me must feel. Of course, on the other side, at least their computers hadn’t achived levels of complexity that start making them look like thinking beings.

By the way, a quiz question for the geeks out there – if you were going to hypothetically build a video editing workstation, what type of PC would you use, these days?

So, Peterbilt (mail and file server) had a bad video card that was resulting in spontaneous crashes and reboots. After about fifteen of these, the RAID array showed a failed port, and couldn’t rebuild. Finally, after much tearing out of hair (and a attempt to build a backup server that failed – although I hope to get it done soon – I’ve ordered another 3Ware card after attempts with motherboard fakeraid and rocketraid card both ended badly) I backed it up – almost all of the data was intact – and am now restoring it after reformatting.

I ended up formatting with a 256K stripe. Last time I was running 64K, and I had some trouble with mp3s skipping on playback when users were accessing their mail spools – and I found somewhere (http://www.pcguide.com/ref/hdd/perf/raid/concepts/perf.htm) that suggested that filesystems with lots of small files should use a large RAID stripe, somewhat counterintuitively. We’ll see what happens. I’m certainly not wanting to go through the whole backup-format-restore process again.

Interestingly, the restore process went much faster than the backup process even though the array was initializing at the time.

I have this problem.. once I start spending money on computer equipment, I really don’t know when to stop. Today I had to bite my fingers to avoid buying:

* 2 extra disks for PB, to replace any bad drives that come up

* A 3.0Ghz PC to replace the 1.8Ghz one that currently drives our TV set

* A 14-drive rack of 15k RPM 18GB SCSI drives.. (I was just thinking, what a swwweeet video editing array that’d make..)

The sad thing is, I don’t yet know that I’ve totally resisted – all three sound good. But I’m still not out of debt…

maybe when I get back from the U.K.

For those of you interested, we’re going to the UK on the 6th of Aug and returning on the 14th. We will be visiting Enyc (in Swindon) and Madrory (in Gasglow)

Yet another health insurance provider has declined me.

I’ve pretty much decided to stop applying. Among other things, I don’t particularly feel like supporting the current system, and that’s what I’m doing if I pay in – in the unlikely event that I can find anyone willing to underwrite me.

I keep finding things, over and over, which make me think that money is a power for the side of the universe that I would describe as evil. Hm. I guess someone did say money was the root of all evil. And yet, very few embrace my desire to do away with money entirely.