[sf-lug] Rick's explanation of his internet setup.

Adrien Lamothe alamozzz at yahoo.com
Sun Jan 1 13:13:14 PST 2006


Happy New Year!

I'm very optimistic about 2006; I think this will be a banner year for open-source software (with most of the good stuff happening later in the year.)

Thanks, Rick, for explaining uncle-enzo and the network setup and the reference to Raw Bandwidth Communications.

Have you ever monitored RAM usage on uncle-enzo? RAM is really inexpensive these days, even the older PC-100 (you can also usually use PC-133 safely in a PC-100 system.) UPSes have also come way down in price.

I would ask what kernel uncle-enzo is running, but that is probably an improper question due to security concerns.

Cheers,

Adrien


Rick Moen <rick at linuxmafia.com> wrote: Quoting jim stockford (jim at well.com):

>(this message also tests that the sf-lug mailing list
>is working on the linuxmafia.com host.)

You might have noticed that the machine had ten hours of downtime 
(5 am to 3 pm), courtesy of the longest storm-induced power outage my
household has seen in about five years.  More below.

First, just a few words about the current situation, since we-all simply
popped up on linuxmafia.com like mushrooms in the spring.  Jim and a
friend are still busy working on restoring the _real_ SF-LUG mailing
list and Web site, and I believe are considering hosting options.
Meanwhile, I offered to give the mailing list a temporary home on my
Mailman-equipped Linux server, linuxmafia.com.  (Yr. welcome.  Glad to
help.) 

linuxmafia.com (aka "uncle-enzo.linuxmafia.com", for you Neal Stephenson
fans) is the antique 1998-era VA Research model 500 2U server[1] in my
living room in suburban West Menlo Park just north of Stanford U.,
connected through home aDSL service provided by Raw Bandwidth
Communications (which company I recommend strongly).  I am _not_ the
listadmin of this temporary mailing list incarnation; Jim is.  I'm just
the current hardware's bit janitor (sysadmin) and pink-slip owner.  ;->  

As SF-LUG has already learned somewhat painfully, keeping a group's
Internet presence on a machine in someone's residence has its
disadvantages.  Even for the resident, it's a tradeoff:

On the plus side:  Bandwidth uptime is exceptional, thanks to Mike
Durkin at Raw Bandwidth, who is fanatical about such things.  Also, 
I get complete control of all aspects of my machine including physical 
custody.  And it's pretty cheap for what I get, considering all the
other needs simultaneously served by the incoming pipe and 5 routable
IPs.  (If you want business details, consult http://www.rawbandwidth.com/ 
and talk to Mike.  Not me, please.  I'm just a customer.)

On the minus side:  Effective throughput is merely adequate compared to 
what's available at your average colo -- enough that I've survived a
couple of slashdottings, but with some strain.  (I'm spoiled, having
gotten used to the T-1 in my old building in S.F.)  Also, power outages
and other physical-plant issues are _my_ problem.

My long-term solution to the power-outage problem is mostly to punt:  
I've never bothered putting uncle-enzo on a UPS.  The software's
completely bulletproof and the hardware's disposable.  Power outages
knock it offline for maybe half an hour a couple of times a year:  When
the power comes back, all service reliably come back.  If I get unlucky
and a surge burns out a PSU or hard drive, I can bring the machine back
in about an hour on substitute hardware that's ready for that
purpose.[3]

Back in 2001, you may recall that our {cough} friends at Enron and kin
visited upon us an entire summer of rolling blackouts.  _That_ created
some challenges for uncle-enzo, because all of the journaling
filesystems were still pretty beta, and I was at the time still 100%
ext2.  Because of the power crisis, I did an urgent rebuild onto SGI's
XFS filesystem in May 2001[2], thereby fixing the immediate problem.
The next full site rebuild after that, in 2003, I switched to ext3 --
except still using ext2 for /usr (normally mounted read-only) and for
/tmp and /var/log, for performance reasons.

Anyhow, SF-LUG is welcome to use my machine's facilities as long as it
wants, but I anticipate that we'll want to move it to SF-LUG's _own_
machine in good time.  I just wanted to reassure you that 10 hours of 
downtime is something of an outlying case, but is within expectation for
storm season and nothing to get alarmed about.  Because uncle-enzo is 
_my_ principal home on the Internet, I am motivated (and able) to bring
it back up within about a day somewhere, no matter what happens -- root
compromise, hard drive failures, whatever.

[1] Get a load of these 31337 specs:  single-proc PIII/500, 256kB PC100
SDRAM, 2 x 9GB SCSI-2 HDs.  Go, me!

[2] See:  "XFS Conversion" on http://linuxmafia.com/kb/Filesystems/ .

[3] In fact, I really should swap in a slightly better machine and two
less ridiculously small and old hard drives, that is waiting.  I just
haven't had time to do the swaparound.

-- 
Cheers,             
Rick Moen                 "Anger makes dull men witty, but it keeps them poor."
rick at linuxmafia.com                                   -- Elizabeth Tudor

_______________________________________________
sf-lug mailing list
sf-lug at linuxmafia.com
http://linuxmafia.com/mailman/listinfo/sf-lug



		
---------------------------------
Yahoo! Photos
 Ring in the New Year with Photo Calendars. Add photos, events, holidays, whatever.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://linuxmafia.com/pipermail/sf-lug/attachments/20060101/fd6075e5/attachment.html>


More information about the sf-lug mailing list