This item is for system announcements (new computer equipment on Grex, 
system upgrades, Grex meetings, etc.).  Personal announcements should go 
back in item 2; Grex system *problems* belong in the next item (#4). 

23 responses total.

   Grex now has a new 36G disk to replace the dead 18G one.  Changes
to Grex include more space for the /var partition, such that we won't
run out of space for things in the future.  We're also running "process
accounting" now, which uses lots of disk space.  With this we are able
to see what commands have been run, helping determine who ran something
like a forkbomb if another kills the system.

   To help prevent nasty programs from hurting Grex the number of
processes that a user can run has been set to 24.  This means you
can have 23 different things running on Grex at once.  An example
might be a couple of watch's, bbs and mail all at once.  That
would be 4 or 5 processes running, which is a long ways off from
24.  I don't think this is going to be a limit for any normal
usage of Grex.  Before I brought the system online yesterday, I
wrote a little forkbomb and let it run for a while as I did other
things.  Basically, the system didn't notice the merry forking
that was going on, limited to 23.  I briefly had the forkbomb
running with a limit of 128 processes and the story was different.
There was a huge delay for other things to get done, so with 24
(or even 32) processes I think we'll be in much better shape.

   Grex is now creating four copies of the /etc/master.passwd
and /etc/group files at 2am, 8am, 2pm and 8pm each day, so we have
four copies of the passwd file on Grex.  We had this on SunOS.  I
wish I'd gotten that running before.  I have a machine that is
going to be making sftp (secure FTP) transfers of these files
every day, and for further paranoia store the files on a machine
which isn't routed on the net.  I intend for Grex to never have
password file (or other critical config stuff) problems again.

stunning! and immense thank you-s! thank you!

re #1:

 > Grex is now creating four copies of the /etc/master.passwd and
 > /etc/group files at 2am, 8am, 2pm and 8pm each day, so we have
 > four copies of the passwd file on Grex. 

 Perhaps this goes without saying, but the backup copies are stored
 on a totally different disk (that is, different hard drive, not just
 different partition) from the master copy in /etc, right?

Hooray, and thanks.

   They're stored on all four physical disks.  There is a copy on sd0;
this isn't useless, since it isn't on the a partition, and might be
useful if staff ever wipes out /etc again.  Anyway, for the paranoid
all four disks have protected copies now.

36G??? I didn't think you could still get hard drives that small. I would hope
you dropped a zero there.

   No, 36G.  Grex runs on Ultra-320 SCSI disks.  These disks blow IDE disks
out of the water in terms of performance, and in erms of overall quality.
Grex isn't so much a disk space monster, as an I/O hog.  Having a u320 
disk means that we're able to increase our disk throughput by getting a
ultra-320 scsi disk controller, if we want/need to.  And these are 15K
rpm disks, too.

Whew. Thank goodness (and the staff) that grex still lives!

I have a 250MB hard drive on a ten-year-old computer.

Actually, I think it's 11 years old.

SCSI disks are still uncommon and expensive in larger sizes, for some
reason.  I think maybe SATA is slowly taking away the market.

   Right.  You can get 147 u-320 15k disks, but those are the biggest
that I know of.  We'll see of sata disks prove reasonable.  So far, I
don't think they have proven themselves.

In what way?  Reliability?  I don't know if it's true or not, but I've 
heard claims that a lot of new SCSI disks are actually SATA disks in 
disguise -- basically the same hardware, but with a SCSI interface 
layer. 
 
I'm just starting to experiment with them, myself.  The new voice mail 
system I set up at work has a pair of SATA drives in a RAID-1 
configuration.  (Software RAID, that is.  Most "SATA RAID" controllers 
are actually software RAIDs anyway.) 
 

   Things could be changing, but scsi disks get more careful scrutiny
on the way out the door.  I got this from a technical representitive
from a disk company last year.  The sata disks are the up and coming
thing however, and really, one shouldn't be more reliable than the
other.  All the mechanial work on disks is the same, as is the way
you make them clean.  Electronics are well, electronics.  But ide is
the biggest until recently in terms of size, and scsi still has
the best overall reliability ratings.

   Grex now has a uptime log as we did before.  /var/log/uptime.log,
updated every 6 minutes.

This response has been erased.

 >  I don't know if it's true or not, but I've heard claims that a lot
 >  of new SCSI disks are actually SATA disks in disguise -- basically
 >  the same hardware, but with a SCSI interface layer.

 That doesn't make sense to me for two reasons:

  1)  SATA drives seem to come in sizes comparable to PATA drives.
  2)  Although they may exist, I've never seen a 15krpm SATA drive,
      or even very many 10krpm SATA disks.  SCSI drives pretty 
      routinely come in those speeds. 

tigerdirect.com has a 300GB 10k Ultra-320 hard drive for $1229.99.

Whats the sound of Grex going bankrupt buying hard drives?

   Non existant.  Grex wouldn't do that.

cLEAR.YL  non-existant.

I applied a fix to backtalk that should end the crashes when you read
anonymously.

I hate children.
That is all.

 Can't take the competition?

can't fight the bullies :(

Response not possible - You must register and login before posting.

- Backtalk version 1.3.30 - Copyright 1996-2006, Jan Wolter and Steve Weiss