|
|
| Author |
Message |
| 25 new of 130 responses total. |
garya
|
|
response 100 of 130:
| 
Mar 19 02:02 UTC 1996 |
Thanks to everyone! My $60.00 is/will be enroute to Grex!
|
mcpoz
|
|
response 101 of 130:
|
Mar 19 02:07 UTC 1996 |
Hey, this Grex is smokin'.
|
garya
|
|
response 102 of 130:
|
Mar 19 02:15 UTC 1996 |
Woo-Hooo I made 100 !!!!!
|
srw
|
|
response 103 of 130:
|
Mar 19 02:51 UTC 1996 |
Congratulations, Gary, and thanks for your support.
THere are a lot of wonderful things we can do with the money that comes in
from people who are willing to help out.
|
steve
|
|
response 104 of 130:
|
Mar 19 03:07 UTC 1996 |
And, the neat part about getting to the Sun-4 is that we're on
the bottom-of-the-line SPARC CPU chip. Next stop, 4/400's...
|
srw
|
|
response 105 of 130:
|
Mar 19 05:26 UTC 1996 |
That's a great concept, STeve, but your getting ahead of yourself a bit.
We are greatly in need of other upgrades first. Newsfeed, more bandwidth,
faster modems. These are all coming, but we need money for them. They're
all ahead of the even *faster* cpu STeve is talking about.
Eventually, though, we're ready for it.
|
rcurl
|
|
response 106 of 130:
|
Mar 19 06:33 UTC 1996 |
I was going to post
1:04am up 1 day, 4:43, 55 users, load average: 3.70, 3.02, 2.25
but I won't as it doesn't trump #99.
|
tsty
|
|
response 107 of 130:
|
Mar 19 06:50 UTC 1996 |
i tell'd srw with 56 users and 3.45 load earlier .... Extra Beans!
this is wonderful - thakxx to all who contributed, sincerely.
|
gregc
|
|
response 108 of 130:
|
Mar 19 09:33 UTC 1996 |
Here, I'll trump myself:
8:57pm up 1 day, 36 mins, 48 users, load average: 1.47, 2.09, 2.55
|
steve
|
|
response 109 of 130:
|
Mar 19 12:52 UTC 1996 |
Oh, I *know* we ain't gonna upgrade the CPU/memory for a while
yet. That wouldn't be good--too many other things to do, first.
But when the time comes, and it will, the 4/400 series of CPU card
makes a lot of sense. It turns out we may also have to upgrade
our SCSI controller at that time too; we need to do some more
investigation on that.
But we won't have to recompile the world!
|
popcorn
|
|
response 110 of 130:
|
Mar 19 17:56 UTC 1996 |
...for which there was *much* rejoicing.
|
srw
|
|
response 111 of 130:
|
Mar 19 18:43 UTC 1996 |
Indeed. Although "recompiling the world" understates by a huge amount
the complexity of the sun-4 upgrade. We are running a new OS, and spent more
time getting all of the details of that right than all of the compiling put
together.
STeve's point is not only that we don't have to recompile the world, but we
also don't have to change OSes. This is a major big deal. Yup. It'll be
real easy compared to what we just went through
|
tsty
|
|
response 112 of 130:
|
Mar 20 17:32 UTC 1996 |
Sometihg seems strange ... now the load average isin the high teens
again ... is this a problem withthe new arrangements or can these
high load averages still happen frequesntly?
|
arthurp
|
|
response 113 of 130:
|
Mar 20 18:49 UTC 1996 |
I saw 30s. :( Went away before I could snoop enough to figure it out,
though. :)
|
srw
|
|
response 114 of 130:
|
Mar 20 18:51 UTC 1996 |
Spikes of high load average can usually be traced to mail.
Mail will not stop processing its queue until the load average hits 8, and
will not stop accepting incoming mail until the load average hits 12.
I saw an interesting phomenon yesterday or the day before. The load average
spiked, and so I scanned all of the processes to see why. There were 50
sendmails running, almost all connected to msn.com.
My suspicion is that there had been a period of substantial downtime at the
Microsoft Network, and when they came back up the barrage began. We could
reduce the impact of that by lowering the sendmail load average shutdown
points. I don't know if we want to, though.
|
janc
|
|
response 115 of 130:
|
Mar 21 07:19 UTC 1996 |
It appears that the high loads were due to various users doing absurd things.
|
srw
|
|
response 116 of 130:
|
Mar 21 07:23 UTC 1996 |
yes, well, "usually"
|
popcorn
|
|
response 117 of 130:
|
Mar 21 14:49 UTC 1996 |
Also, it looks like we may have a problem with sending/receiving mail to msn.
|
kaplan
|
|
response 118 of 130:
|
Mar 21 20:21 UTC 1996 |
I'm much less concerned about load averages than stability. Uptime is
currently over 2 days. Knock wood downtime from now on will be scheduled.
|
srw
|
|
response 119 of 130:
|
Mar 22 06:20 UTC 1996 |
Both load speed and reliablity are important.
People have given up in disgust because we were so slow.
Speed was the most commonly cited objection M-netters had to Grex.
(Galactic boringness came in second)
|
ajax
|
|
response 120 of 130:
|
Mar 22 07:02 UTC 1996 |
The unreliability didn't annoy me nearly as much as slowness, but I'd
think from the perspective of the dedicated staffers who resuscitated
Grex daily, the unreliability would be the bigger irritant. Sure is
nice to have both aspects so much improved...thanks to the many people
who worked on this!
|
scg
|
|
response 121 of 130:
|
Mar 22 07:10 UTC 1996 |
re 117:
That's because we need to switch to Microsoft TCP/IP, and a Microsoft
message transfer agent. There's nothing nonstandard about either MSN or NT.
It's the rest of the Internet that's set up wrong. ;)
|
steve
|
|
response 122 of 130:
|
Mar 22 16:15 UTC 1996 |
I've been watching the system for a while now and not noticed
really sharp peaks at times, and usually, they are due to news.
When, in the course of two minutes 33 sites started trying to give
us mail the load average jumped up about 12, to 19 or so. Watching
the connections close went right along with the falling load average.
|
adbarr
|
|
response 123 of 130:
|
Mar 22 20:28 UTC 1996 |
I have been away for a while, did you guys get a new extension cord or
something? Seems like this is faster somehow?
|
tsty
|
|
response 124 of 130:
|
Mar 23 01:28 UTC 1996 |
about microsoft.com ... i got some MIME encapsulated stuff the other
day .... so i ran (erp) pine to separate the enclosure neatly. At least
that was the idea .... Obviously it failed.
then i thought, well, it's just a kind-of wrapper-thingie for uuen/decode
so i'll do it manually. That failed, no "begin" statement.
Has Grex lost it's MIME-aware tools? Is there an un-documented trick
hiding out there somewhere? I kinda need this stuff.
|
