On Thu, 15 May 1997, Scott Armstrong wrote:
> Here at saluki.com we have 5 lists running on majordomo v1.94.
> Four of these lists are small >200 but one of the lists is up
> to about 2400 subscribers. There are about 30 posts per day on
> the large list. My question is do we have what it takes to run
> these lists? About half of those on the large list do not
> receive much from the list. The small lists are running ok.
>
> Here is what we are running:
> Red Hat Linux v4.1
> Sendmail v8.7.6
> Majordomo v1.94.3
>
> This is all running on a p120 w 96mb ram. The internet
> connection now is ISDN @128kbs but next week a T-1 is being
> installed. Will the upgraded line speed fix our problems? Any
> recommended tweaking in sendmail?
Well, that doesn't sound excessive in terms of load for the
machine, but you really might be overloading your
connection bandwidth.
Here's an apples vs oranges comparison:
I have a machine that hosts 54 lists (each has digest and
archives) that serve a total of 3464 subscribers. Message
traffic varies from list to list and with time. My biggest and
most active list has 755 list subscribers and 914 digest
subscribers and traffic of 60+ messages daily.
I am running:
Software: SlackWare Linux v3.1
Linux kernel version 2.0.30
Sendmail 8.8.4/8.7.3
MajorDomo v1.9.3
perl 5.003
Hardware: cpu 486 (Am5x86-WB) at 133 MHz
bogomips 66.15, 48MB DRAM, 64MB SWAP
hda: Maxtor 71260 AT, 1204MB w/256kB Cache
sda: SEAGATE ST12400N
Note: I was not fully servicing my mail queues when I
had them located on the IDE drive. I measured a
throughput of only 2,220 messages per hour.
When I moved the queues to the SCSI drive, I got
a throughput of 3,750 messages per hour.
Majordomo and Sendmail are disk I/O intensive.
Network: EtherNet 10BaseT at machine.
T3 from Campus backbone to world.
Here are some of the mailstats program output for single days of
usage (week of May 9 to May 15 1997):
Statistics from Fri May 9 00:00:27 1997
M msgsfr bytes_from msgsto bytes_to Mailer
0 0 0K 612 1465K prog
3 1558 7542K 9 99K local
4 1690 27792K 47126 174604K smtp
========================================
T 3248 35334K 47747 176168K
Statistics from Sat May 10 00:00:33 1997
M msgsfr bytes_from msgsto bytes_to Mailer
0 0 0K 223 466K prog
3 406 1493K 9 99K local
4 196 2026K 5252 32715K smtp
========================================
T 602 3519K 5484 33280K
Statistics from Sun May 11 00:00:27 1997
M msgsfr bytes_from msgsto bytes_to Mailer
0 0 0K 248 487K prog
3 769 2052K 9 97K local
4 252 749K 18404 39971K smtp
========================================
T 1021 2801K 18661 40555K
Statistics from Mon May 12 00:00:33 1997
M msgsfr bytes_from msgsto bytes_to Mailer
0 0 0K 395 831K prog
3 1065 3263K 9 97K local
4 1892 65432K 23100 146759K smtp
========================================
T 2957 68695K 23504 147687K
Statistics from Tue May 13 00:00:28 1997
M msgsfr bytes_from msgsto bytes_to Mailer
0 0 0K 496 1018K prog
3 1549 4843K 77 261K local
4 2491 21071K 38846 136504K smtp
========================================
T 4040 25914K 39419 137783K
Statistics from Wed May 14 00:00:28 1997
M msgsfr bytes_from msgsto bytes_to Mailer
0 0 0K 474 1193K prog
3 1694 41442K 73 18524K local
4 1706 31359K 43397 149585K smtp
========================================
T 3400 72801K 43944 169302K
Statistics from Thu May 15 00:00:20 1997
M msgsfr bytes_from msgsto bytes_to Mailer
0 0 0K 506 973K prog
3 1251 4340K 38 170K local
4 1729 14366K 36566 97665K smtp
========================================
T 2980 18706K 37110 98808K
In terms of assessing capacity as a mailer and mailhost, I once
(*blush*) had a message that looped peculiarly. I moved 11
Gigabytes of mail outbound that day before I caught it. However,
I shut the computer down and destroyed the mailstats log for
that day. Here are some of the high usage days that I *can*
find mailstats for:
Statistics from Fri Feb 14 00:00:18 1997
M msgsfr bytes_from msgsto bytes_to Mailer
0 0 0K 16187 16576K prog
3 16344 21827K 15 1839K local
4 3335 1110404K 23048 1152534K smtp
========================================
T 19679 1132231K 39250 1170949K
Statistics from Wed Feb 19 11:53:24 1997
M msgsfr bytes_from msgsto bytes_to Mailer
0 0 0K 413 809K prog
3 702 13714K 7 20K local
4 5915 1633184K 28002 1683110K smtp
========================================
T 6617 1646898K 28422 1683939K
Statistics from Thu Feb 20 00:00:31 1997
M msgsfr bytes_from msgsto bytes_to Mailer
0 0 0K 538 1171K prog
3 1066 8630K 14 950K local
4 4264 1758008K 51615 1879994K smtp
========================================
T 5330 1766638K 52167 1882115K
> TIA
> Scott
> saluki.com
>
My best guess is that your connection bandwidth is your weakest
link.
-------
Paul M. Sittler (root) | / / (_)__ __ ____ __ | The choice
Leviathan System Admin | / /__/ / _ \/ // /\ \/ / | of a GNU
email: p-sittler@tamu.edu | /____/_/_//_/\_,_/ /_/\_\ | Generation
A 486 is a terrible thing to waste...
References:
|
|
