Howdy,
I am planning on moving my SBBS to a low powered PC - but then I noticed it takes around 5-8%% CPU on it constantly - even while there is nobody on, and no mail transferring.
On this (old) system - with much faster CPUs it takes 2-3% constantly (which probably wouldnt normally be a problem to most).
I'm wondering why its taking up so much CPU, and if there is something I can do to reduce it? (ie: smtp/nginx uses negligble CPU while waiting for work.)
I've disabled a few services that I dont use, like web/ftp, etc with no real difference in CPU usage.
An strace, shows it doing this iteratively:
[pid 850] <... nanosleep resumed> NULL) = 0
[pid 850] select(21, [20], NULL, NULL, {0, 0}) = 0 (Timeout)
[pid 850] select(21, [20], NULL, NULL, {0, 0}) = 0 (Timeout)
[pid 850] nanosleep({0, 5000000}, <unfinished ...>
[pid 18352] <... nanosleep resumed> NULL) = 0
[pid 18352] ioctl(0, TCGETS, {B38400 opost isig -icanon -echo ...}) = 0
[pid 18352] select(1, [0], NULL, NULL, {0, 0}) = 0 (Timeout)
[pid 18352] nanosleep({0, 1000000}, NULL) = 0
[pid 18352] ioctl(0, TCGETS, {B38400 opost isig -icanon -echo ...}) = 0
[pid 18352] select(1, [0], NULL, NULL, {0, 0}) = 0 (Timeout)
[pid 18352] nanosleep({0, 1000000}, NULL) = 0
[pid 18352] ioctl(0, TCGETS, {B38400 opost isig -icanon -echo ...}) = 0
[pid 18352] select(1, [0], NULL, NULL, {0, 0}) = 0 (Timeout)
[pid 18352] nanosleep({0, 1000000}, NULL) = 0
[pid 18352] ioctl(0, TCGETS, {B38400 opost isig -icanon -echo ...}) = 0
[pid 18352] select(1, [0], NULL, NULL, {0, 0}) = 0 (Timeout)
[pid 18352] nanosleep({0, 1000000}, <unfinished ...>
[pid 850] <... nanosleep resumed> NULL) = 0
Are their ways to reduce this?
Probably. Have you compared running sbbs both as a daemon and not as a daemon? It might take some more experimentation on your part to find where the CPU utilization is coming from exactly.
Re: Idle CPU usage
By: Digital Man to Alterego on Sun Dec 15 2019 03:22 pm
Probably. Have you compared running sbbs both as a daemon and not as a daemon? It might take some more experimentation on your part to find where the CPU utilization is coming from exactly.
So I have 2 running instances.
One starts with just "sbbs" in the foreground.
The other starts via supervisor with "sbbs -syslog"
Are you suggesting that starting SBBS with "-d" would result in less CPU usage?
Howdy,
I've been playing with mailling lists - with my board receiving maillist emails - however, I've noticed they are not posting:
mail 0036 SMTPS !luv-main-request@... (user #0) cannot post on LUV Main (reason: 13)
Where can I find out what "reason: 13" is?
In my ACL I have "LEVEL 99 OR USER = 0" as per the example in the wiki...
Configuration step 6 at http://wiki.synchro.net/module:listgate
On 12-16-19 13:20, Alterego wrote to Digital Man <=-
Howdy,
I've been playing with mailling lists - with my board receiving
maillist emails - however, I've noticed they are not posting:
mail 0036 SMTPS !luv-main-request@... (user #0) cannot post on LUV Main (reason: 13)
Re: Idle CPU usage
By: Digital Man to Alterego on Sun Dec 15 2019 08:32 pm
Configuration step 6 at http://wiki.synchro.net/module:listgate
Argh, thank you. I've read that wiki a few times, and I must have missed that.
BTW: Without setting a posts per day for level 0 - is it possiblet to set a "particular email address" (ie: the list from address), can post an unlimited number of times?
And thereby any emails from non list addresses, destined for this sub are rejected?
On 12-16-19 10:27, Digital Man wrote to Alterego <=-
And thereby any emails from non list addresses, destined for this sub are rejected?
Can't think of a way off hand. Posted messages are actually handled by
the mail server, not the listgate module.
On 12-16-19 10:27, Digital Man wrote to Alterego <=-
And thereby any emails from non list addresses, destined for this sub are rejected?
Can't think of a way off hand. Posted messages are actually handled by the mail server, not the listgate module.
I did notice that. I can see potential issues if someone does work out what's going on - they could spam the echo. Not likely anyone would bother, but it's possible. I am making use of that property to effectively funnel 2 mailing lists into one echo, but of course, listgate can only post to one of the lists.
That's fine for my situation, because one of those two mailing lists is being phased out, the other (the one that listgate.js psts to) is its replacement, and I only need to be able to post to that, but want to catch posts from the old list.
On 12-16-19 14:36, Digital Man wrote to Tony Langdon <=-
I did notice that. I can see potential issues if someone does work out what's going on - they could spam the echo. Not likely anyone would bother, but it's possible. I am making use of that property to effectively funnel 2 mailing lists into one echo, but of course, listgate can only post to one of the lists.
That's fine for my situation, because one of those two mailing lists is being phased out, the other (the one that listgate.js psts to) is its replacement, and I only need to be able to post to that, but want to catch posts from the old list.
There's probably some method todo what you want using an external mail processor (ctrl/mailproc.ini).
Can't think of a way off hand. Posted messages are actually handled by
the mail server, not the listgate module.
I did notice that. I can see potential issues if someone does work out what's going on - they could spam the echo. Not likely anyone would bother, but it's possible.
On 12-17-19 09:22, Rampage wrote to Tony Langdon <=-
we had almost the same thing with gigo gating into echos and mailing lists... there was a "secret" poster that we configured for each echo and/or mailing list... if someone figured what it was and the proper format of the options, they could do the same...
we had almost the same thing with gigo gating into echos and mailing
lists... there was a "secret" poster that we configured for each echo
and/or mailing list... if someone figured what it was and the proper
format of the options, they could do the same...
Hmm, I didn't think GIGO could post messages from a single user, or perhaps that came later? I remember people having to individually subscribe and be set to NOMAIL. Wasn't a big issue, because I was
the only user of gated mailing lists on my gateway.
On 12-18-19 07:44, Rampage wrote to Tony Langdon <=-
like these from my maillist.cfg:
Sysop: | BrokenMind |
---|---|
Location: | Central Pennsylvania United States |
Users: | 60 |
Nodes: | 4 (0 / 4) |
Uptime: | 17:28:46 |
Calls: | 175 |
Files: | 2,017 |
Messages: | 20,380 |