2:37pm up 30 days, 9:06, 1 user, load average: 32.01, 29.86, 30.82
472 processes: 464 sleeping, 8 running, 0 zombie, 0 stopped
CPU states: 4.5% user, 26.7% system, 15.5% nice, 53.1% idle
Mem: 1031184K av, 883552K used, 147632K free, 0K shrd, 84052K buff
Swap: 1052248K av, 0K used, 1052248K free 458552K cached

It is busy 24/7 :)

This is a server from rackshack by the way.

processor : 0
vendor_id : GenuineIntel
cpu family : 15
model : 2
model name : Intel(R) Pentium(R) 4 CPU 2.00GHz
stepping : 4
cpu MHz : 1992.571
cache size : 512 KB
fdiv_bug : no
hlt_bug : no
f00f_bug : no
coma_bug : no
fpu : yes
fpu_exception : yes
cpuid level : 2
wp : yes
flags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm
bogomips : 3961.24

Uhm... if its redhat it will most likely tank... if you managed to but BSD on it... BSD is a trooper, it can take that :D .

Originally posted by alapo
Uhm... if its redhat it will most likely tank... if you managed to but BSD on it... BSD is a trooper, it can take that :D .

forgot to add... it's dedicated for mysql :)


Rackshack doesn't have BSD so :)

Originally posted by inteltechs
Rackshack doesn't have BSD so :)

People have gotten BSD on their systems though :).

Originally posted by alapo
People have gotten BSD on their systems though :).

:) forget about BSD solution :)...get back to the topic :)...do you think it is going to last for a while?

Originally posted by inteltechs
:) forget about BSD solution :)...get back to the topic :)...do you think it is going to last for a while?

RH might tank, but I bet MySQL will first.

I've seen a RH box with a 400.00 load and it didn't crash, but it was so lagged you couldn't do anything on it really.

Originally posted by DarktidesNET
I've seen a RH box with a 400.00 load and it didn't crash, but it was so lagged you couldn't do anything on it really.

woo 400.00 load ...what a record!!! does it run like that 24/7 or just a short period of time?

Oh, it was 24/7 unfortunantly. Some guy used to come here and I bought a reseller from him. This was like 1 year ago. His site was bigsiteofhosting.com (I believe)

The machine always had a load of 82-400 constant because they overloaded it to hell.... trust me, it was constant and I suffered. Quickly left after that...

I don't think yours will crash anytime soon. I've seen a BSD box at 1300+ load (or was it 1800) posted to this forum before... but I don't know how long it was up or what it was running because that was months ago.

How exactly is that load number calculated? The processor can be at 100% use for a long time, but not spike load.

I'm not exactly sure to be honest. I do know that the time the process(es) take affects the load.

For example, if you run a mysql query that lasts 5 seconds but uses 100% cput the load will go up a bit then drop, if it eats 100% cpu and lasts 10-50 minutes, your load will be huge.

Each of the loads indicate a time frame, I believe the first is 5 min, second 15 min, and 3rd is 30 min (correct me if I'm wrong).

How it exactly calculates overall load I am unsure, but wouldn't mind knowing myself.

download mytop and see how busy your mysqld is doing. It seems you still have lots of CPU idle + memory, so I think there are just too many processes sleeping. Try to optimize it to reduce the loads.

Originally posted by alapo
How exactly is that load number calculated?

It's supposed to be the number of processes/threads waiting for CPU time at any given time - ie currently active rather than sleeping processes. So a load average of 1.0 for a certain period should mean that there was one process ready to run every time a CPU was ready to schedule time for processes. If the load average is equal to the number of processors you are, in some sense, using the server optimally (no processes needs to wait longer than necessary and no processors are idle).

In practice, the load average may include more than that; it's not completely accurate. But it gives you some indication of server load.

it will survive. Just get lag a bit.

Just curious.. if BSD is more stable then RedHat (especially because big boys like Yahoo! and Tucows use it), why then to people tend to go with RH?

RH is easier to use probably.

Linux/Unix is generally more stable than most Windows os's but people still use Windows because it's easy.

I have servers here that regularly have loads in the 100+ range that have been chugging along for 5 years now.

It's funny.. if my clients would pay just a few hundred dollars to get a new motherboard/chip, and update their software.. their load would probably never go above 10.

A client of mine has an UltraSparc that can handle loads way above 200. Very amazing machine.

If you have a high load web site, and you can't do anything to reduce the load (ie-use mod_perl instead of normal CGI).. try an Alpha with BSD.

Originally posted by userfriendly
I have servers here that regularly have loads in the 100+ range that have been chugging along for 5 years now.

It's funny.. if my clients would pay just a few hundred dollars to get a new motherboard/chip, and update their software.. their load would probably never go above 10.

A client of mine has an UltraSparc that can handle loads way above 200. Very amazing machine.

If you have a high load web site, and you can't do anything to reduce the load (ie-use mod_perl instead of normal CGI).. try an Alpha with BSD.

impressive :)....Thx for the information.

hehe, is that server being used to p4host backup SQLDB or something

Originally posted by Hoobastank68
hehe, is that server being used to p4host backup SQLDB or something

This rackshack server belongs to one our clients.

Here is p4host server load:

9:14pm up 7 days, 2:12, 1 user, load average: 0.71, 0.84, 0.83
205 processes: 191 sleeping, 5 running, 0 zombie, 9 stopped
CPU states: 4.7% user, 3.1% system, 0.0% nice, 92.1% idle
Mem: 902868K av, 853952K used, 48916K free, 0K shrd, 35872K buff
Swap: 2097136K av, 16656K used, 2080480K free 547468K cached

I have seen one of our redhat servers hit 700 before. No joke! It's real high quality (SCSI and Dual p3 tualitins) and took about 10 minutes to login via SSH, but I recovered it without a reboot, and everything was good to go after that!

Anyway, just fyi.... i'm not trying to brag or anything :D

Well, here it is - working horse of my friend. It is running Dual PIII, 1GB ddr, SCSI raid in mirror. Check screenshot attached.

I have seen higher loads on 1Ghz durons, they were still responding to commands, but it takes too long :)

That's crazy... :stickout: what's that machine doing... ? :)

crazy but yet true. pretty busy server running adult sites with mysql and a lot of interactive stuff :)

I had a busy Plesk BSD box last weekend that had been UP since October 3rd I think--someone destroyed that when they reorganized the cabinet and shut it for for a minute :(

The load average is more of an approx indication as to what's going on with your server rather than anything more scientific...

It is the average number of processes running or ready to run over the interval(s), or something along those lines. The three numbers are actually averages for the past 1 minute, 5 mins and 10 mins (taken from `man uptime` ;)).

Miha, you missed the hostname in the title bar! :).

I'm reading this thread and I'm thinking "And here I get worried when the load gets anywhere NEAR 1"!