Load averages


#1

Hi folks,

I’m relatively new to Dreamhost, so I thought I’d post a couple of questions here and see if any of the more experienced members can point me in the right direction.

  1. What is an “acceptable” load average to have on our server? I can understand that, due to it being a shared server, there will be some fluctuations, but I’m curious as to what a good upper limit is before I should bring it to Dreamhost’s attention. Is 5 considered excessive? How about 10?

  2. How long should I wait for the problem to persist before reporting it to Dreamhost? 5 minutes? An hour?

Hoping I can get some answers on these, as I’ve been having some slowness issues on my site in the last couple of days, but I don’t want to be too much of a pain to Dreamhost support either…

Thanks,

– Doug

My Website - My Blog


#2

Each core should normally have a maximum load of 1.0 - most (if not all) server here are dual-core systems, so a load of 2.0 should be fine. However I wouldn’t worry as long as the load doesn’t creep above 5. If it stays above 5 for a longer period contact support, but they also monitor server loads, at least that’s what they are telling us. They might offer to move you to a knew server. I did that move a few months ago and since then I’m on a server which has a load of around 1 and 2, often even below 1 and sometimes it spikes up to 8. In my experience the biggest bottleneck here at DH is the MySQL server.

You might want to do a search, this topic has been discussed here often.


Promo-Code: [color=#CC0000] SAVEMONEY97 [/color] - Save [color=#CC0000]$97[/color] on your first year of hosting.


#3

It’s also good to remember that your filesystem is NFS mounted, so when you see loads shoot way up (say, above 10), chances are it’s a problem with the connection to the NFS server.

Free unique IP and $67 off with code [color=#CC0000]LMIP67[/color] or use [color=#CC0000]LM97[/color] for $97 off. Click for more promo code discounts


#4

load average isn’t any kind of useful performance metric - it just means that a lot of processes are in running state. popcorn sits between 20 and 80 pretty much all the time.

CPU utilization is what you really want to be looking at.


#5

In fact, I’ve been seeing a lot of NFS failures over the past week. (I run my own version of SpamAssassin and when there are NFS problems spam scanning doesn’t complete.) IIRC, the last time this happened, it did cause load averages to increase.

Larry


#6

[quote]load average isn’t any kind of useful performance metric - it just means that a lot of processes are in running state. popcorn sits between 20 and 80 pretty much all the time.

CPU utilization is what you really want to be looking at.[/quote]
Both have their purposes, but if the load average on your server sits between 20 and 80 all the time, your server is not a happy one.

Have you submitted a support ticket about that load average? What was the response?

Free unique IP and $67 off with code [color=#CC0000]LMIP67[/color] or use [color=#CC0000]LM97[/color] for $97 off. Click for more promo code discounts