Load average in the 300s


#1

I am on the chex server and all of my sites are down and I am unable to log in via SSH to my main account. I am able to log in with a sub account, and I see this:

23:07:12 up 1 day, 12:02, 2 users, load average: 383.73, 365.71, 330.32

Usually the load is in the single or double digits, so that seems way too high. Of course the DH status page has no information. I don’t know why I can log in with one account but not the other. My sites were down over 5 hours yesterday and another two hours today. This is just unacceptable. My account renewed last month for a year. Does Dreamhost give a prorated refund for cancelling, or am I stuck with this crappy service for another 11 months?


#2

If it’s at 300+, it just means there’s an NFS issue. Perhaps you should file a support ticket so that they can actually work to fix it or move you to another server, rather than being another typical whinny baby. Otherwise I’ll direct you to the TOS to find out about refunds. You should have read this before even signing up.

Good day.


Chips N Cheese - Custom PHP installs and the like!


#3

There’s a big difference between being a ‘typical whinny baby’ and incurring the same problem over and over and over.

Here’s the current load average on yoda:

top - 06:26:12 up 1 day, 14:17, 5 users, load average: 203.18, 151.12, 107.21
Tasks: 3 total, 1 running, 2 sleeping, 0 stopped, 0 zombie
Cpu(s): 22.4% user, 35.7% system, 41.9% nice, 0.0% idle
Mem: 3615716k total, 3495100k used, 120616k free, 8556k buffers
Swap: 6313512k total, 87500k used, 6226012k free, 902596k cached

PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
29542 coolgeek 9 0 1152 1148 928 S 0.0 0.0 0:00.64 ksh
21678 coolgeek 9 0 1176 1172 952 S 0.0 0.0 0:00.38 ksh
2608 coolgeek 9 0 1036 1036 844 R 0.0 0.0 0:40.82 top

My site is currently throwing blank pages, as it was last night when I filed a Sev 1 ticket (my first Sev 1 ever, I believe). The last time I filed a ticket for same (Sev 2, I believe) was March 29, and I’m sure I’ve been filing tickets for the same problem for over a year.

At that time, support tried to tell me that these NFS issues, and correspondingly absurd load averages, weren’t really a problem, and that they had seen load averages of 800 without problems.

Bullsh!t!

Though the problem was ‘resolved’ last night, they didn’t even respond to the ticket, which leads me to believe they’re kind of sick of dealing with the problem themselves.

I understand the trade-offs of $20/month hosting. But this is a problem that has been recurring, and it’s obviously not just my server.

If, in fact, the problem is truly NFS, then either their NFS configuration, or NFS itself, is unstable and they need to architect a new solution.

This is completely unacceptable.


#4

LOL, I got a response from support 9 hours later and they said they can’t find anything wrong or reproduce the problem, and that no one else on my server is reporting a problem! Maybe if they looked into it when I actually opened the ticket! The load was in the 400s and they say that nothing is wrong. Give me a break. And you have some nerve calling me “whinny baby”. I opened support tickets and I don’t get anywhere with them. They are useless. Must be nice for you to sit back on your problem-free server and call people names when they have repeated problems with their server.


#5

Here I go being a whinny baby again. Sites just went down, load just jumped from 5 to 60 in a couple of minutes and continuing to rise. Now I’ll open another ticket so they can respond in 9 hours and tell me nothing is wrong.


#6

I’ve actually had many problems with the two servers I was on. Currently I’m seeing around 10-20 as the load average for the last week. I’ve seen 100-900 occasionally, but if you actually took the time to TALK to someone and find out why, it’s simply due to a timed out NFS mount. So essentially, you’re server itself isn’t actually experiencing that high load, it’s just an inflated average due to the timeout of the NFS mount.

And yes, you’re a whinny baby just like most of the ones who post on the DHStatus page. "Waaaaaaah! Where is my refund!? WAaaaaaah support sucks!"
I’ve been with DH since December 2006, I’ve gone through a few issues, including being moved to another server due to some issues with the one I was originally placed on. It’s expected to happen with shared hosting and I’m very happy to deal with support on it, who have ALWAYS (minus one occasion) been helpful and have resolved my problems as quickly as possible. I make sure to give them as much information as possible, I’m nice and courteous, and I thank them when things are working properly.
If you’re on a server that’s constantly having issues, then it’s time to ask them to move you somewhere else. Just complaining about it and demanding some form of refund, when you clearly never read the TOS when you signed up (which if you had, would explain their refund policy), clearly dictates you’re a whinny little baby who needs to take their business to a host that provides and SLA (Service License Agreement, which essentially says “We’ll pay you XX Cents per hour or such of downtime”). Don’t like being called a “whinny baby”? Then grow up, act responsible, and WORK with the people you’ve essentially employed to host your website. If they aren’t working with you and you’ve exhausted your abilities to work with them, then it’s definitely time to move somewhere else.


Chips N Cheese - Custom PHP installs and the like!


#7

First of all, show me where I ever demanded a refund. Second, the DH staff are of no help to me. When I get a response to my trouble ticket it has been non-personal responses that are cut and pasted and sent to everyone, many hours later. So you tell me how I can actually TALK to someone and I will jump at the chance. But instead my sites are down for hours and hours with no response or indication that they are working on it. I’ve been patient with them, but when my site is down for over 11 hours in two days I start to get upset. And it’s funny that you tell me to grow up when you’re the one going around name calling like a 10 year old calling people “whinny babies”. Horses whinny, moron.


#8

Depending upon your plan level, you can request a call back from the Control Panel->Support -> Contact Support screen. The very low grade plans don’t over this (and you shouldn’t expect it for that price!) :wink:

–rlparker


#9

And it’s also funny that you’ve been with Dreamhost for a whopping 6 months and have already had, as you say, many problems on two different servers, and that is perfectly acceptable to you.


#10

[quote]
And it’s funny that you tell me to grow up when you’re the one going around name calling like a 10 year old calling people “whinny babies”. Horses whinny, moron.[/quote]
lmao. Just proves my point further :wink:

And yes, from a shared host that’s clearly overselling as much as they are, I expect some bumps in the road. What I don’t expect, is my site to be down constantly, and it hasn’t at all. If I expected more, I would have never signed up with DreamHost, because I did my research in advance :slight_smile:


Chips N Cheese - Custom PHP installs and the like!


#11

How do you get this information about the server load? What is the shell command?


Happy Dreamhoster
Promo Codes:
$97 off: [color=#00CC00]TOBEMAX[/color]
2 free domains: [color=#00CC00]TODOMAINS[/color]


#12

w (“man w” for instructions) or, for even more info use top (“man top” for instructions)

–rlparker


#13

/usr/bin/top continuously updates load, memory and process statistics

(You’ll hear complaints about the load that the top process places on the system, but you’ll note that it doesn’t register in terms of memory or CPU)


#14

That said, please don’t just leave top running in a shell for extended periods of time. :wink:

You can always set a cron job to pipe the results of w to a text file at regular intervals if you want a representative idea of load over an extended period of time.

–rlparker