Web Hosting Talk







View Full Version : Nocster & Burst won't listen


mainarea
09-03-2002, 04:36 PM
I'm on a Nocster virtual plan, and server loads have been spiking all over the place (see http://www.ukee.com/status/stats.php ). Uptime for the previous month was below what they promised (I'll get a partial refund once everything else is sorted out). Live help (AIM) can only tell me my bandwidth usage sometimes, or else I have to email Matt (and for some reason, I never get an answer from him) - Cpanel never calculates the bandwidth.

I also was billed for the incorrect amount, contacted them about it, and haven't heard back since (it's less than $10, I can wait). Why does it seem like Burst just doesn't care anymore?

- Matt

edit: Server load is at 7.60 right now, it's usually at 1.50, I assume that isn't good...

Samuel
09-03-2002, 04:39 PM
Hrmm, what are you hoping to gain by posting here?

Take your business elsewhere....

Or... is this just a complaint thread?

Akash
09-03-2002, 04:40 PM
Originally posted by mainarea
I'm on a Nocster virtual plan, and server loads have been spiking all over the place (see http://www.ukee.com/status/stats.php ). Uptime for the previous month was below what they promised (I'll get a partial refund once everything else is sorted out). Live help (AIM) can only tell me my bandwidth usage sometimes, or else I have to email Matt (and for some reason, I never get an answer from him) - Cpanel never calculates the bandwidth.

I also was billed for the incorrect amount, contacted them about it, and haven't heard back since (it's less than $10, I can wait). Why does it seem like Burst just doesn't care anymore?

- Matt

edit: Server load is at 7.60 right now, it's usually at 1.50, I assume that isn't good...


while 7.6 isn't too terrible (i've heard of some in the hundreds at some hosts) it should still be low enough for you to hang in there while burst gets things sorted out...but you are right, i hear of more and more people saying that burst is getting too big, which is rather unfortunate :rolleyes:

mainarea
09-03-2002, 05:14 PM
7.6 isn't too terrible, but after the beginning of the month (when server loads soared after an customer ran many abusive scripts at once), I was promised by Burst that they would closely monitor the server load to make sure it didn't happen again.

For $25/month with 500MB space (which i will never use) and 50GB transfer (right now I use ~20GB), where else can I take my business? Besides, Burst has been relaible in the past.

Richard Ward
09-03-2002, 05:29 PM
That doesn't mean anything. I have two servers upstairs that have an average load of 25.00 and I'd only get complaints if my customers saw that and something like 1.50 as "normal," because they don't understand how the load numbers are calculated. Control panel software, such as Ensim, Cpanel, consumes a huge amount of CPU/MEM power with any real user base. While 25.00 is *NOT* good, you'll see no "real" speed effects on protocols from this (HTTP, HTTPS, SMTP, POP3, IMAP4, etc.) because of the load balancing and constant human monitoring for problem processes.

(Edited here:) I thought I'd throw in the uptime reply from one of those machines, just for the hell of it. -- 5:39PM up 154 days, 21:25, 5 users, load averages: 23.12, 23.19, 23.22

mainarea
09-03-2002, 05:33 PM
In the past, I've checked into the server load issues, and 99.9% of the time, it is another user or a runaway process (that they finally catch after 8 hours). Runaway processes have gotten to the point of shutting access to the machine down for almost 8 hours, without Burst noticing (they claim that they monitor their servers with virtual plans).

- Matt

Mango
09-03-2002, 05:35 PM
Originally posted by Richard Ward
That doesn't mean anything. I have two servers upstairs that have an average load of 25.00 and I'd only get complaints if my customers saw that and something like 1.50 as "normal," because they don't understand how the load numbers are calculated. Control panel software, such as Ensim, Cpanel, consumes a huge amount of CPU/MEM power with any real user base. While 25.00 is *NOT* good, you'll see no "real" speed effects on protocols from this (HTTP, HTTPS, SMTP, POP3, IMAP4, etc.) because of the load balancing and constant human monitoring for problem processes.

(Edited here:) I thought I'd throw in the uptime reply from one of those machines, just for the hell of it. -- 5:39PM up 154 days, 21:25, 5 users, load averages: 23.12, 23.19, 23.22

Very true.

It all depends on hardware specs and such.
Example : we have one server (AMD 1Ghz, 512MB of RAM, slow disk), which starts running mysqld very, very slowly once hitting 3-4.00. On the other side, we have a PIII 1.4Ghz with 1.5GB of RAM and a SCSI disk, which doesn't show any sign of slowness, even when running at a constant 20-30.00. A lot also depends on WHAT is causing the higher load. Is the server swapping intensively (>harddisk 'overload') ? Which processes are pushing the load (>processor 'overload') ? etc etc etc
It's all so relative and you need to know the server inside out before judging about it's load and consequences.

Just my 2 cents,
Carl

HRBrendan
09-03-2002, 07:51 PM
a p3 1.4 ghz with 1.5gb of ram running good with a load of 20-30 is an impressive feat.

-Brendan

UmBillyCord
09-03-2002, 07:57 PM
Originally posted by Richard Ward
That doesn't mean anything. I have two servers upstairs that have an average load of 25.00 and I'd only get complaints if my customers saw that and something like 1.50 as "normal," because they don't understand how the load numbers are calculated. Control panel software, such as Ensim, Cpanel, consumes a huge amount of CPU/MEM power with any real user base. While 25.00 is *NOT* good, you'll see no "real" speed effects on protocols from this (HTTP, HTTPS, SMTP, POP3, IMAP4, etc.) because of the load balancing and constant human monitoring for problem processes.

(Edited here:) I thought I'd throw in the uptime reply from one of those machines, just for the hell of it. -- 5:39PM up 154 days, 21:25, 5 users, load averages: 23.12, 23.19, 23.22

Sure it will run *most* things fine, but the issue is what happens when it spikes, can it handle it? I doubt it. I would rather run low and be ready for those spikes. Also, I would be willing to bet I could easily detect the difference between a .10 average server with the same specs and one running at 25.00. It maybe sutile, but thatis my middle name. ;)