Bandwidthd totally different than SolusVM?

[coreyman]

I don't remember who but someone told me about bandwidthd and said it would be great for monitoring traffic. Several people have told me 'solusvm is inaccurate'. So I installed bandwidthd and I've been running it for a while. Today I decided to look at the numbers and compare top traffic with solusvm.

First thing I noticed about bandwidthd, there are 4 index files all dated differently but they all have the same data.

I decided to use the 4th index file dated Apr 17th. I looked at top 20 for this month and found them in solusvm. Some people transferred less than what it says in bandwidthd in solusvm, but some people transferred WAY more (over 250GB of traffic unaccounted for in bandwidthd)....

Has anyone else here used bandwidthd and can shed some light on this situation?

 

[SkylarM]

Bits vs Bytes perhaps? Are you sure both solus and bandwidthd are tracking in the same format?

 

[WebSearchingPro]

First thing I noticed about bandwidthd, there are 4 index files all dated differently but they all have the same data.

One is 2 day one is 8 day one is 40 day and the last is 400 days. Since I assume you just set it up, you would see the same data in those files.

 

[KuJoe]

We've always found SolusVM bandwidth tracking to be inaccurate. We started using our own bandwidth tracking software early last year so we had something accurate to show our clients when they opened a ticket complaining about the bandwidth total not equaling what their monitor showed. Oddly enough our new control panel uses the same method SolusVM uses (a different method than our bandwidth tracking software) and the bandwidth usage actually matches up between both of our software. I'm not sure what SolusVM is doing differently since I can't see their code but it doesn't make sense why they would be so different.

 

[coreyman]

We've always found SolusVM bandwidth tracking to be inaccurate. We started using our own bandwidth tracking software early last year so we had something accurate to show our clients when they opened a ticket complaining about the bandwidth total not equaling what their monitor showed. Oddly enough our new control panel uses the same method SolusVM uses (a different method than our bandwidth tracking software) and the bandwidth usage actually matches up between both of our software. I'm not sure what SolusVM is doing differently since I can't see their code but it doesn't make sense why they would be so different.

You're right it doesn't make any sense whatsoever so I'm pretty confused.

 

[Damian]

What methods are you guys using? The way i've done it is to use an iptables chain as a counter, then clear it every time I read from it.

 

[Francisco]

What methods are you guys using? The way i've done it is to use an iptables chain as a counter, then clear it every time I read from it.

This should work fine.

Solus always had derpy accounting back when we used it. A VPS flooding in/out would cause their counters

to roll over and the flood wouldn't get counted against the customer.

There's been MANY stories of bw accounting failing in Solus and hosts getting dicked for 10's of TB's in overages that should've never happened.

Francisco

 

[KuJoe]

We use 2 methods but only the iptables counting matters when it comes to client's VPS accounting.