Godaddy VM hosting and TCPIP problem

Markfre1

We are hosting AwareIM initially at GoDaddy on a Virtual Server. Quite often it seems that the TC PIP connection is lost because AwareIM is running but nothing works because the data connection is lost as well. We have to shutdown AwareIM and restart it and its up and running fine.

Sometimes it will go days with no problem and sometimes a couple of times a day. AwareIM says its the virtual environment which I agree and should go to a dedicated server. We do not want that extra expense yet.

Has anyone experienced this and resolved it? Or is anyone running in a Virtual environment and does not have any problems and can recommend a service?

Thanks for any feedback.

Mark

nlarson

yes, I had the exact same thing, and I believe it has to do with their VPS infrastructure and hw it handles resource re-balancing. Aware helped me as much as they could, but it's not really an Aware issue. The bottom line is that godaddy VPS sucks.

My solution was to move to a Dedicated server (I chose Codero), it cost me a little more but not too much.... never a hiccup since.

nlarson

just checked in on Codero's current offerings, you can get a dedicated physical box for as little as $70/month; it's dated, but will run Aware just fine.

For what is worth, they have been rockstar's for the $150 I am paying them.

Markfre1

Yes, AwareIM made me aware the it was a VPS problem but it's good to hear that confirmation. It's unusable.

Do you think another VPS might work differently? One of the other reasons we wanted VPS was the easy scalability.

It's kind of a hassle to deactivate and reactivate licenses so I would like to make this move once.

I will check out thsi Cordero as well. So you never go down? Since we are new to AwareIM, we weren't 100% sure of the stability but it is such a well put together product, we were pretty confident it was not their problem.

On a dedicated, can we expect the system to pretty much stay up?

Thanks in advance for your response.

Mark

nlarson

yes, should just stay up. Stability is not a short coming in this product.

RLJB

For what it is worth:

we have co-located our own servers in Australian data centres, 100% stable , Aware is rock solid
we have used and are using Storm on Demand cloud servers in US, it's more recent but so far it's also 100% stable with Aware

We did a pretty big search from Aus to US to choose our cloud IAAS provider. Went with Storm due to a) cost - it's good value b) they scale big (you can get 96gb RAM accounts) c) their self management is great (you can resize, backup images, clone etc) d) the support is good.

One thing... I'm not 100% certain yet if latency is an issue, something in the back of my mind tells me our server should be in Aus for better speed (since all our users are here) but all the cloud providers in Aus are 200% more expensive

Markfre1

Thanks for all the feedback. We are going to install on another vps maybe storm an evaluation copy and let it sit and see if it stays connected. Then if it's stable, transfer the license.

Thanks again for the feedback.

Mark
Us

GaryBartlett

Hi Rod,

Are you still using Storm on Demand? If so, what package are you using and do you have an idea on how many users/logins you're running on it?

Thanks very much - I'm new to Aware IM and trying to get an angle on the various costs associated with moving our sites and services on to it.

Regards

Gary

RLJB wrote
For what it is worth:

we have co-located our own servers in Australian data centres, 100% stable , Aware is rock solid

we have used and are using Storm on Demand cloud servers in US, it's more recent but so far it's also 100% stable with Aware

We did a pretty big search from Aus to US to choose our cloud IAAS provider. Went with Storm due to a) cost - it's good value b) they scale big (you can get 96gb RAM accounts) c) their self management is great (you can resize, backup images, clone etc) d) the support is good.

One thing... I'm not 100% certain yet if latency is an issue, something in the back of my mind tells me our server should be in Aus for better speed (since all our users are here) but all the cloud providers in Aus are 200% more expensive

RLJB

Hi Gary

Yes - we are still with them, and they are still great (the best of the cloud providers we have tried, very good support, well priced, and very stable with Aware).

BUT latency has proven to be an issue (most of our customers are in Aus), so we are planning on moving to Aus based provider (we use Zettagrid here), which has been good and bad. I think our inital set up with VMWare wasn't right, so it was unstable, but we have that fixed now and it is very fast. I think their support is slow and they are lacking some features that Storm provides (can't clone an instance) but the speed is very good.

We over provision our production servers, so we always deploy with 8GB of RAM, user numbers vary greatly - so it isn't really a meaure as they don't/aren't all logged in at the same time (varies wildly depending on bus space).

Hope that helps.

GaryBartlett

Thanks, Rod!

rocketman

My two penneth. I'm with 1and1 in Europe, have a dedicated located in Germany with all my users in the Uk. Very reliable. I used to have an old style vps which exhibited the same crash symptoms as everyone else reports but since moving to the dedicated I've not has an hour's downtime in over 12 months (touch wood).

Later cloud servers are ok since they handle the resources differently and I've been told that if a cloud server goes down, the ISP can switch to new hardware which then takes on the complete persona of your original server which in theory should overcome the AwareIM licensing issue since AwareIM will still think its the machine it was originally licensed to.

It would be helpful to me if anyone could confirm this as I think I will need to make my next jump soon and like the idea of the scalable nature of cloud servers.

On a different topic - maybe this should be a separate post - has anyone hosted a MySQL database on Amazon S3 space whilst running AwareIM on another ISP server. I do like their 99.999999999% uptime SLA.

RLJB

Aws ec2 just launched in aus last November. We're testing now and will report back. So far, very impressive. Good prices. MySQL sits on an EBS volume not a separate server. The ebs is elastic and can be snapshot backed up in a few seconds. The backs ups reside on S3. It's stable and fast so far.