Server Down?

Customer Support H-Sphere Shared Hosting
21

[QUOTE=brawney]
This is interesting because Invision Portal is a commercial software package. It is not freeware. If that is all you’re using on your site then we’re talking about an entirely different problem then. We’re not talking about home grown software that is poorly made.
[/QUOTE]

I believe the key here, in my case, is Resource Utilization. The software (again, in my case) is fine and has been running without issue for the past month with Jodo host, and for 8 months previous to that on another host, all fine. I won’t say any more until I give Yash and Support a chance to respond to all this.

I’m very frustrated though that our site visitors have been unable to reach us for 1 full day now, which could turn into many more if I have to rush out to try to setup shop with another host.

22

[QUOTE=Hawkeye]
TekniDude, so you have received the same letter and been asked to clear your files? Is your service being cancelled?
[/QUOTE]

?( No clue, I have no clue whats going on. ?(

I got this email before they said that they would look into it:

Dear Customer

We have identified two of your websites (*.com & teknidude.com) responsible for repeatedly crashing IIS. We were forced to deactivate them. We suggest you clear all your files off your account and put a simple “This site is facing technical difficulties”" page as your default. We’d then reactivate it as you search for a new host

Best Regards
Customer Support

and then I got a message saying they are looking into it.
It’s almost been 24 hours that my site has been down.

I just want to get a copy of all the files on my site just incase something happens.

-Jason

23

[QUOTE=Hawkeye]
This is RIDICULOUS. So, if your site is getting high traffic, which WE ARE PAYING FOR, we are in danger of being shut down?!!!

Here is the letter I got today:

"Dear Customer

We have identified your website responsible for repeatedly crashing IIS and high resource usage. We were forced to deactivate it. We suggest you clear all your files off your account and put a simple “This site is facing technical difficulties”" page as your default. We’d then reactivate it as you search for a new host

Best Regards
Customer Support"

There are a couple of problems here - first off, I have been paying for the TITANIUM plan since I became a subscriber (about 5 weeks ago). I can’t believe this is how you treat your TITANIUM customers in particular.

It stands to reason that TITANIUM customers will use more resources, whcih is why we are paying for these resources. You should have TITANIUM CUSTOMERS on servers which can handle the load we are paying for. It should also be noted that my site is using on ONLY HALF the bandwidth that is included with the TITANIUM plan! ! !

I find it very strange that my site has been running fine for all these weeks, then with no changes in the past two days, was suddenly identified as taking down the server. The traffic on my site has been consistant the ENTIRE TIME.

So, basically I think that I have been targeted because I was using more resources than other sites on that server (thus the reason I have paid for the TITANIUM plan). How else could my site be suddenly crashing the server when it has not changed and was not crashing the server 1 week or 3 weeks ago? I think another customer installed something on THEIR site which has been crashing the server and I (along with others?) were caught up in the sweep since my site has high utilization.

I think they have simply taken the sites which have the top utilization on this server and sent these letters to the owners. So, the message here seems to be, that it’s fine to purchase a plan with high limits (and pay for it), but just DON"T USE IT. X(

Changing web hosting services is a HUGE HASSLE and one I just went through early last month. I settled on Jodohost because I had heard GOOD THINGS about the service. I don’t WANT to switch or be forced out, especially since I don’t believe my site was responsible for this.

If anyone is interested, like many of you, my site uses the ASPInvision portal (MadDog’s portal) just like a LOT OF OTHER SITES HERE.

So, I would like the following questions answered:
1.) Why are TITANIUM customers, who can be expected to have a higher utilization, not on servers which can handle the load? And I’m only using HALF the alloted load for the TITANIUM plan and I’m being told that I’m using too much???
2.) With no changes to my site, how am I suddenly crashing the server now?

To say I am angry about this is an understatement.
[/QUOTE]

Win5 has been slow and unstable for the last 6 weeks. Detecting websites reponsible for high usage and crashing IIS is very difficult. We were finally able to isolate your website.

You were getting a high volume of daily traffic. You were running MadASP which runs probably a large number of SQL queries for each visitor. Your site was sustaining extremely high CPU usage. The CPU usage of the entire system has been going above 40% at times. At least 15% to 20% of this CPU usage we have been able to determine was from your site and database queries. Now that is 1/5th of the processing power on a DualP4 system. That is extremely high.

24

Yash, are you saying that the Mad Dog Portal (also known as MadASP, ASPInvision or just plain Invision Portal) can cause these problems because of the amount of database access it is doing? Drew says that his code is fully optimized for speed, so how can this be? I’m running the current released version of this app (v2.5) on one of my sites, but I’m just in testing mode right now. I was planning to use the new version (v3.0) when it was released. If you think this app can cause problems please let us know. I don’t want to lose my hosting account due to this app.

I suspect that any SQL Server app that is a message board (forums) is going to create a good amount of SQL Server traffic. That’s what SQL server is there for.

In any case, I don’t suspect my site would have much traffic even once it is ramped up and in full operation. :slight_smile:

25

You should be able to access your FTP folder fine. We’re willing to help you in any and all parts of your shift. To begin with, download your files and put a simple HTML file saying that you are temporarily down due to technical difficulties. We’d then reactive your site while you carry out the shift

Additionally, we’ll refund both of you for at least 2 month’s of service.

26

[QUOTE=brawney]
Yash, are you saying that the Mad Dog Portal (also known as MadASP, ASPInvision or just plain Invision Portal) can cause these problems because of the amount of database access it is doing? Drew says that his code is fully optimized for speed, so how can this be? I’m running the current released version of this app (v2.5) on one of my sites, but I’m just in testing mode right now. I was planning to use the new version (v3.0) when it was released. If you think this app can cause problems please let mek now. I don’t want to lose my hosting account due to this app.

In any case, I don’t suspect my site would have much traffic even once it is ramped up and in full operation. :slight_smile:
[/QUOTE]

No I’m saying that if you are running a high traffic MadASP portal, it generates alot of CPU usage. It just doesn’t belong on a shared environment. Semi-shared or VPS is the way to go.

27

if any of you have probnlem with FTP access, open a ticket

28

[QUOTE=Yash]
You should be able to access your FTP folder fine. We’re willing to help you in any and all parts of your shift. To begin with, download your files and put a simple HTML file saying that you are temporarily down due to technical difficulties. We’d then reactive your site while you carry out the shift

Additionally, we’ll refund both of you for at least 2 month’s of service.
[/QUOTE]

Once I ftp in:
I can see my folders. Ex. teknidude.com support.teknidude.com etc…
If I go into the foler the only thing in them is another folder called webalizer.

None of MY files are in their. Thats what I’ve been trying to say.
** OK I opened a ticket. **

I’m not sure how MadASP works but I’m guessing it uses MySQL or MSSQL.
Now arent the database servers seperate from the web servers?
If so then the real load/processing of the SQL wouldnt effect the web server. It would effect the mysql.jodohost.com right??

-Jason

29

[QUOTE=TekniDude]
Once I ftp in:
I can see my folders. Ex. teknidude.com support.teknidude.com etc…
If I go into the foler the only thing in them is another folder called webalizer.

None of MY files are in their. Thats what I’ve been trying to say.
** OK I opened a ticket. **

I’m not sure how MadASP works but I’m guessing it uses MySQL or MSSQL.
Now arent the database servers seperate from the web servers?
If so then the real load/processing of the SQL wouldnt effect the web server. It would effect the mysql.jodohost.com right??

-Jason
[/QUOTE]

mssql5 is on Win5.
Open a ticker, we’ll look into it right away

30

[QUOTE=Yash]
Win5 has been slow and unstable for the last 6 weeks. Detecting websites reponsible for high usage and crashing IIS is very difficult. We were finally able to isolate your website.

You were getting a high volume of daily traffic. You were running MadASP which runs probably a large number of SQL queries for each visitor. Your site was sustaining extremely high CPU usage. The CPU usage of the entire system has been going above 40% at times. At least 15% to 20% of this CPU usage we have been able to determine was from your site and database queries. Now that is 1/5th of the processing power on a DualP4 system. That is extremely high.
[/QUOTE]

I’m still confused - as a TITANIUM customer, it is expected that my site would have higher utilization…that’s what I pay for and why you have such a plan. With that in mind, why am I not on a more powerful server, or a server which has less sites total? You’re still basically saying that my site is being removed because it is busy, even though I am within the boundaries of the plan I have paid for.

Yash, removing my service isn’t the right thing to do and you know it. If I need to be moved to another server, that’s fine, and the sort of response I would expect from a professional service. Jodohost does NOT portray themselves as a discount Hosting provider who can only support small sites, but we still have a larger site here which is being penalized, why?

You advertise a TITANIUM plan because you have a place for sites of higher utilization and that is why I purchased such a plan.

Again, this is the chance, in a public forum, to do the right thing and avoid a bad public relations situation. I have also been trying to get my default page re-activated (with a message I have posted) so our visitors will have some idea of what is happening, but that hasn’t occured thus far.

31

[QUOTE=Hawkeye]
I’m still confused - as a TITANIUM customer, it is expected that my site would have higher utilization…that’s what I pay for and why you have such a plan. With that in mind, why am I not on a more powerful server, or a server which has less sites total? You’re still basically saying that my site is being removed because it is busy, even though I am within the boundaries of the plan I have paid for.

Yash, removing my service isn’t the right thing to do and you know it. If I need to be moved to another server, that’s fine, and the sort of response I would expect from a professional service. Jodohost does NOT portray themselves as a discount Hosting provider who can only support small sites, but we still have a larger site here which is being penalized, why?

You advertise a TITANIUM plan because you have a place for sites of higher utilization and that is why I purchased such a plan.

Again, this is the chance, in a public forum, to do the right thing and avoid a bad public relations situation. I have also been trying to get my default page re-activated (with a message I have posted) so our visitors will have some idea of what is happening, but that hasn’t occured thus far.
[/QUOTE]

Win5 has the least number of websites - 200.
We had to close it early because it was sustaining high usage at such low numbers.

As a titanium customer, you are entitled to get more bandwidth disk space and databases. But you are not entitled to utilized 1/5th of the processing power available on a machine. And this is sustained usage. You really don’t understand how high that sort of usage is.

32

[QUOTE=Yash]
Win5 has the least number of websites - 200.
We had to close it early because it was sustaining high usage at such low numbers.

As a titanium customer, you are entitled to get more bandwidth disk space and databases. But you are not entitled to utilized 1/5th of the processing power available on a machine. And this is sustained usage. You really don’t understand how high that sort of usage is.
[/QUOTE]

What bothers me is that you do not have that written on your services page. Although I am not one of those high usage customers, when I look at your Titanium package, I assume that I should be able to use the allotted resources without any problems - and that includes CPU usage. When you introduced SDU, this was the risk you took - people using lots of database queries, thus having high CPU usage.

33

I didn’t know that anybody put the SQL server on the same computer as the web server.

34

[QUOTE=brawney]
I didn’t know that anybody put the SQL server on the same computer as the web server.
[/QUOTE]

Agreed. I’ve never heard of that either.

35

[QUOTE=Yash]
Win5 has the least number of websites - 200.
We had to close it early because it was sustaining high usage at such low numbers.

As a titanium customer, you are entitled to get more bandwidth disk space and databases. But you are not entitled to utilized 1/5th of the processing power available on a machine. And this is sustained usage. You really don’t understand how high that sort of usage is.
[/QUOTE]

Well, maybe Titanium customers should also be put on more powerful servers.

Having my website pulled with NO NOTICE is causing me quite a bit of trouble. I may also add that my substitute default.asp page has STILL not been put in place as you said you would do in order to let our visitors know what is happening. I have it right there in my root directory…it is default.htm.

At this point I am looking at a MINUMUM of 1 week of complete outage as I attempt to switch hosts, which will probably extend to almost 2 weeks before things are ready, all with NO NOTICE to our visitors. This is unbelievable.

After the way this has happened, I wouldn’t recommend that ANYONE host a business website with Jodohost. I will also be recommending people running personal or hobby sites to NOT use this service either.

You have cancelled the accounts of two individuals today, without due notice, (an email saying I’m already disabled is a bit late) myself being one, causing us major interruptions in service. If we were effecting other users, we should have been moved to more isolated servers and allowed to continue operating while we transitioned out.

I can’t tell you how angered I am by all this. X(

36

Amazing! The number of quick replies from Yash is going down while the number of queries are increasing.

It seems that jodohost is currently having a severe capacity problem. No new domains, shutting down well utilised sites, etc. The sites I host with jodohost have limited traffic (Traffic since Oct 24, 2003 - 17 Nov 2003 is 292.1 MB out of 10.0 GB limit) so I’m possibly regarded as an ideal customer. However, I’m concerned at the jodohost website’s why us section at http://www.jodohost.com/whyus.htm under “superior hardware” it states that

“..We ensure that our servers are always running below capacity. Sites requiring high resources are shifted to less used servers on a regular basis by our staff…”

This obviously will be interpreted that you can handle capacity issues and would do what Hawkeye suggests move high capacity users like him to less used servers. I guess that the problem is that all your servers are up to capacity and thus don’t allow for those necessary shifts.

You also state under “superior service” that “We offer great services such as Microsoft SQL Server…” Also from the customer testimonials at http://www.jodohost.com/testimonials.htm it transpires that MS SQL Server is a major attraction point, meanwhile you seem to close down sites that interact frequently with those databases (especially discussion forums/portal sites) blaming your customers to have such interactive sites.

The latter is my major concern, it seems that as long as you have a static website (which also doesn’t have big files for download) you’ll be ok with any plan, as soon as you make a more dynamic site, especially with discussion forums, the servers do not seem up to scratch irrespective of package size, which as “Bliss” pointed out in this thread should be made clear in the package descriptions, ie “acceptable processor usage estimate”.

Although I’ve problems with backing up my SQL database on my “real” site (test site based on ASP Invision Portal beta 3.6) with more than 100 interactive users hosted by tacticalsystems.com I’m glad that in the end I didn’t transfer this account to jodohost, as I fear that once it would go “final” jodohost would close it down, being an interactive database driven website.

Finally I’ve advocated jodohost on the ASP Invision Forums (so has Bliss) and I recognise quite a number of names in this forum, including of course Drew Gauderman, ASP Invision CEO. I’m surprised that if indeed Asp Invision Portal was/is the problem that this does not seem to have been discussed with Drew, who since recently is member of the jodohost forums team. This affects two businesses, Jodohost and Asp Invision and more importantly the businesses of quite a number of customers of both suppliers.

The decent thing to do is either make transparent that you actually do not like to host dynamic, database driven, interactive sites (often using portal software PHP, ASP and ASP.NET portals) or that you try to enhance your capacity to deal with those sites and when problematic portals are identified try to resolve the problems with your customers first before sending them away. The latter is a very insulting activity as I can imagine that some of your customers might feel prosecuted like criminals while in reality they probably didn’t know they were “bringing the server down” as their bandwith and diskspace allowances were not met and no “cpu usage meter” is available.

In legal terms it seems that your customers can be found guilty, charged and sentenced without any warning, defence or probation. However excellent your customer service has been this is a serious flaw that you must rethink if you don’t want to jeopardise your business. I also sincerely hope that in the end jodohost proves to be a serious business and not a “scam”. I don’t like to use the word “scam”, but I increasingly feel that customers were drawn into competitive packages (which changed, but still are competitive) that ultimately can’t be serviced to their demand. In PR terms this is negatively associated with “overpromise but underdelivery, and all spin but no substance”.

Please show us that we can trust you to care for our business as well by not immediately closing down sites, but by contacting the customer and discuss a remedy (ie not sending them away, but constructive advise). Taking away the possibility to run discussion forums should be clearly stated in your packages (ie we don’t allow portal sites on our servers).

37

[QUOTE=brawney]
I didn’t know that anybody put the SQL server on the same computer as the web server.
[/QUOTE]

Incorrect. Companies like GearHost.com put ms sql on the same server (at least that is what Drew from MadASP told me)

38

[QUOTE=Bliss]
What bothers me is that you do not have that written on your services page. Although I am not one of those high usage customers, when I look at your Titanium package, I assume that I should be able to use the allotted resources without any problems - and that includes CPU usage. When you introduced SDU, this was the risk you took - people using lots of database queries, thus having high CPU usage.
[/QUOTE]

Customers can use upto 5% or 10% of the CPU but 20% over a sustained period of time is down-right harmful for the server and other customers on it. It doesn’t matter what sort of load is on that server.

39

  1. We do move sites between servers. But a site using 20% on a Dual P4 system simply has no place on our servers. That is simply too high for shared hosting.

  2. There are plenty of windows hosting companies that put 500 or 800 domains on each server. We have restricted the number of sites to the minimum we can support profitably - 200 to 300. We take a good amount of time to recover software costs on each server. You may have noticed much faster execution times, speeds and uptime with us. We have only been able to achieve that by being so proactive on these sort of issues. Win5 crashed 12 times. That is a serious issue for us and I hope customers understand that.

20% usage means that the site is taking 1/5th of the resources. That means we could only put 5 such sites on one server. If the customer is assuming titanium customers can consume that much, it means 5 titanium customer or $150/month gross for us on that server (our monthly costs are more than that) Even so, usage would go up to 100% at times. Result - the server crashes, execution times are extremely slow,etc.

Such a site simply doesn’t belong on a shared evironment. It takes ALOT of traffic and simultaneous SQL queries to push usage up that high. Customers shouldn’t worry about this. I can’t comment on exact transfer figures for any customer but I went over this guy’s stats and I was amazed to see how many hits he had sustained, that too on a website that depends entirely on MS SQL.

What we are doing isn’t uncommon in the hosting industry. I know someone who runs a unix hosting business. He had to shut down some forums for high usage. And I’m talking about UNIX here.

We will be revising our current termination policy. Our cluster will be up very soon and we’ll be working on more effective ways to track customer resource usage and warn them before finally disabling their site. Right now however we had no choice. We just had to disable these sites to keep that server up.

40

BTW, on the new cluster, we’ll be having stand-alone ms sql servers. That should put some load off the web servers. Our capacity problems will also be solved with that.

site’s for teknidude and hawkeye were reactivated as they search for a new host.