Serious problems on reseller hosting in eleven2.com

Portal Home > Knowledgebase > Articles Database > Serious problems on reseller hosting in eleven2.com

Posted by B_Dark, 01-15-2012, 05:40 PM
Hello. I thought that I should share my issue with you to see if I could finally find a solution. I have a reseller hosting(5 sites: 1 wp single site active in Internet, 1 site with one simple static html page and a wp multi site with 3 sites under constructions no active i Internet) for 3-4 months with eleven2. I was for 3 years a client for SharkSpace that the last few months it has merged with the company above. When it merged with eleven2 I started to have the most serious problems that I ever had. Some time before Christmas (12/03/2011) without any reason one of my wp sites was ruined. All of the photos and videos from the items, one and half's year work was lost.After they troubled me for a week (12/09/2011) without regaining my data base, sayint to me at first that their backup was crashed and after that they told me that they can't find my data because a part of my base is innodb and not myisam and that's they couldn't trace it. After I was exposed to my clients and advertisers and lost money in a such a hard period to make money, I let it pass and me and another person made from scratch everything that it was destroyed in a very small time period. At the same time (12/23/2011) I notice that the difference that I payed (11/01/2011) to go to the package R-100 in a european server. And for more than a month I wasn't moved to european server and if haven't told them that am pretty sure that they wouldn't notice it. And after they moved me to the european server((To have ping because I am in Europe and I address mostly to European visitors), three days after I said that I started to have again BIG problems in my account. I always had errors 404, 500, 503 and fatal errors out of memory... Is 1/15/2012 and still they are unsolved!!! I have sent countless of tickets, I have talked with them a load of times in the chat without any result. From the side they were telling that they had a problem with the european server in london and I should be patient and on the other side they were telling on the same on the chat that an associate of theirs that the problems I had in my account are because of the big traffic of the visitors and if I want to see results I have to move to a cloud server. Please tell me if the same you talk on the chat with an associate of theirs and he says to you that admin informed him that if I want to resolve my proms I have to move to a cloud server and at the same the same or another admin replies to my ticket that is their problem and they will fix it and I should be patient, then what impression it makes to you? Wouldn't you think or that they are amateurs and have a very bad communication between them and give a very bad image to their clients in result they will believe that they aren't professionals and not to be trusted or all of these "sudden" problems have happened in purpose to make pay more money to them. Also they told me, because I all of my sites in one account are the cause of my problem, because of the limits they have and to resolve the problem they told me to separate the sites to different accounts or to move to a cloud server. So I ask him that if he guarantees that if I move my sites to different accounts I won't have any problems. And he replied that he can't guarantee for that(THEN WHY HE TOLD ME?) but he can if I move to a cloud server(SO HE IS SURE IF I PAY THEM MORE MONEY). So I decided to ask them what kind of setup the server has, how many clients they have on it and what limits I have. And their associate told me these on the chat: Intel(R) Xeon(R) CPU E5506 @ 2.13GHz/24GB ddr3 CloudLinux Server release 5.7 64 bit/litespeed webserver And overall are 800 clients and my limits are 1gb of memory and 25 concurrent connections At the link below of their site http://www.eleven2.com/web-hosting-technology/ say that the characteristic of a reseller account are these and it has nothing to do with the things they told me. Maybe they were upgraded Web Hosting & Reseller Hosting Our most popular hosting plan is packed with the top of the line servers featuring the following technologies: Intel or AMD Dual Quad Core Servers 12GB DDR3 Memory RAID 10 Hard Drive Arrays 1000mbit Network Uplinks Dual-Redundant Backups (we have 2 different backup methods). 24x7 Proactive Monitoring 24x7 Live Technical Support CentOS 5.5 x86_64 Also, my memory in my Cpanel is 2GB and not 1GB and my Entry Processes is 30 and not 25... http://img580.imageshack.us/img580/4391/statsda.jpg Aren't different? Am I being wrong somewhere? See my Resource Usage on cpanel and tell me your opinion: http://img718.imageshack.us/img718/6375/part1nvx.jpg http://img716.imageshack.us/img716/1551/part2bd.jpg I can understand that my site is getting evolved and it has more requirements now, so more or less I must my hand in my pocket, but it doesn't even let me activate or deactivate or save some settings of a plugin(trying to do some optimization) .. I see that this moment I have 4 visitors(including me) at the and 4 bots and when I try to any of the above I get this: 503 Service Unavailable The server is temporarily busy, try again later! Powered By LiteSpeed Web Server LiteSpeed Technologies is not responsible for administration and contents of this web site! This isn't logical. I am not writing this message on the forum because of some hate on the company. But am out of my limits and am disappointed for getting to the point for having to make my issue with this company public so I will get any kind of results. I wish to be wrong and have just wrong impression... but all the above show the opposite.
Posted by Eleven2 Hosting, 01-15-2012, 06:01 PM
This is clearly due to you being over your CPU usage as shown in the screen shots you uploaded. Please send me your ticket ID to rodney (at) eleven2 dot com and I will gladly check this out for you. We do have several customers who were severely abusing our servers and once we put some very very conservative settings in place (with Litespeed and Cloudlinux), it 'caged' those customers and fixed a majority of problems that were affecting other customers.
Posted by B_Dark, 01-15-2012, 06:17 PM
For that reason i ask to remove all my web stat programs (except Bandwidth, Error Log and Resource Usage) from my reseller account, so i don't have a lots of tracking programs that use my cpu and work only with google analytics. I sent you a mail with my ticket id, i hope very soon we found a solution in my problem..
Posted by lastminutehosting, 01-15-2012, 06:26 PM
That would be a good reason because of high CPU usage I think it may be time to go dedicated if you need a lot of CPU.
Posted by Forward Web, 01-15-2012, 10:09 PM
Sounds like a mess, approximately how many unique visitors are you getting per day? If you are maxing out your allotted CPU usage, its probably due to plugins you have installed on WP (it happens all of the time).
Posted by leckley, 01-15-2012, 10:23 PM
The stats should be running nightly under root, so they should not be causing any load to be shown under CloudLinux. If you are hitting your CPU limit then you have a script causing you to use all your CPU.
Posted by Mad_matt, 01-15-2012, 10:43 PM
you say you lost 1 1/2 years of work on one site? you dont believe in doing the occasional backup? you dont have to even backup often just after major upgrades. seriously no excuse for not having at least 1 backup done in 18months.
Posted by rahulkapoor, 01-16-2012, 08:12 AM
absolutely, there is no reason for not backing up your data on your own. and, not backing it up for more than an year is a cardinal sin. all the best getting your data back now.
Posted by B_Dark, 01-16-2012, 12:16 PM
I didn't do any major upgrade in this site. It was destroyed without any reason, all of this happened in one night.
Posted by SpaceStation, 01-16-2012, 12:52 PM
I've bought the r-100 account with eleven2 some time ago to host our websites off our network to be available to our customers if something goes wrong. Everything was working fine for the few first days but somebody was clearly abusing the service as on a 8 core server they had loads of 6-5-7 with a few moments of the server being completely unreachable with whm and whmcs timing out. I have sent a ticket and got a reply that really some things were change and that everything should be back to normal soon. Well the sites really did become more stable but seems to me that the abusers are not completely contained as I did experience some hiccups today and yesterday. The thing is pinging the servers gives REALLY low times but sites with php(including whmcs) are SLOW. I won't be bothering them again as my sites are just being redesigned so the presence isn't crucial. But would be nice if it was completely sorted.
Posted by SpaceStation, 01-16-2012, 01:04 PM
Just as I was writing this the server became completely unavailable and the requests are timing out. Load spiked to 8.09 I know how it is to have a user or two ruining the experience for everyone else, we had it happen to us on several occasions, the abusers would simply be suspended/offered a vps/dedi or given a time period to find a new webhost...
Posted by BreakFree, 01-16-2012, 01:32 PM
Honestly, it's pretty common to blame WordPress plug-in's without really looking more into the issue before making the call. I've see installations with 20-30 plug-in's that barely use more than 64-128MB's of RAM, which would be clearly within his allowance. In such a case, I would look to see if there was something that could be done to reduce CPU usage, since that's the cited issue. Perhaps enable caching (W3 Total Cache) and off-load the images to a CDN (as last I recall, Eleven2 does offer one, or used to). Serving static, cached pages to those not logged in would greatly reduce CPU usage, as well as RAM consumption. Off-loading images, JS, and CSS to a CDN would reduce HTTPD requests on the server level, and increase performance. The less HTTP requests you're generating, the more RAM that is available for use. Beyond that, to expand on the caching, see if there is some sort of OP-Code caching available to use with W3 Total Cache, such as APC or eAccelerator (even xCache). Combine this with Minification of files, page caching, GZIP compression (which is available though CPanel's intergace), and setup the expire times for files. That's just a few suggestions. I don't know how much traffic you're generating, but it'd be easy to cut the resource usage in half using the above. I know as I work with WordPress on a daily basis, beyond that, I work on optimizing WordPress websites, so I know it works. It's just a matter of implementing it, or getting someone to help you do it.
Posted by SpaceStation, 01-17-2012, 11:44 AM
UPDATE: To be fair to the provider in question, everything is back to normal. Server is as fast as the first day and whm is more than responsive. Good job
Posted by B_Dark, 01-17-2012, 12:22 PM
Rodney sent me a message : "We will be tweaking a number of things here on our cloudlinux install before adding this back in including an extensive blog post explaining what is being throttled, etc, Your account is constantly going over, so once the limits are back in place, you will most likely have issues again as we are not changing the limits. I would recommend splitting all your addon domains up so that they are running on individual accounts, rather than inside of 1 cpanel account as you are setup now, that is the whole purpose of a reseller account." I did this and when I activated it it kept having 503 lightspeed error every time I tried to save draft or publish a wordpress post. Also the same happened when I added a video in "external media" and when I was trying to upload a photo I got an "http error" without showing me the progress bar of the uploading process. Even though I had an "http error" and it showed me the progress bar the photo was uploaded but it didn't let me to make a "set as featured image". All of these happened when I activated the new account with no visitors at the time. It is not logical. I sent a message to Rodney and told him that it looked like they blocked some "action/progress scripts" in my account, I wish to be wrong, Right now I don't have any problem. Also I asked from the tech support to deactivate many of the web stats programs that they have activated on the Cpanel because one big reason that I am reaching the limits of the CPU, me and surely many other of their clients, is because of those programs that are tracking our sites are using a big part of our CPU to do this process. They sent me this message: "Hello, We cannot deactivate the following stats programs as the custom packages is not allowed in the shared server. I suggest you to move to cloud server so that you can make the changes there. You can check the details from the link http://www.eleven2.com/cloud-servers/" Everybody now are using Google analytics, I believe that is very unnecessary to have activated on accounts all of these programs. They can keep awstats that is most common thing, all the rest are not needed. Also I want you to suggest a monitoring program to see analytically what is holding back my CPU and my memory.
Posted by B_Dark, 01-17-2012, 12:29 PM
I have activated on my site all of this time that having those problems the plugin W3 Total Cache and I have followed the guide of http://gtmetrix.com/. APC and eAccelerator is deactivate of the option of W3 Total Cache, only disk basic and disk enhanced is enable. I will activate again Cloudflare because the first time it didn't worked well and we can see. a.v visitors per day 150-280 4k-6,5k per month visitors Last edited by B_Dark; 01-17-2012 at 12:38 PM. Reason: i forgot to mention something
Posted by Eleven2 Hosting, 01-17-2012, 01:18 PM
I had a really long and great reply here a hour or so ago and accidentally hit the back button on my computer. Opps So you are getting the short response here from me. Yes, we are running limits in place, this will upset the 1% of our clients whom are using 99% of the server resources (CPU, Memory and IO), we are going to embrace that. The other 99% of our customers (most of which are silent and not going around saying 'wow things are so much faster and better') are going to love this as we are ultimately eliminating the 'problem' accounts on our shared servers. We are working with each of those customers to make sure they are on the proper hosting setup for their site. We are in the process of rolling out CloudLinux and Litespeed to all servers, this is going to be a multi-month project as we have so many servers, but this is moving along quite well. We did encounter a issue when not using the new CL 6 Kernel and Litespeed, we are working with CL to get this resolved and should have it shortly, but this problem that was giving the 503 errors will not reoccur now until this is resolved. We have numerous mechanism in place with Nagios to ensure things are functioning properly with all the configurations working between CL, LS and cPanel. We are excited about this upgrade and additional as it is ultimately going to give all of our customers better uptime, response time, etc. We are not spending thousands and thousands and probably 10's of tahousands more dollars per month for no reason when this is all completed, we know our customers will love this and the best part of it all, we are doing all this without raising costs for anyone. I will be posting a very very detailed blog post on our site shortly outlining allot of the details of this change and what is accounted for and what is not. Mail and Stats processing is NOT accounted for in CloudLinux, that WILL NOT count against your resources. I wanted to make that clear as there has been allot of missinformation and non-factual information in this thread. As always, I am always available via e-mail (rodney at eleven2 dot com) shall you have any questions about anything.
Posted by B_Dark, 01-17-2012, 02:29 PM
Am glad for your upgrade and am looking forward to experience it as your client. I understand that we must show patience from our side but you must understand that to me and many others our work is going backwards instead of going forward with all of these problems.