Rackspace Alternative

Portal Home > Knowledgebase > Articles Database > Rackspace Alternative

Posted by ibejohn818, 06-22-2009, 03:11 PM
Hello All! Just recently experienced my second power failure with rackspace. Power failure resulted in my MySQL database servers (Linux) hard resetting indicating the lack of a UPS. (4 other servers also hard reseted) Has happened twice in the past 9 days with no attempt to move my rack, or correct the problem. I'm really pissed this time, because it resulted in corrupted data on my Master MySQL server and is causing a whole bunch of problem in my application. The Alternative networks that I am looking at are -Media Temple -Liquid Web I would need fully managed support as I had at Rackspace. My setup is pretty complex, consisting of 4 load balanced clusters (5 Win3k boxes per cluster ) 5 Linux DB's on Redhat Cluster all load balanced with a hot-failover setup and all approriate networking switches and firewalls. Are there any rackspace converts that have gone with Media Temple or Liquid. How would you rate them compared to Rackspace? It just seems that my account is not big enough for Rackspace to care about really supporting my hardware.(Even though is costing $$$$) I mean, my machines hard reset twice and they still haven't put me in a new cabinet or moved me to a different power block. Thanks for your feedback!
Posted by SSD-Greg, 06-22-2009, 03:45 PM
Try them out for another time if it happens ask for another rack or a diffrent power block.
Posted by cristibighea, 06-22-2009, 03:55 PM
Make some demands, like ToNeZ recommended, as that will probably make them do something about it, since many companies only start caring when you start complaining too much . I don't think MT does Windows management, and LiquidWeb requires Plesk from what I remember, but be sure to ask them if you do plan to move in the end.
Posted by HostingASPNet, 06-22-2009, 03:58 PM
Hello, Yes, is will be much better to talk with them. Regards
Posted by ibejohn818, 06-22-2009, 04:00 PM
I have complained, multiple times. That's what really sucks about this whole situation. I love their personnel, it's just that my account does not move the needle over there. Don't get me wrong, I'm spending in 5 digit range a month (6 digits yearly) for my hosting. (Can't say how much due to NDA) It seems that if me being mad does not affect their standing on the NASDAQ, they could care less how much I complain. I just don't want to sit around and have another power issue make my MySQL boxes crash. I mean, im sure there dell desktops have UPS's to ensure their damn twitter messages don't get lost mid-post!!!!!
Posted by DATARTIM, 06-22-2009, 04:13 PM
Have seen these types of issues more and more with them, albeit more in the UK. If your spending five figures you should expect them to properly manage your systems and setup. I'm not sure if either of the ones you have chosen are a direct replacement for RS though, liquidweb maybe but MT have a slightly different want of doing things.
Posted by AquariusStorage, 06-22-2009, 04:20 PM
RackSpaces major competition in the "premium pricing/premium service field" is Liquidweb. The nice part about Liquidweb, is that they are a little cheaper then RackSpace yet you get the same great service Check them out.
Posted by Scott.Mc, 06-22-2009, 06:06 PM
I wouldn't say that liquidweb are any real competition to rackspace, not that rackspace are *that* good but they are superior to liquidweb in terms of network and support. Liquidwebs ticket responses are quite slow compared to their phone support. As system administrators generally when we are reviewing those companies it's generally always after a new customer has become frustrated with them and I would say if I had to choose between them both that rackspaces support knowledge is generally better and generally you will be escalated to someone who actually knows what they are talking about easier than with liquidweb. For example the vast majority of customers we look at generally are frustrated because of downtime caused by application level issues and the common response from reading their previous tickets with liquidweb is either buy more hardware, it's a ddos, "needs a configuration change" and things to this effect whereas rackspace will generally after a bit escalate to someone that actually understands what is going on and indicate at what the issue is (weather they assist in resolving it or not is another matter but they don't necessarily have to offer any assistance with it). Now weather the above means anything to you or not is another matter, just thought I'd share it. You might want to have alook at datapipe/inetu instead. As an interesting note, received this from rackspace about 1-2 hours ago which is probably in relation to what you have been experiencing,
Posted by ibejohn818, 06-22-2009, 06:28 PM
Yes, They've emailed, read, mailed me that statement. I wouldn't be this concerned if my front end webservers just crashed, but my MASTER mysql box failed along with all my slaves and my hot failover. That's dooms day crash. I'm lucky that my whole transactional log file didn't corrupt. Even after that, we found short comings in the MySQL config that was preformed by their DBA's. My log files for my slaves were being saved/referenced from the tmp directory, thus when rebooting from hard reset mysql couldn't even re-start and took a level of trouble shooting. Disaster! I understand that everyone has issues, but the same issue in 9 days, and doing nothing to correct this. They could've have just plugged my MySQL boxes into a separate rack mounted UPS to give me some assurance. It just makes me and my company look like idiots. And if I stay on for a third power debacle, then the only idiot is myself.
Posted by Scott.Mc, 06-22-2009, 06:31 PM
Agreed pretty much their mySQL configuration is always the same and they simply change the myisam key buffer to the size of the indexes that is pretty much their entire optimization, which may be fine for the majority average setups but not for anything that actually requires alot of mySQL activity. Not sure I understand this, first as to why the temporary directory would be used but during a reboot they shouldn't be cleared out unless I am missing some part of the setup. You are correct, it's clearly a fundamental design flaw that your slaves are setup on the same power source.
Posted by LiquidWebTravis, 06-22-2009, 06:40 PM
You are certainly entitled to your opinion but our help desk ticket response times average between 5-10 minutes. I'm not sure that can really be described as "quite slow". Perhaps you used us long ago?
Posted by Scott.Mc, 06-22-2009, 06:44 PM
Nonsense. In fact there is a thread from just the other day, right here on WHT which shows pretty much the exact same type of thing I expressed as my overall experience with liquidweb: http://www.webhostingtalk.com/showthread.php?p=6232469
Posted by ibejohn818, 06-22-2009, 06:45 PM
Travis. I assume you work at Liquid. PM me the contact information for a sales rep so I can email my configuration and get a FRIENDLY quote. thanks!
Posted by LiquidWebTravis, 06-22-2009, 06:50 PM
Scott.Mc. As you will notice I said "average" and the customer you are talking about had 1 ticket over 30 minutes. That's 1 ticket of the 10,000+ we handle in a month. We have a 30 minute response time SLA. If you have a personal experience that you are talking about I would be glad to look into your tickets or accounts and provide SLA credits if you had delayed response times.
Posted by ibejohn818, 06-22-2009, 06:51 PM
My replication logs files were being saved/ref'ed from the tmp var directory. Every time the machine reboots the tmp dir's contents are deleted. Making mysql hang on trying to start-up. I went with Rackspace so didn't have to worry about that stuff.
Posted by Scott.Mc, 06-22-2009, 06:57 PM
I don't really want to keep de-railing the OP's thread but I don't doubt you answer the majority of the tickets in the time you specified however the vast majority of tickets I get to review the responses are always slow, bare in mind the only time I will be reviewing a ticket at most datacenters is if the customer is unhappy there. The other part I highlighted was the general response when there actually is an issue and I feel the general level of support during an issue is poor but I will not continue on with that, the OP is free to search these very forums and in fact just search these forums for liquidweb ddos and similar keywords and just compare the number of times when a customer has been having consistent issues with their system which is accounted down to a ddos/flood when a fair number of them are probably are not.
Posted by ibejohn818, 06-22-2009, 07:02 PM
Well this is very insightful. Maybe it's back to the One Whilshire Colo and super uber nerd admins to run my stuff. ugh
Posted by Scott.Mc, 06-22-2009, 07:03 PM
Which OS is this? Since I am struggling to understand the relationship between replication not running and mySQL not starting or the tmp partitions being cleared during a reboot. You mentioned above your databases are RHEL so I assume you mean /tmp which should not be getting cleared during a reboot (by default) although it doesn't take away the fact they should not be using the temporary directory for the binary/relay logs in the first place, I am just curious as to what the actual issue was.
Posted by LiquidWebTravis, 06-22-2009, 07:24 PM
Does this speculation have a factual basis?
Posted by ibejohn818, 06-23-2009, 01:59 PM
Are there no other alternatives? I'm gonna get a quote from liquid, but it seems as if their tech-support is lacking in logistics and organization. Has anybody use the Planet's Managed solutions?
Posted by Scott.Mc, 06-23-2009, 02:02 PM
Datapipe or inetu are worth considering, as for ThePlanet's while I am not a huge fan of ThePlanet I do wonder what their new managed line is like as they have many, many great techs working for them and if they live up to the marketing they might be a good choice, so if you do use them I would be interested to hear how it works out for you.
Posted by ibejohn818, 06-23-2009, 02:11 PM
I'm getting a quote right now. I'll update this thread with my first impressions of both Liquid and The Planet (Which i guess is EV1)
Posted by mainpipe, 06-23-2009, 02:29 PM
I don't know if I would put LiquidWeb in the same category as RackSpace. I'm not saying LiquidWeb is bad... but just not in the same league. Take a look at datapipe.net. They are in the same premium category.
Posted by 01globalnet, 06-23-2009, 05:43 PM
You can also check Cartika - they can support such setups.
Posted by lockbull, 06-24-2009, 01:31 AM
Based on some RFPs and followup discussions, LogicWorks and Planet Northstar have pretty competent MySQL DBAs. They are definitely more at the Rackspace price point (probably higher actually) than at the LiquidWeb price point.
Posted by neuman, 06-25-2009, 04:32 AM
WiredTree hands down is the best to go with. I stay away from liquidweb due to an issue I had with them a few years ago and they did nothing to fix it. wiredtree has went out of there way to correct problems for myself.
Posted by Kaumil, 06-25-2009, 05:30 PM
Hello, I personally do not have any experience with MediaTemple or LiquidWeb, but I do with Singlehop. I used them a while back for servers that were needed by Vistapages (my old hosting company, now sold) and can definitely recommend them. Here's just one scenerio: I had really heavy users that paid a lot of money per month to host and I needed a great solution for them to ensure servers ran fast, was reliable and that I had admins that I could count on. I left to cuba, came back and my entire migration process was done for me. I had their two dual quad core 5410's at the time, with a dedicated MySQL backend. Everyone there, especially Chris has been a pleasure to deal with and I suggest that you give them a ring! Mention my name and he'll treat you even that much better Best of luck!
Posted by RelativeDesign-Jerret, 06-25-2009, 07:42 PM
I moved a couple servers from Rackspace to a newer server at HandyNetworks.com about six months ago. I couldn't be happier with their service, they've been wonderful. I also saved quite a bit of money to boot. Regards, Jerret Kinsman
Posted by tunitech, 06-26-2009, 03:40 AM
contact inetu and gigenet and you might find what you are looking for
Posted by mhalligan, 06-26-2009, 02:03 PM
On the very high end, there's Terremark. On the mid-end there's Inetu who I've known for 12 years, and I highly respect them. Then there's the option of going with a very stable provider like SoftLayer, and a fully-outsourced management firm to perform the management.
Posted by SingleHopChris, 06-26-2009, 02:40 PM
I think that based on the complexity of his clusters, it would be best if he had the server and the management with the same company.
Posted by ibejohn818, 06-29-2009, 04:50 PM
Hey everyone! Rackspace crashed again! So i'm the idiot that didn't move in time. this is the 3rd time in 15 days that rackspace has crashed my DB's
Posted by SingleHopChris, 06-29-2009, 04:58 PM
Sorry to hear that you are again having a problem, did you find any other provider that meets your needs? Last edited by SingleHopChris; 06-29-2009 at 05:10 PM.
Posted by ibejohn818, 06-29-2009, 05:43 PM
I'm probably going to move to the planet. I'm still talking to people over there, but now, i must accelerate the move. I'm just not at that magic number for rackspace to give a ****. 200k+ a year just isn't enough for them at least get MY racks on a valid backup power supply. I'm sure they will start loosing many more customers over these pathetic power issues. Wish me luck
Posted by Scott.Mc, 06-29-2009, 05:56 PM
Pity you didn't move already, I assume you are still in contract (this is the only reason many of our customers still use rackspace) but at least you were not alone this time since it seems like large portions/all of the DFW datacenter was out this time, first time the power issue has effected any of our systems. Another provider suggestion is bitpusher (who posted in this thread) if they support windows you should consider them since it seems like exactly the type of provider you are looking for.
Posted by SingleHopChris, 06-29-2009, 06:47 PM
Is The Planet able to provide the complex clustering setup you require? I have not heard much about that aspect of their services.
Posted by KHazard, 06-30-2009, 11:48 AM
For more information about The Planet's complex infrastructure offerings, visit our Virtual Rack and Private Rack pages.
Posted by eSited, 07-07-2009, 01:50 AM
I would also recommend gigenet. We have been a customer with them for quite a while now, never had a problem with uptimes nor power outages. Their support has been quick in terms of responses and problem resolutions in the few times that I emailed them. Furthermore, all dedicated servers are fully managed.
Posted by neuman, 07-07-2009, 02:54 AM
I would stay far far away from liquidweb. Coming from a rackspace setup and moving to a liquidweb setup I feel that you would have more troubles with liquidwebs managed support then rackspace. I used liquidweb a while back for a simple server migration. directadmin to cpanel. Every mysql database they transferred was corrupted. My server load went from maybe 10% cpu usage to 100% and they could not fix it. Long story short I managed to get a full refund after 8 days of service. I would consider contacting gigenet.com for a quote on both the migration and server setup. They are very friendly people and always answer the phone.
Posted by ibejohn818, 07-07-2009, 01:17 PM
Rackspace went down AGAIN! Lucky i've moved my most important data to another network. However, they've scheduled me to move out of their DFW data center to their Virginia data center. I'm now reconsidering the move, even though the VP of operations has assured me that there would be no further power outages as they have corrected the issues. Seems like Rackspace is sub-contracting to a very ****** management company for their data centers now as their support technicians have no idea what is going on, and also no idea of the exact infustucture being implemented. I've requested a co-located UPS, but that cannot support that. I guess rackspace is ANTI UPS.
Posted by mainpipe, 07-07-2009, 05:16 PM
Why would they move you from Dallas to Virginia? What is their reasoning? They are HQ'd out of Texas... don't they have faith in their TX setup? Seems really strange to me.
Posted by lockbull, 07-10-2009, 12:36 AM
maipipe, Rackspace has a few datacenters in Texas, including San Antonio (which is their first location) and Dallas. Rackspace's DFW facility certainly has had a bad run over the last 2 years, and I can't blame you for wanting to get out of there. FWIW, from what I know the uptime at the Virginia datacenter has certainly been much better than at the Dallas facility. The Virginia facility currently operational is in Herndon, and they have another facility coming online this summer in a Dupont facility in Ashburn. If you still want to stay in Texas, you might see if they can relocate you to San Antonio. Also, it's not unusual for a datacenter to not allow third party UPS systems.
Posted by mhalligan, 07-10-2009, 04:20 AM
Can you quantify this, because my experience is otherwise. We regularly manage complex infrastructures for customers who host outside of our facilities, and many of which were originally managed by the dedicated server/colo providers who just weren't in the business of managing complex applications. I've even taken business way from Rackspace, IBM Global Services, and Terremark who are in the business of managing complex applications, just not in an efficient enough manner.
Posted by mhalligan, 07-10-2009, 04:21 AM
Hey, tell Dev I said hi!
Posted by SingleHopChris, 07-10-2009, 10:05 AM
Sorry for the delay, I was on vacation, I base this on conversations with clients that moved from other hosting companies where they have had problems that were not resolved quickly or at all as a result of having third party management. I see it quite a bit, I also think from a practicality standpoint that it makes more sense to be able to talk to the people that have hands on access to the servers, as opposed to making the client the middle man between the management company and the D.C.
Posted by mhalligan, 07-10-2009, 02:37 PM
Dedicated servers are a commodity, nothing more. As long as the ISP provides for good remote access, and the management company is competent, it's likely that any management firm who "specializes" solely in remote server and application management is going to be more effective than a managed server provider whose sole impetus is convincing the customer to consume more hardware. My guys currently manage about 600 servers, for a couple of dozen customers, only 78 of which are in our own datacenter. We manage both colocated servers and dedicated servers in another 7 locations, plus Amazon EC2. That includes understanding our customers applications as well as in-house systems administrators would. We can do this because we are constantly optimizing our processes and tools to manage not servers and applications, but to simplify complexity and understand the delta between those servers and applications. Hardware & VMS are just that, another resource to be programmed by a configuration and application management tool.
Posted by SingleHopChris, 07-10-2009, 03:00 PM
The obvious fallacy is in assuming that a managed server provider's sole impetus is to convince a client to consume more hardware. Many providers do take providing the best possible service to their clients seriously.
Posted by mhalligan, 07-10-2009, 04:03 PM
At the very least, their motives are suspect. If you call up your rep at Rackspace and tell them, "Hey, I got a great deal on 50 servers over at SoftLayer, I need your guys to setup a DR cluster for me there" they will do their best to push back on that statement and try to convince you to host in one of their other datacenters instead. They will not do this because it's the best business decision for you, they will do this because they are in the business of making margin off of dedicated servers. In the end, I doubt that they will manage your resources at another provider. This means you need to build a relationship with a second management vendor, or find a management company who is provider-agnostic.
Posted by DATARTIM, 07-10-2009, 04:27 PM
That may be the case with rackspace, but certainly not any and all managed providers.
Posted by Scott.Mc, 07-10-2009, 04:39 PM
Which many are, be it through simply convincing the customer more hardware is better or be it through inexperienced administrators who do not truly understand what they are actually managing (which many, many, many are guilty of). Ultimately it is going to depend on the clients requirements and the company itself. The above is so good that it needs to be mentioned again. Obviously I am very biased towards using separate administrators that actually understand the product being offered, just like mhalligan and much the same way as anyone offering managed hosting will be biased in the other direction. The argument of being able to access the systems directly is almost always the same one quoted but the reality is when there is no financial gain to be made from the systems themselves then it's naturally more likely that clients are only going to be paying for systems they actually need.
Posted by Layershift Damien, 07-14-2009, 12:17 PM
This argument also assumes that it's more profitable for a provider to switch a customer onto bigger and better hardware instead of keep them on the existing kit. I think the fact is there are good management companies, and good hosting providers offering good management. To suggest that one is inherently better than the other is way too much of a generalisation! There are arguments either way, such as benefits of a single point of contact for any/all issues: if you have a problem with your site then you call the hosting provider. This problem might turn out to be anything from server software/hardware/network and they can deal with it all. The management company is only able to deal with the software portion, so in that sense they cannot help with everything. On the flip-side there is the argument already presented here that some server providers are only interested in selling more hardware rather than providing the best solution for the customer. (It should be noted that in a similar way that there are examples where a management company might suggest the easier management route rather than the best technical option etc.).
Posted by SingleHopChris, 07-15-2009, 04:45 PM
I agree, I can definitely see situations where having a separate management company is better, I am not against that, It depends completely on the service provided by the hosting company. Assuming that a managed hosting provider's main motivation is selling more hardware seems a little off considering that, in my experience, profit margins are much higher for the management services than they are for the hardware itself, to sell it you, have to buy it first.
Posted by mhalligan, 07-22-2009, 02:25 PM
Hardware is a one-time cost that depreciates. Labor is an ongoing cost that just increases over time. Lease out a $3,000 server for $250/month and the profit margin on that is going to be a lot higher of a percentage than whatever it is you can mark up an engineer's time.