Web Hosting Talk







View Full Version : Expected Performance with a VPS Plan


TurboJones
02-25-2005, 06:01 PM
I recently migrated my shared/reseller plan to a VPS. I selected the VPS1 plan from Liquid Web. First of all kudos to the support provided by Liquid Web. It is first rate so far.

My concern is that the performance I'm receiving doesn't even come close to what I had with the shared/reseller plan I had with HostForWeb that cost $24.95.

When I called support at Liquid Web to discuss this I was told that the $60 VPS1 plan is just a really basic plan and that I won't get good performance from it. The VPS2 or dedicated was what they said I would need.

Now I have 11 sites configured with this plan. I have several Mambo sites and 2 static sites. The Mambo sites do not get any real traffic yet as they have not even been launched/published yet. 1 of the static sites get a minimum amount of traffic, about 30 visitors a day. I would think with this light of a load that I would not be seeing performance problems. By performance problems I mean it takes 30 seconds to login to FTP (and yes I'm using Pure FTP).

What I'm asking here is... Should I be expecting better performance from this basic VPS plan? Or do I really have to spend $100 mo. to get decent performance for so few sites?

I really like this company and it was rated very highly on WHT. But I'm beginning to think perhaps they are overloading their VPS servers that host the VPS1 plan. Can anyone provide me some insight here? I am in no way bashing Liquid Web here, I just need to understand what is reasonable to expect regarding performance.

Best regards.
TurboJones

TurboJones
02-25-2005, 06:28 PM
Correction, the previous post should have said 130 visitors per day for the static site.

TurboJones

layer0
02-25-2005, 07:25 PM
You may want to try to optimize your VPS here are some things you can do.

1. Enable PURE-FTPD for faster FTP logins and security.
2. Disable mailman if you do not need it.
3. Disable spamassasin if you do not need it.
4. You can disable some webmail services if you wish.
Basically you can disable anything you are not using on the VPS for the most part.

If you need help with this just holler.

davebytes
02-26-2005, 12:39 AM
I'll be interested to hear other experiences, as that's the 'maximum' plan I'm shooting for at the moment. And I run a blog that typically gets >1K readers per day, and has been slashdotted even. I'm looking at VPS over dedicated basically for redundancy, but if the performance of the average 256MB-guaranteed plans are struggling... well...

More opinions and optimization commentary would be much appreciated (btw, I plan on something like clamav & spamassassin, imap email, etc., so there is other stuff going on... and a dynamic PHP site, but I can tweak the SQL queries if I have a good managed provider with some linux gurus onboard...).

-d

gilbert
02-26-2005, 02:12 AM
im suprised but good to hear that you had a good experience with hostforweb cause i was on the overbooked sirius server almost over a year ago from now

ive noticed vservers are kinda popular with some people who want a dedicated server but i kind fruggle (no offense)

TurboJones
02-26-2005, 10:55 AM
Zeon - thanks for the feedback. Tuning I'm sure will be absolutely necessary with such sparce resources.

I'm still really interested in understanding if I'm just expecting too much performance from such a plan or if indeed it may just be a result of too few resources split between too many VPS accounts. I can't seem to find any guaranteed CPU in Liquid Web's plan description but one of their techs told me it was 256mhz. So with 256mhz and 256mb RAM minimums, should it still take 30 seconds to log in to FTP? Should I be expecting to get more Burst resources for the money I'm paying?

When I decided to go with Liquid Web it was a toss up between ServInt and them. I ultimately went with Liquid Web as the dedicated pricing is much better and as my business grows I plan to migrate to dedicated. I just didn't plan to have to migrate with only 2 paying customers. I thought I'd be able to have more active sites with the basic VPS plan and still get decent performance. I'm beginning to think I should have gone with ServInt but I really don't have any facts to back that up.

Please help me by sharing your experiences with VPS performance.

Regards,
TurboJones

8inet-Johnathan
02-26-2005, 11:49 AM
Turbo,

I've heard nothing but great so far about ServInt, and for less money you are assured 1gb RAM. With this extra ram you only get half the space you have now.

But if you want stability over space i recommend you switch to servint

Yaroslav
02-26-2005, 12:59 PM
Turbo,

I'm not extremely experienced in VPS resource allocation, but IMHO you're being screwed. I've got a tiny VPS for $30 with 200MHz/128M guaranteed, and logging in through SFTP (note the 'S' which applies Diffie-Hellman encryption) inside an HTTP tunnel (yes, this eats even more) takes me only a second or two...

TurboJones
03-01-2005, 12:27 AM
I'm still hoping for some experienced feedback with VPS. Thanks to those that have contributed thus far but my question is really still unanswered.

I know ServInt is a great provider, and I thought Liquid Web was as well. But I'm still not sure if I am indeed getting the short end of the stick with my VPS performance. Experienced input please... and thank you.

TurboJones

Haddy
03-01-2005, 02:29 AM
Originally posted by TurboJones
When I called support at Liquid Web to discuss this I was told that the $60 VPS1 plan is just a really basic plan and that I won't get good performance from it. The VPS2 or dedicated was what they said I would need. What did they base that on...Surely they said more than "this plan wont work, you must upgrade", atleast I hope they did...

TurboJones
03-01-2005, 12:43 PM
Haddy- When I called to ask about performance and the long delay in loging into FTP the tech started to tell me that the plan I was on was really a very basic plan and that I should expect low performance. He then said that their $100 plan was on higher end machines and would be a decent plan for performance but that a dedicated server was really what I needed. The thing that got me most was that he said something like "what do you expect from a $60 plan?" like that was really cheap or something. The thing was that I had just migrated from a shared/reseller plan for $24 that way outperformed it. I know shared plans are a mixed bag and that's why I went with VPS but I would think that even a "basic" $60 plan should at least host 10 sites with minor traffic. I only have 1 site that has any real traffic and my performance is horrible IMHO.

I'd move over to ServInt to try that out but I expect to migrate to a dedicated plan as soon as my business picks up and justifies the expense and Liquid Web has much better pricing on dedicated plans.

Ideas... Feedback...

Thanks.

Haddy
03-01-2005, 01:16 PM
Originally posted by TurboJones
Haddy- When I called to ask about performance and the long delay in loging into FTP the tech started to tell me that the plan I was on was really a very basic plan and that I should expect low performance. He then said that their $100 plan was on higher end machines and would be a decent plan for performance but that a dedicated server was really what I needed. The thing that got me most was that he said something like "what do you expect from a $60 plan?" like that was really cheap or something. The thing was that I had just migrated from a shared/reseller plan for $24 that way outperformed it. I know shared plans are a mixed bag and that's why I went with VPS but I would think that even a "basic" $60 plan should at least host 10 sites with minor traffic. I only have 1 site that has any real traffic and my performance is horrible IMHO.

I'd move over to ServInt to try that out but I expect to migrate to a dedicated plan as soon as my business picks up and justifies the expense and Liquid Web has much better pricing on dedicated plans.

Ideas... Feedback...

Thanks. Is it just FTP that is slow or are other operations slow and is it slow all the time??

I just dont see how that plan would not fit your requirments unless theres something running on these sites that is just eating up resources...

TurboJones
03-01-2005, 01:28 PM
It is all performance. The Virtuozzo control panel often reports 100% cpu utilization.

Like I said above, I have quite a few Mambo sites that use mysql configured but they have no traffic on them. The only traffic being generated is from a static site. My concern is that if I launch a new customer on one of the Mambo sites that I will be wringing the neck of my VPS. I just fitured I'd have more cpu resources to spare with such minimal usage.

When I run the services report on the virtuozzo cpanel it does not show any processes hogging the CPU. It's really kind of frustrating and again I'm just getting an upgrade suggestion from support.

If indeed my site is really using that much resource and moving to another VPS would not solve that, then I understand and will pay more to upgrade. The problem is that I can't justify in my mind that I am taxing this VPS plan that much already. I am very new to non-dedicated hosting and linux based hosts so I'm swimming a little here. I'm sure I'm just a bit nieve.

Can anyone share with me how many sites & types (static vs dynamic) they are running on their VPS with decent performance?

Cheers.

Yaroslav
03-01-2005, 01:40 PM
Can you log in to SSH and run top? It shows processes and CPU usage in realtime and also accumulates values while it runs, maybe it'd give you some picture.

P.S. Just my 2 cents: if I would get such responses from the support team, I would seriously doubt before hosting ANYthing more with them, let alone go Dedicated...

TurboJones
03-01-2005, 01:43 PM
I have run top but it seemed to be reporting total machine resources (at least on the ram) and still did not point out any processes that are hogging resources.

Thanks for the suggestion though.

Yaroslav
03-01-2005, 01:47 PM
For CPU it reports only your processes, AFAIK.

Maybe leave it running and go surf through your own sites? Then get back and see if some CPU time accumulated on httpd processes...

layer0
03-01-2005, 09:22 PM
Hi,

Did you try my optimization?
I would be happy to have a look at your VPS free of charge, (I love these things). If you want me to PM me.

TurboJones
03-01-2005, 09:29 PM
Originally posted by ZeonFx.com
Hi,

Did you try my optimization?
I would be happy to have a look at your VPS free of charge, (I love these things). If you want me to PM me.


Yes I did. I was already using pure-ftp. I was only able to remove a few services because I need the others. No major improvement in performance.

It must be a shared resource thing (burstable resources) as performance varies through the day. I think the machine is just bogged down with too many VPS's and I'm being held to the ~250mhz min cpu range most of the time. Just guessing.

davebytes
03-01-2005, 10:10 PM
I'd be interested to hear from LW on this one -- given that I'm trying to decide amongst LW, ServInt, and one or two others for VPS service... ;)

In theory, all the VPS providers really SHOULD post their guaranteed MHz specs, make it easier on the rest of us doing research.

But, if you're being pegged at 256Mhz (or so), that'd mean that EVERYONE is. And I'd think LW, like the other top-tier VPS companies, would see a machine is maxed out and shift some VPSes around. A number of providers specifically say they leave 'headroom' on every server.

What's the point of being on a taxed VPS box. The redundancy becomes, well, redundant if you don't get bursts of at least a gigHZ now and again so you perform above-the-minimum if an average is taken.

I might have to reconsider going dedicated... my shared hosting is lacking that extra 'oomph' I'm looking for, don't want to go to something 'worse'. ;)

-d

layer0
03-01-2005, 10:26 PM
Originally posted by davebytes
I'd be interested to hear from LW on this one -- given that I'm trying to decide amongst LW, ServInt, and one or two others for VPS service... ;)

In theory, all the VPS providers really SHOULD post their guaranteed MHz specs, make it easier on the rest of us doing research.

But, if you're being pegged at 256Mhz (or so), that'd mean that EVERYONE is. And I'd think LW, like the other top-tier VPS companies, would see a machine is maxed out and shift some VPSes around. A number of providers specifically say they leave 'headroom' on every server.

What's the point of being on a taxed VPS box. The redundancy becomes, well, redundant if you don't get bursts of at least a gigHZ now and again so you perform above-the-minimum if an average is taken.

I might have to reconsider going dedicated... my shared hosting is lacking that extra 'oomph' I'm looking for, don't want to go to something 'worse'. ;)

-d

I have a ServInt VPS would absolutely no problems at all. I host quite a few clients on it and some run resource-intensive websites. I never got a complaint from a single client on that VPS. LiquidWeb probably overstacks their servers and does not leave any headroom. I think you should really check out ServInt as there is basically always room to burst more than your restrictions.


TurboJones, you may have configured something wrong on your VPS. I would love to take a look at it as I said above. If you want me to, PM me :).

Regards,

WiredTree Zac
03-01-2005, 11:07 PM
I've already PM'd TurboJones to see if I can help him out.

We don't overload our VPS plans, and there is plenty of headroom on them. I am logged into all of our servers and looking at them right now. I've run benchmarks on full, active servers (including the Unixbench WHT version), and the performance results have been very good.

And to clarify a few other things, minimum CPU on our VPS1 plans is ~333Mhz, minimum CPU on VPS2 is ~666Mhz, however I have never seen a case where there wasn't "burst" CPU available. The two plans go on seperate servers. For scripting intensive sites, we recommend VPS 2 plans as the extra RAM and CPU help out.

TurboJones
03-01-2005, 11:46 PM
I see some of you are jumping to conclusions when it may just be something unique to me, such as my expectations of performance :-D

Again I'm not trying to raise a stink here I'm just trying to understand and then make business choices.

I see a PM from Zac at Liquid Web, I'll keep you posted on what comes up.

Turbo

Lynx
03-02-2005, 01:24 AM
Sounds way too slow. You may need to upgrade to a plan with more cpu but you ought to do it at Servint instead. Great support board with helpful members sharing scripts to help monitor your site.

People have to keep in mind that when they get a VPS with 128 or 256 ram, a certain percentage is used for the control panel, cutting down on the amount that is really available.

davebytes
03-02-2005, 01:58 AM
Originally posted by Lynx
People have to keep in mind that when they get a VPS with 128 or 256 ram, a certain percentage is used for the control panel, cutting down on the amount that is really available.

Now THAT's a useful line for further questioning. Can someone provide more memory details? i.e., which control panels actively use (approx) how much memory? DirectAdmin, Plesk, CPanel/WHM, Webmin, etc... That would be extremely useful!

Actually, if some people WITH say 128M or 256M VPSes, with various control panels and services installed, could provide some memory footprints/benchmarks (when active might be as useful as when idle), I think many of us in the community could benefit greatly from some deeper details.

Certainly, getting an idea of which OSes (under which VPS system), with certain services/etc. running leave how much remaining active memory (since many of the services might swap out -- i.e., sendmail/postfix, etc.) -- this would seem the 'next logical step' beyond Carlos' current VPS companies spreadsheet.

If there's a guru-type out there that can think of a simple script that would capture some relevant data (might just be top output), that might help.

Actually, I think this would be of use to a lot of folks -- resellers moving from shared environments to VPS would certainly benefit as much as independent site owners doing the same... ;)

THANKS!

-d

Yikes2000
03-02-2005, 03:25 AM
TurboJones,

Have you run the Unixbench (WHT variant) yet? If so, what is your score? Search for Unixbench in the Dedicated Server forum to see the thread in which a lot of results have been posted. It will give you a better idea of the performance of your VPS.

If your web sites are 'responsive' and slow FTP login is the only problem, then I'd suspect some sort of DNS hangup during the FTP login process. You should monitor CPU load (using 'top') during the 30 seconds of FTP login to see if the FTP server process is indeed maxing out.

On the other hand, 256 MHz (or 333 MHz) of CPU cycle isn't much. The Unixbench-WHT uses Pentium 233 for its base index. Thus a score of 10.0 means a server perfermance equivalent to Pentium 233. The Unixbench-WHT thread has several ServInt VPS scores. Compare them to your score to decide if such a move would be prudent. IMHO, ServInt offers less performance for better support.

One second - let me put on a flame suit for making that last comment. :)

TurboJones
03-02-2005, 12:17 PM
Yikes- that was a great suggestion. Exactly the kind of tool I was looking for. Here are my results (note this was in the morning with a user load hitting the static site)

==============================================================
BYTE UNIX Benchmarks (Version 4.1-wht)
System -- Linux host.thebizstarter.com 2.4.20-021stab022.1.777-smp #1 SMP Sat Jun 26 16:57:13 MSD 2004 i686 i686 i386 GNU/Linux
/dev/vzfs 18874368 2877860 15996508 16% /

Start Benchmark Run: Wed Mar 2 08:42:52 MST 2005
08:42:52 up 12 days, 6 min, 1 user, load average: 1.67, 1.32, 1.20

End Benchmark Run: Wed Mar 2 08:56:24 MST 2005
08:56:24 up 12 days, 19 min, 1 user, load average: 17.04, 7.25, 3.93


INDEX VALUES
TEST BASELINE RESULT INDEX

Dhrystone 2 using register variables 376783.7 6221201.1 165.1
Double-Precision Whetstone 83.1 572.6 68.9
Execl Throughput 188.3 1198.8 63.7
File Copy 1024 bufsize 2000 maxblocks 2672.0 14173.0 53.0
File Copy 256 bufsize 500 maxblocks 1077.0 5743.0 53.3
File Read 4096 bufsize 8000 maxblocks 15382.0 256841.0 167.0
Pipe Throughput 111814.6 302428.8 27.0
Pipe-based Context Switching 15448.6 77829.1 50.4
Process Creation 569.3 3047.0 53.5
Shell Scripts (8 concurrent) 44.8 179.9 40.2
System Call Overhead 114433.5 209703.6 18.3
=========
FINAL SCORE 56.4


Well from the results, compared to what I read in the Dedicated forum results, my performance doesn't look too bad. I wonder how much of a difference this benchmark gets based on different server loads? I'll try running it again later during the peak time.

Turbo

rasputinj
03-02-2005, 04:01 PM
Originally posted by davebytes
I'll be interested to hear other experiences, as that's the 'maximum' plan I'm shooting for at the moment. And I run a blog that typically gets >1K readers per day, and has been slashdotted even. I'm looking at VPS over dedicated basically for redundancy, but if the performance of the average 256MB-guaranteed plans are struggling... well...

More opinions and optimization commentary would be much appreciated (btw, I plan on something like clamav & spamassassin, imap email, etc., so there is other stuff going on... and a dynamic PHP site, but I can tweak the SQL queries if I have a good managed provider with some linux gurus onboard...).

-d

I had a vps plan with dinix on Virtuozzo and VPSColo on UML, I was getting at the time around 1,700 uniques a day. Both were on a 256MB plan, and the vps with Vpscolo handled it quite well even with a few other sites up and running, I had some problems with Dinix, might have been more of a load issue.

You can do stuff, with your site like mmcache and or jpcache, which should let your site handle thousands of hits a day without problem and speed up things a bit.

layer0
03-02-2005, 04:32 PM
I have another suggestion. You may want to enable the MySQL Query cache as that really helps with CMS based websites. To learn how, check out an article on http://www.databasejournal.com

davebytes
03-02-2005, 07:49 PM
rasputin - thanks for the comments. always helpful to hear what people can do in 256MB.

Zeon - the query cache is a good suggestion. I'm currently doing a php-based caching system in my intermediate database handler, for some specific queries that I know weren't changing often. Easier to pull off disk, and then heavy stuff will sit in the disk cache. ;) But when you don't know PHP, or don't have real control over the queries, the QC is a great way to 'spend some RAM' for performance.

I'm really struggling with the cost-benefit analysis between 256MB VPS and a 512MB dedicated box... I've seen managed VPS down to maybe $45/mo for a 256MB provisioning... but once I get up toward the $60/mo for VPS, not sure that $70-80 for a cheap dedicated wouldn't do better. Dell has $500-600 entry servers (mini-tower or 1U...), which I could cover up-front, and colo it with someone who will give me basic management (I'm not a linux geek, but I'm no idjit either.). But there's something to be said for VPS enterprise redundancy...

-d

TurboJones
03-02-2005, 11:09 PM
Zeon - thanks for the MySql tip. I'll check that out. I've also enabled caching in Mambo which should help alot for pages that don't change often.

TurboJones
03-03-2005, 12:05 AM
I ran another UnixBench on the server now that traffic has died down. Seems pretty impressive compared to other VPS scores I read about. I guess LiquidWeb really IS a great provider and my concerns are a result of my lack of experience with VPS's. I've only used shared & dedicated in the past.

Zac at LiquidWeb has been helping me to find out about the FTP delays and the other connection oriented delays. Again great service from them.

I'll do some tuning and launch a couple of Mambo sites and see how it goes. I'll also be moving to a dedicated box in a few months which I'm sure will help alot. Wish me luck and thanks to all that have contributed to my experience & understanding.

Best regards,
TurboJones

Here are the UnixBench results during a low load:

LiquidWeb VPS1 Plan $60


==============================================================
BYTE UNIX Benchmarks (Version 4.1-wht)
System -- Linux host.thebizstarter.com 2.4.20-021stab022.1.777-smp #1 SMP Sat Jun 26 16:57:13 MSD 2004 i686 i686 i386 GNU/Linux
/dev/vzfs 18874368 2881196 15993172 16% /

Start Benchmark Run: Wed Mar 2 19:59:22 MST 2005
19:59:22 up 12 days, 11:22, 1 user, load average: 1.01, 0.75, 0.80

End Benchmark Run: Wed Mar 2 20:12:38 MST 2005
20:12:38 up 12 days, 11:35, 1 user, load average: 17.42, 6.98, 3.54


INDEX VALUES
TEST BASELINE RESULT INDEX

Dhrystone 2 using register variables 376783.7 6647374.7 176.4
Double-Precision Whetstone 83.1 573.2 69.0
Execl Throughput 188.3 1806.2 95.9
File Copy 1024 bufsize 2000 maxblocks 2672.0 15172.0 56.8
File Copy 256 bufsize 500 maxblocks 1077.0 6262.0 58.1
File Read 4096 bufsize 8000 maxblocks 15382.0 312251.0 203.0
Pipe Throughput 111814.6 328156.5 29.3
Pipe-based Context Switching 15448.6 77876.1 50.4
Process Creation 569.3 3215.5 56.5
Shell Scripts (8 concurrent) 44.8 274.9 61.4
System Call Overhead 114433.5 245214.5 21.4
=========
FINAL SCORE 64.9

layer0
03-03-2005, 04:47 PM
Nice Benchmark! That is excellent considering it is coming from a VPS.

TurboJones
03-03-2005, 04:57 PM
Hi Zeon, agreed. It was really unnerving not knowing how it was performing and only using my response times as feedback. That benchmark tool helped my confidence tremendously.

I still get some latency when I first visit one of the Mambo sites after some dead time since they are not active. I think the delay is just waking up the resources (sleeping processes, swapped memory, etc) because it starts to respond decently after some use.

Thanks again for your feedback and help.

TurboJones
03-03-2005, 06:17 PM
Well, just an update for everybody...

All of my performance issues have been resolved (for now <wink>). No more FTP delays YESSS! I am now very confident in launching my new business on the LiquidWeb VPS platform.

Thanks to Zac and the LiquidWeb team. I'm extremely happy with my choice of hosting provider.

TurboJones

davebytes
03-03-2005, 09:21 PM
If Zac is willing, I'd love to hear some of the issues that were (are?) there, and solutions. I'm trying to decide what the difference is between a $45 VPS and a $60 VPS (obv, LW is the upper one...), and would like to understand more of performance issues and resolution.

I've always had the impression that LW looked like they'd be a good provider, but I'm tryin to justify an extra 100-200/yr in total cost.

-d

WiredTree Zac
03-03-2005, 10:08 PM
davebytes,

I'm not sure what the issue was completely. I didn't have any delay when logging in through FTP on the machines I tested with. There wasn't any other obvious indicators like failed resource limits, and the machine was definitely not overloaded (as you can see from his benchmarks). We did upgrade the kernel on the server TurboJones is on, and that seemed to take care of the issues for him.

I have to watch what I say because of self-promotion rules in the forums, but when you look at a provider, if you're only looking at account specs and pricing, you're not getting a complete view of the service that will be provided. Support and management should be included in the overall calculation - does the company have onsite level 3 techs you can phone 24/7? Is the account unmanaged or fully managed? Is a control panel included? Are the bandwidth providers they are using quality? Are addons like Fantastico included or are they extra? What other services are provided? Are backups included? Also, performance. If you compared TurboJones's VPS benchmarks to some others, you may be surprised. What I'm trying to say that is that you need to take everything into account when looking at a plan - not just strictly memory/disk space/transfer/monthly fee. If you're already doing that, great! :)

TurboJones
03-03-2005, 10:39 PM
Davebytes- I on the otherhand don't have any self promoting issues ;-)

When I was shopping for a VPS provider, service was my main concern and I was willing to pay a premium for it. I actually had a $40 VPS plan with HostForWeb before switching to LiquidWeb and the reason I moved was service. I woke up one day to find that apache & mysql had been down for about 12 hours and my "managed" VPS was still broken when I got to it. A prompt email (no phone support) to their tech support resulted in the issue being fixed 2 hours later. Needless to say I cannot afford for that to happen when I have a business that depends on quality of service.

My choice then was between ServInt and LiquidWeb based on the testimonials here on WHT. Both companies have sterling histories of service. ServInt was $5 cheaper (after adding cPanel) but I had also planned to get a dedicated box after my business increased. LiquidWeb is quite a bit lower in cost for dedicated boxes and so I went with them. Turns out that I not only got great service but the performance numbers from UnixBench are 5 times better than those reported by other users on ServInt. Check out the Dedicated forum and search for UnixBench to see other results.

Both companies are great and there are others too. But I can tell you that I've had several learning issues configuring and tuning and LiquidWeb tech support being available 24/7 by PHONE is fantastic. I definetly made the right choice for my business. It's worth the few extra bucks each month to have peace of mind (as compared to lower cost VPS's). It does cost money to have qualified staff 24/7 so a company offering plans for less have to cut somewhere. It's simple business arithmatic.

ChrisLM2001a
03-04-2005, 10:50 AM
It was because of a phone call at 3am that made my decision to signup with Liquidweb. As someone who likes working in the wee hours of the night, it's nice to know there's service at that hour.

I've been searching for a small shared account for a simple site, and got the run around with another host just to activate one. So I saw LW's special and thought it wouldn't hurt to give them a ring. Glad that I called because I got someone who not only spoke clear English, he clarified some points, didn't shirk at special requests for a VPS config, and he wasn't asleep too. ;)

If you run a business, or have a rush project or having server problems relating to the network or server itself, you can't wait 3 days to activate an account or get a problem resolved. I know hosts got to protect their bottom line, but a timely response goes a heck of a long way. I lost faith in a host I was a client and recommended highly for over a year, and I know I'm not the only client that felt like that. Good hosts are hard to find, but when even the one you knew was good goes South, it feels like losing family. :(

UPDATE: ordered a plan @ 6:30am, and by 9am got a call confirming the billing and config, saying it'll probably take 8 to 9hrs to finish (which is fair as it's not a stock config, and was honest on what they couldn't do). That's a fast turn around, even faster that what I got from The Planet for a much more expensive server.

Don't know how they are after the sale, but for someone needing an account today done your way, I'll seriously look them up!

Chris

breaume
03-11-2005, 06:51 PM
TurboJones,

I have to tell you this thread was like a good movie playing at my emotions. I just signed up with Liquid Web a few days ago. I ran across this post. I thought great someone on the same plan as me. Then I start reading about the apparent performance problems – Oh no. Then it has a happy ending. This was like a rollercoaster ride.

Anyway I chose liquid web for many of the same reasons you did. Good reviews here at WHT and Future growth. I went With a VPS plan because I was tired of other the other hosts clients sucking the life out of my semi-dedicated server and killing my site. I don’t really need the bandwidth or storage yet. But I am planning on moving Ecommerce site there and want it running like a top.

Let me say that I have been working on Central office environment nearly 10 years. But have had little exposure to servers and host. I am not a beginner but a novice. So this was a little nerve racking. I know enough to get into real trouble.

Tech support was priority one. I have to say that the people at Liquid Web have been as helpful and courteous as any tech support I have dealt with (including Nortel, Tellabs …). I had a certain set up, that I know no as a server alias that I need done. But could not really explain. They asked me ‘what do you need done your not really being clear.’ I had to give them ‘If I knew I would have told you already.’ They bit their tongue and helped me through it.

I just wanted to give credit where credit is due. We live in a society of complaints and expectations. Its kid of funny the amount of service we expect for $60/ month. Great Job Liquid Web.

** After reading my post I sound like the owners relative. I don't have any affiliation with LW just a happy customer so far.

TurboJones
03-11-2005, 07:32 PM
Sorry about the roller coster ride LOL

I thought I was clear in my initial message that it could be my perception of performance... and indeed it was (except for the FTP login which was resolved). The LW VPS is performing well above many other providers reviewed here (in the benchmark thread).

Good luck to you.

hostingshelf
05-13-2005, 11:59 AM
I have been using Liquid Web's Basic VPS Plan for about 3 months now and give kudos for support, although I wish they would explain how they resolved some problems without me having to pry it out of them. I am having one heck of a time getting ftp logins to speed up. I have done all the things I've found on the forums, use PureFTP, stuff like that but it still takes 30-60 seconds for ftp logins to execute. I ran unixbenchmark with the following results, which I understand are pretty good. Any more ideas about how to speed up the ftp?

BYTE UNIX Benchmarks (Version 4.1-wht)
System -- Linux host.hostingshelf.com 2.4.20-021stab028.7.777-smp #1 SMP Thu Apr 14 15:04:58 MSD 2005 i686 athlon i386 GNU/Linux
/dev/vzfs 18874368 3397940 15476428 19% /

Start Benchmark Run: Fri May 13 10:12:13 CDT 2005
10:12:13 up 2 days, 58 min, 1 user, load average: 1.21, 1.01, 0.85

End Benchmark Run: Fri May 13 10:25:17 CDT 2005
10:25:17 up 2 days, 1:11, 1 user, load average: 19.26, 7.55, 3.78


INDEX VALUES
TEST BASELINE RESULT INDEX

Dhrystone 2 using register variables 376783.7 4682294.5 124.3
Double-Precision Whetstone 83.1 734.7 88.4
Execl Throughput 188.3 1298.6 69.0
File Copy 1024 bufsize 2000 maxblocks 2672.0 23522.0 88.0
File Copy 256 bufsize 500 maxblocks 1077.0 7177.0 66.6
File Read 4096 bufsize 8000 maxblocks 15382.0 213561.0 138.8
Pipe Throughput 111814.6 421464.1 37.7
Pipe-based Context Switching 15448.6 79758.3 51.6
Process Creation 569.3 2392.7 42.0
Shell Scripts (8 concurrent) 44.8 229.2 51.2
System Call Overhead 114433.5 376602.8 32.9
=========
FINAL SCORE 64.8

TurboJones
05-13-2005, 01:40 PM
I use SFTP now and it's faster and secure. I'm not sure why standard FTP is slower. My FTP connection speed used to be about 30 seconds as you read in my posting but they did something to speed it up to about 8 seconds. But SFTP/SSH is much faster still and takes about 3 seconds to connect.

Hope that helps.

Regards,
TurboJones

hostingshelf
05-13-2005, 02:46 PM
I like the sound of that. Questions:

How did you get Liquidweb to properly address the issue? I just keep getting the runaround with them on it.

This SFTP - Where do I get and how do I install it to work with cPanel - or is this just a protocol or something similar?

Thanks for any help you can give me.

TurboJones
05-13-2005, 02:56 PM
SFTP is Secure FTP. It's a protocol. Most major FTP clients support it such as WSFTP and FileZilla.

PM Zac at Liquidweb for assistance. He responded to my post here and that is when I noticed the FTP performance increase. See the prior responses for his info.

I've since upgraded to the VPS 2 plan. It's WAY faster and I have no delay issues. Well worth the money for my business.

ikeo
05-13-2005, 04:38 PM
Use winscp3 if you're going to use scp. scp is just ftp over the ssh port. So if you have access to ssh you can use scp.

For the slowness with ftp, are you guys using iptables? If you could do a simple test and turn off iptables temporarily try to FTP and see if it works. Then turn it back on and try again... See if that has any effect.

nucleotide
05-13-2005, 06:14 PM
Turbo jones, you say you only have a few light traffic sites and yet apparently the plan you chose wasn't sufficient, so how would somebody planning to resell hosting cope on a plan like this.

Or put another way 'How does a VPS compare to a reseller in terms of ram, cpu usage, or are the main benefits of VPS those of being secluded from other peoples actions on the same reseller server affecting you and of having greater control of your environment?'

Can somebody enlighten me, please?

Cheers

TurboJones
05-13-2005, 06:29 PM
Originally posted by nucleotide
Turbo jones, you say you only have a few light traffic sites and yet apparently the plan you chose wasn't sufficient, so how would somebody planning to resell hosting cope on a plan like this.

Can somebody enlighten me, please?

Cheers

Part of my issue was education. The basic VPS plan was sufficient for my needs at that time. I have since launched my business and have some heavy traffic and so I upgradded to the next level to insure good service.

If you look at the performance stats from my previous basic VPS it was more than competitive with others out there. In most cases much faster.

As to VPS vs Shared the greatest difference is guaranteed resources and control. You are at the whim of the hosting company with a Shared account as you could easily end up on a machine with heavy loads and there is no guarantee of performance/resources. VPS provides a guarantee and also allows for bursting of performance and resources as the resources are available. Think of VPS as a small dedicated server.

For reselling I wouldn't choose anything but a VPS or Dedicated server because at least I will know what resources I have to offer and can plan for growth more reliably than can be done with Shared hosting.

Hope that helps. Regards, TurboJones

nucleotide
05-13-2005, 06:41 PM
Thanks for the clarification.