1. Update November 2009
2. Signing Up
3. MoxieVM
4. Plan Upgrades
5. Performance
6. Conclusion

Looking at my history of past reviews and found that I have not done a hosting review for the last 8 months! Meanwhile, I probably have used half a dozen different hosting providers since my last review (my review on Gandi.net’s Xen VPS, which I will hopefully write about them again sometime soon). So, it’s the last day of August, and let me do a quick review on one of the VPS hosting company that I have been using for the last 3 months and are quite happy with — Fivebean Media.

I first came across Fivebean Media on LowEndBox.com of their $5/month VPS deal (which I do not think is currently on offer). According to the about page, the company started as a “technology/web consulting company” back in 2004 (XlogicGroup, a service/consulting company), and Fivebean Media was created in 2008 to offer web hosting. It is still a relatively young company, and it has a weird name for a hosting company, but I do like their clean design :)

My next encounter with Fivebean is from their advertisement on OzBargain — yes a Michigan company posting “deals” on a website which the main visitors are all the way across the Pacific. It was one of those offers that was “too good to be true” (and yes I am a sucker on bargains which was the reason I started that site in the first place) — an OpenVZ VPS plan with

• 768MB dedicated memory, 1,200MB burstable memory
• 35GB disk space
• 275GB monthly transfer
• Equal share CPU
• $17.51/month AUD

It’s the “Bite” plan on their VPS plan page at more than half the price. Their servers are hosted at Databank in Dallas. Since I have not yet had a VPS in Dallas before (yeah, pretty lame excuse), I went for the impulse buy to took the bite…

Signing Up

Signing up was almost instant. Ordered on 23 May at 11:02pm AEST, and the VPS was paid for, activated and the welcome email was received by 11:19pm AEST. In this day and age some VPS providers, especially on the budget end that have no expertise in writing their own automation, still have not been able to provide instant activation. Big plus for Fivebean on this one.

Welcome email contains URL and credential to HyperVM’s control panel as well as IP address and root password to the newly activated VPS. Note that like almost all the other budget VPS providers back in May, Fivebean used HyperVM to power their virtualisation infrastructure. That was before the hacking incident at FsckVPS and the tragedy at LxLabs. On 7^th of June, prior to FsckVPS’s hacking incident but after LxLabs’ last vulnerability report, Fivebean decided to turn off HyperVM access completely for security reason. A few days later a new control panel project was announced, and on 18^th of June, their new VPS control panel was released to the customers — a very impressive turn around time.

Name of their own control panel? MoxieVM.

MoxieVM

MoxieVM is Fivebean’s VPS control panel to replace HyperVM. It has almost all the features you need to control a VPS — restart, bandwidth/disk space stats, change root password, reverse DNS, etc. I won’t call it ugly, but certainly has a very simple design.

It lists out all the VPS under your account and some basic stats when you log in, and when you click on “Manage”, it then goes to a page with all the operations you can do with that VPS.

As you can see it is pretty straightforward no-frills kind of design. The only lacking in my opinion is some kind of console access built into the control panel, in case of getting firewalled from your own VPS. MoxieVM offers recent version of Debian, Ubuntu, CentOS, Fedora, OpenSUSE and Slackware templates. Some are in 32bit and some 64bit. On the rebuild page it clearly states the rebuild process can take up to 7 minutes to complete if there is a queue.

I think MoxieVM is good enough for most of us, even though it is still a work in process. You can find notes in the applications that state out there’s a bug and “will be resolved in the next update”. Some features (TUN support) are currently turned off.

Plan Upgrades

Another big plus is the plan upgrades that I have been receiving. I am already paying 1/2 of listed price for the Bite plan, and then on 2^nd of June the spec got upgraded again. Now logging into MoxieVM I can see that my VPS has:

• 768MB guaranteed/1.5GB burstable memory
• 60GB disk space
• 550GB/month transfer quota

All for USD$13.71/month. I guess it is the power of OpenVZ which makes overselling easier. As I am still in the process of migrating some service over, my VPS is using only around 170MB privvmpages out of 1.5GB available. I am sure many are in the same shoes on the same box, which means they can squeeze a few more users on that box.

Nevertheless, occasional plan upgrades make customers happy. Take a note on that, SliceHost :)

Performance

Now let’s come to probably the most important bit — actual performance of the system. I am using Debian 5 32bit here. Dumping the /proc/cpuinfo revealing that:

• Core 2 Quad CPU Q9300 at 2.5Ghz was used on my VPS
• No CPU-limiting has been set for OpenVZ, so equal-share CPU time.
• Access to all 4 cores

Q9300 is actually a desktop class CPU, although it is available for many rack mounted servers at configuration time. It does however led me to believe that they are opting for cheaper servers with Core 2 Quad instead of Xeon, RAID’ed SATA instead of SAS drives, etc. Nothing wrong with that in this end of the market I think, as some providers choose to have less-beefy boxes but fewer VMs per box. Just don’t be surprised that it’s not a Xeon (or Opteron) though.

Here is a dump of Unixbench that I just ran:

                     INDEX VALUES
TEST                                        BASELINE     RESULT      INDEX

Dhrystone 2 using register variables        376783.7  9116420.8      242.0
Double-Precision Whetstone                      83.1     1368.0      164.6
Execl Throughput                               188.3     5134.4      272.7
File Copy 1024 bufsize 2000 maxblocks         2672.0   142437.0      533.1
File Copy 256 bufsize 500 maxblocks           1077.0    38563.0      358.1
File Read 4096 bufsize 8000 maxblocks        15382.0   685888.0      445.9
Pipe Throughput                             111814.6  1560446.2      139.6
Pipe-based Context Switching                 15448.6   443351.5      287.0
Process Creation                               569.3    18032.8      316.8
Shell Scripts (8 concurrent)                    44.8     1417.8      316.5
System Call Overhead                        114433.5  1350673.5      118.0
                                                                 =========
     FINAL SCORE                                                     264.2

Not bad.

Network connectivity and bandwidth is pretty good as well. I created a 30MB file and tested out downloading them from various places on the Internet:

• From Fremont CA (my Linode VPS): 8.6MB/s
• From St. Louis MO (my Slicehost VPS): 3.3MB/s
• From London: 1.1MB/s
• From Sydney: 550KB/s

I think the test result would of course vary depending on location and time but I am generally happy with what I got.

Conclusion

Initially I thought I would just test it out for 3 months but I think I am getting pretty good performance from the amount of money I’ve paid, and are happy to continue to renew. Other reviews I found on the net (here at WHT) are also positive.

I will be gradually migrating some services over to this VPS at Fivebean in the next few months, and might follow up a review when it’s finally “fully loaded”.

VPS Plans and Pricing

You can find out more about the pricing of Gandi.net Xen VPS on their server pricing page. For each share, it’s currently at $14/month (launch price until Dec 31 2008, and I have completely no idea what the price will be afterwards). You can buy “shares” on daily basis, so you can scale up or down without paying for the capacity for the whole month.

Now, what is a “share”? It defines a fixed amount of server resource, and according to Gandi, it’s:

• 1/64^th of a Quad Quad-Core AMD.
• 256 MB guaranteed memory + 512 MB of Swap
• 5 GB hard disk storage
• 5Mbps bandwidth + unmetred traffic.

So a 256MB Xen VPS for $14/month is actually not bad (I am currently paying SliceHost $20/month for 256MB). However there is no discount when you scale up. Therefore a 512MB VPS costs $28. A 1GB VPS costs $56. A 2GB VPS costs $112 — the cost starts to become less favourable when you scale up the number of shares.

It does have very generous bandwidth though — unmetred 5Mbps per share. That means your 1GB VPS will have 20Mbps bandwidth, enough to run a small video sharing websites.

Control Panel

Gandi’s server control panel is probably the highlight of their service. There are multiple ways to configure your new Xen VPS at Gandi.net, and they all starts with dragging a slide on the number of shares.

Then you are presented with three ways to create your new server:

1. Use Gandi MI (manual install) — just like other providers’ control panels, where you get to choose a Linux distribution and that’s it! Currently Ubuntu, Debian, Fedora, Mandria, OpenSUSE and CentOS are offered.

2. Use Gandi AI (custom mode) — where the real gem of the Gandi control panel is. You start with a Gandi-customised Ubuntu 8.04 install, and then you get options to add and configure various software. For example, configurations for FTP server, HTTP server, mail server and database server.

3. Use Gandi AI (easy mode) — like Gandi AI above, but you choose the functionality of this VPS instead of customising various software packages. Currently there’s only one predefined configuration, which is “A PHP/MySQL web server”. Gandi AI will then install Apache, PHP, MySQL, phpMyAdmin and a FTP server onto VPS.

The list of servers that can currently be installed by Gandi AI are:

• Apache 2.0 (plus options such as Awstats, mod_php, mod_perl, mod_python, mod_rail, Django, RoR, etc)
• Subversion
• ProFTPd
• TeamSpeak
• Postfix
• MySQL 5.0 (plus optional phpMyAdmin)
• Dotclear 2.0 (with debatable description “Dotclear 2.0 is probably the best blog application on the web”)

After you picked the software you wish to install, you get to choose the configuration on the next page (MySQL password, Apache virtual hosts, system users/groups, etc). Then you can submit the server-creation request, and wait a few minutes before a new system is set up and deployed.

Gandi AI is pretty much useless to me personally because I prefer the server to be set up the way I want it. While it is great that you can quickly set up a LAMP server or a TeamSpeak server with a few button clicks for those who cannot work around “apt-get”, there are a few issues that really put me off.

• Speed. Or lack of. Maybe Sydney to France might not have the best route, but operations on Gandi’s control panel just feels so slow. I am not talking about just in terms of responsiveness, but any queued task will take a few minutes to sort out. Actually, creating a Gandi AI customised VPS took almost 1 hour for me. Linode also uses a queue interface, but tasks get cleared out much faster that you don’t even notice a queue is there.

• Complexity. Yes it feels like it’s built by engineers for engineers. In the “customise” mode there are just so many parameters that you need to fill out before a server can be created.

I have still not yet fathomed why you want to pre-configured software on a VPS with a web interface, especially complicated set ups like Apache virtual hosts. I understand that it’s not targeting those who are “in the know”, but wouldn’t an option to install “webmin” and let it take over from there make more sense?

Gandi Flex – Real Time Scalability

Another unique selling point of Gandi VPS is Gandi Flex that you are allowed to pre-schedule the number of shares your VPS is going to have, i.e. to allow you to have “burstable” resources on a Xen VPS. Moreover, the extra shares are charged by the hours you have them so if you are expecting the once off spikes, it can be quite economical.

For example if you are going to be posting the latest Apple rumour on your blog, you can allocate extra shares for your VPS for the next 24 hours to handle the extra load. Your VPS actually does not go offline — the memory size just magically increases and you suddenly get more CPU power when the schedule kicks in. Or you know your peak traffic is always the same time every single day (like those 1 deal a day sites), you can then schedule a daily power boost for 2 hours, rather than permanently paying for a high-end VPS.

Of course that in practise, you never know at what exact time your site will be slashdotted. You still need constant performance monitoring, but this kind of flexibility provided by Gandi should hopefully provide some automation to many use cases.

Performance

Finally, let me talk about the performance. It is something that someone would put #1 priority in choosing a service provider, and I am afraid to say that with my experience so far, Gandi’s VPS has been disappointing.

Network performance is great. Sydney is around 330ms away from France so doing SSH can be a bit laggy. However network is really fast thanks to the dedicated 10mbps pipe. As I have said before, it would make ideal home for a low CPU usage video streaming site.

The rest of VPS, however, does not look that good. This is the UnixBench score for a 1 share 256MB VPS:

TEST                                        BASELINE     RESULT      INDEX

Dhrystone 2 using register variables        376783.7   571267.5       15.2
Double-Precision Whetstone                      83.1     1185.4      142.6
Execl Throughput                               188.3      274.2       14.6
File Copy 1024 bufsize 2000 maxblocks         2672.0    10814.0       40.5
File Copy 256 bufsize 500 maxblocks           1077.0     4200.0       39.0
File Read 4096 bufsize 8000 maxblocks        15382.0    43346.0       28.2
Pipe Throughput                             111814.6   110113.0        9.8
Pipe-based Context Switching                 15448.6    15650.0       10.1
Process Creation                               569.3      479.3        8.4
Shell Scripts (8 concurrent)                    44.8       59.9       13.4
System Call Overhead                        114433.5   183354.8       16.0
                                                                 =========
     FINAL SCORE                                                      20.4

Check the relevant WHT thread for comparison. When I increase the shares to 2 (512MB), the performance is roughly doubled, which shows that they have put some CPU limits on the VPS, correlates to the number of shares it has. Yes it feels so slow that my testing WordPress blog with only 3 posts cannot do more than 1 requests per second.

Therefore you basically need to look at having multiple shares to host any dynamic website with sub 0.1 second page generation time, which can make it more costly.

Conclusion

Thanks again for Nicolas and Wendy from Gandi for providing me a testing account to try out Gandi servers. However the result is mixed…

• Pros –

  • $14/share/month is good from a named brand and well-known provider.
  • Great network performance. Very cheap unmetred bandwidth.
  • Gandi AI is interesting, and might be useful for those who are not sure how to set up a server.
  • Gandi Flex can be a powerful tool when the time of traffic burst is known.

• Cons –

  • Gandi AI and control panel in general feels a bit sluggish (hopefully you don’t have to use it too often).
  • The general usefulness of Gandi AI is questionable, as it might not be as up-to-date as many free control panels sitting inside the VPS.
  • Slow disk IO. Very slow CPU (or rather, you can’t get any burst on CPU).

So basically I asked a few questions in the forums (I am “scotty” on WHT.au) –

• Q: Virtualisation? A: Combination of MSFT and Xen
• Q: Debian or Ubuntu? A: Yes
• Q: Data Centre? A: Sydney (later I found out it’s Equinix)
• Q: You insane? A: (1) Limited offer (2) Bandwidth is subsidised (I guess subsidised == oversold)

Well. It’s only around 40 bucks so a few days later I signed up to give it a test. Almost 2 months later I am now writing this review.

VPS Plans — Cheap Australian Bandwidth

Rule of thumb #1 — when someone said it’s “limited offer”, never believe in it! At least I don’t mind in this case, as their “special price” is still available today, 2^nd month into our new financial year. Here’s their pricing table (taken from their website):

$40/month for 100GB is hard to beat, and $70/month for 500GB (AUD$0.14/GB) is way below cost. Storage is on RAID, and the amount of memory is more than enough for what I intended to use this VPS for — serving static data to Australian visitors (as an alternative to using a CDN like CacheFly). I planned to just install Nginx and let it serve Javascript, CSS and small images.

Signing Up

Signing up was straight forward although a bit of manual process is involved. Within 10 minutes of paying by credit card, Nathan from NCL emailed me asking which OS I would like to install. “Ubuntu 8.10″ I replied, and then got the welcome email with IP address and “admin” password within two hours — on a Sunday evening at 10pm!

Although the account activation is not instant, NCL and Nathan still got my thumb up for the fast process.

Initial Impression

So I have now received my IP address and the password, what am I waiting for?!

There is no SSH access!

There is nothing listening on port 22. Having used several unmanaged Linux VPS for the last two and half years and it’s the first time that I was given a freshly-installed Linux VPS without SSH access. It turns out that instead of installing the minimum Ubuntu Server, the standard Ubuntu was installed. On top of that they have also installed Webmin + other packages that I don’t intend to use (Apache, mod_php, etc). The password in the welcome email was actually the Webmin password.

So for the next hour I was trying to figure out how to log into Webmin and then install the OpenSSH server. A bit of fiddling around I have finally got the Bash prompt as root, and the first thing I did was apt-get remove all the extra packages to strip the VPS back to bare minimum.

Then I tried to check out what kind of hardware I got.

$ uname -a
Linux localhost 2.6.24-16-server #1 SMP Thu Apr 10 13:58:00 UTC 2008 i686 GNU/Linux
$ cat /proc/cpuinfo
processor       : 0
vendor_id       : GenuineIntel
cpu family      : 15
model           : 4
model name      : Intel(R) Xeon(TM) CPU 3.40GHz
stepping        : 8
cpu MHz         : 3400.777
cache size      : 1024 KB
fdiv_bug        : no
hlt_bug         : no
f00f_bug        : no
coma_bug        : no
fpu             : yes
fpu_exception   : yes
cpuid level     : 5
wp              : yes
flags           : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss constant_tsc up pebs bts sync_rdtsc pni
bogomips        : 6857.87
clflush size    : 64

Well. Looks like the previous era of Xeon processor, and I only have access to one of the cores. It is running 32bit Linux with kernel 2.6.24. Here is the result of disk performance test:

# hdparm -tT /dev/sda1

/dev/sda1:
 Timing cached reads:   782 MB in  2.00 seconds = 391.53 MB/sec
 Timing buffered disk reads:   70 MB in  3.07 seconds =  22.81 MB/seca

# ./seeker /dev/sda1
Seeker v2.0, 2007-01-15, http://www.linuxinsight.com/how_fast_is_your_disk.html
Benchmarking /dev/sda1 [19751MB], wait 30 seconds..............................
Results: 501 seeks/second, 1.99 ms random access time

Interesting result! The throughput is a little bit disappointing for a RAID storage but 2ms seek time on a VPS?! You have to do a RAID10 on 15k RPM drives or SSD to get that! Or maybe the virtualisation layer is doing some tricks on me?!

Not Xen, Not MSFT — It’s a VMWare

It turns out, that my VPS is not on a Xen node. Nor on a MS Virtual Server/Hyper-V node. In fact it is running VMWare.

# lspci
00:00.0 Host bridge: Intel Corporation 440BX/ZX/DX - 82443BX/ZX/DX Host bridge (rev 01)
00:01.0 PCI bridge: Intel Corporation 440BX/ZX/DX - 82443BX/ZX/DX AGP bridge (rev 01)
00:07.0 ISA bridge: Intel Corporation 82371AB/EB/MB PIIX4 ISA (rev 08)
00:07.1 IDE interface: Intel Corporation 82371AB/EB/MB PIIX4 IDE (rev 01)
00:07.3 Bridge: Intel Corporation 82371AB/EB/MB PIIX4 ACPI (rev 08)
00:0f.0 VGA compatible controller: VMware Inc Abstract SVGA II Adapter
00:10.0 SCSI storage controller: LSI Logic / Symbios Logic 53c1030 PCI-X Fusion-MPT Dual Ultra320 SCSI (rev 01)
00:11.0 PCI bridge: VMware Inc Unknown device 0790 (rev 02)
02:00.0 Ethernet controller: Advanced Micro Devices [AMD] 79c970 [PCnet32 LANCE] (rev 10)

When you list out all the processes, you can also see vmware-guestd running. Is NCL Hosting running ESXi? Infrastructure? I have no idea but at least the VPS performs alright over the last 2 months.

I think the main issue I have with VMWare is the leaky vmware-guestd. It uses quite a bit of CPU, and the memory usage grows from 1.2MB to 20MB within a month. I have to run /etc/init.d/vmware-tools restart to claim the memory back every now and then although I have no idea whether it will introduce any ill effect on a VMWare-based VPS.

2 Months Later…

So how is it after 2 months? I installed Nginx to serve up static data so I am not running anything CPU intensive. The load pretty much stays at 0 most of the time. The VPS itself is rock stable — no reboots, no downtime, the network is always up. Basically I got nothing to complain about.

As I am only serving the Javascript, CSS and small images, I used around 30GB last month so I have not yet had a chance to test the limit. Bandwidth wise it seems to be buying from the Wholesale Communication Group, and I am getting Teleglobe.net for international traffic and Optus for Australia-wide traffic. Bandwidth also seems to be capped at 10Mbps which is fine for me. From a work server I had at Optus DC at Ultimo, I can fully saturated the 10Mbps link. However from both my home and work ADSL2+ link I can’t even sustain 1Mbps, for some reason that I have not yet investigated.

Overall it is pretty much “set and forget” and the server has no problem handling the traffic. However I am still hesitated to bring my site back to Australia (after moving it to US earlier this year), as I think I still trust Linode more if a disaster strikes.

However if you have high bandwidth requirement in Australia, I doubt you can find anyone cheaper.

Xen Hosting Plans

VPS Media uses Xen to power its virtual private servers. Currently LxLab’s HyperVM is used for server management, however a custom control panel is in the works. The VPS plans range from 384MB of memory for $20/month, to 2GB for $140/month. Servers are available in either Miami with Vault Networks or San Francisco with ColoServe (currently full).

The plan I tested with is their “The Starter” plan with

• 384MB RAM + 512MB swap partition
• 15GB RAID10 storage
• 150GB/month data transfer

Sign up was easy, and the VPS was instantly provisioned (probably some hooks between WHMCS and HyperVM). Not really sure about the fraud detections in process — as it seems to be the biggest excuse for manually provisioning the VPS. However I can’t see why my sign up request should ring the bell when I was clearly coming from an Australian IP address, has provided a valid AU address, and used Carlos’ special coupon code so credit card wasn’t even needed :)

I started with Ubuntu 8.04 32bit and then reinstalled CentOS 5.1. Thanks to HyperVM, rebuilding the node is trivial and there’s no need of intervention from their support staffs. cat /proc/cpuinfo shows that my node is hosted on a dual core, dual CPU AMD Opteron box running at 2Ghz. Not the state of art processor at the moment (Opteron 2212 — I think it was released in late 2006), but even a low-end plan can burst to all 4 CPU cores.

Pricing wise $20/month is very good for an unmanaged 384MB Xen — it offers more memory than SliceHost (256MB) and Linode (360MB) at this price point. In fact I think VPS Media is indeed targeting directly to what SliceHost and Linode are targeting — web application developers (I’ll talk a bit more about that later). When VPS Media first launched early this year it has only 256MB for its Starter Plan, but has increased the offering in early June to be more competitive.

As its California data centre is currently full, I probably cannot recommend them to my Australian colleagues who would like to have minimum ping from US West Coast. However if you are in US or in Europe, VPS Media provides good unmanaged VPS solutions at great price.

Server Performance

I have already stated that the raw CPU performance of the physical server is not great. I have actually had a few attempts to run the Unix Benchmark 4.1 on that node and the result is a bit disappointing — it seems the older Opterons are just no match to the current generation of Xeons. Performance seems to be stable, i.e. the benchmark yields similar result during different time of the day. Do note that from their hardware page they stated that they are using both Xeon and Opteron servers, and it could be just me who got randomly dropped to this box whereas there might be other more recent boxes around. YMMV.

Do note that “being able to run Hardcore Game Servers” like Counter Strike and Team Fortress 2 is one of the points listed under Why VPS Media. Many VPS providers actually do no like to host game servers because (1) high CPU requirement especially with recent games (2) potential target for DDoS. I am sure VPS Media has thought through all the possible short-falls, but potential customers might still need to be aware that there might be a FPS game sever running in the neighbourhood.

Disk is supposed to be RAID 10 — which means at least 4 disks forming mirrored stripped images. The result of running hdparm and seeker:

# hdparm -tT /dev/sda1

/dev/sda1:
 Timing cached reads:   3660 MB in  2.00 seconds = 1831.90 MB/sec
 Timing buffered disk reads:  422 MB in  3.01 seconds = 140.13 MB/sec

# ./seeker /dev/sda1
Seeker v2.0, 2007-01-15, http://www.linuxinsight.com/how_fast_is_your_disk.html
Benchmarking /dev/sda1 [15360MB], wait 30 seconds.............................
Results: 123 seeks/second, 8.11 ms random access time

Again, the performance is satisfactory but not excellent. I could have been on a busy node. However thanks to the isolation provided by Xen, I have no idea who I am sharing the server with.

Support and Community

From my past reviews of VPS companies I have always stressed the importance of building a community around your customers. I am glad to see that VPS Media is no short of online resources for its communities! Here is the list:

• Blog — WordPress powered blog that is used mainly for announcements.
• Forums — vBulletin powered support forums where members can ask question and show off their work. Service status is also posted here.
• Articles — again WordPress powered site that contains a growing list of how-tos on setting up servers.

Support is provided via either the support forum or WHMCS-based support software. Their contact page also contains toll-free phone number, although I am not sure how much response you’ll get during an emergency (obviously I have never tested).

I can’t seem to find whether real-time chat is provided as a support option. Personally I found being able to message someone on MSN or Jabber is more useful than the iteration of submitting the tickets and waiting for replies. However I also understand that tickets and email-based support is much preferred from administration’s point of view because you want everything to be logged and ID’ed (at work I won’t work on a patch unless there’s a ticket associated with the bug already). Hint: Maybe client management software like WHMCS can automatically convert a related chat messages into a ticket against a client for archiving purpose.

Targeted Market — Designers and Developers

VPS Media markets itself as the VPS solution for “Designers” and “Developers”. Whatever it means, VPS Media is probably not for those who need lots of hand-holding as only unmanaged services are provided. As I have stated before — it is targeting the niche where Linode and SliceHost are targeting — those who have no problem managing their servers and use alternate stacks to run their Django, Ruby on Rail, Seaside or ErlyWeb applications.

I am not a designer (and the template of this blog shows clear evidence) so I cannot speak for them that whether VPS Media is really suitable for the designers. They do have very nice web site(s) though, with integrated design all the way through not just their forums and blogs, but also WHMCS and HyperVM control panel.

I am a developer though, and VPS Media gives what a developer needs — a Xen VPS with a root shell (which is actually not asking much). What VPS Media lacks is becoming a “recommended host” from a developer community (like SliceHost with Rails). I guess it depends on what VPS Media is cooking with their next generation control panel that is set to replace HyperVM.

Short Interview with VPS Media

Well. To get a better perspective on the internal working of VPS Media, I actually emailed Carlos a short interview. In fact I was trying to sus out a bit more details on their new control panel but Carlos has kept the secret pretty tight. I guess we’ll see when it is released.

Here’s the full interview.

HostingFu: So Carlos, can you tell us a bit of background behind VPS Media? For example why did you choose to start a VPS hosting company?

Carlos: You see Scott, when our company started back in ’01, our services were geared towards a select group of companies who needed hosting and maintenance of their ERP (oracle), CRM (siebel & dynamics) and Enterprise Email systems (exchange). After a few years of working with high-availability setups of ESX, and Xen for our customers, we decided it was time to bring this ‘Business’ service, down to the people who need it most.

HostingFu: Is there a particular reason why you chose Xen instead of other virtualization technology?

Carlos: Definitely, Xen rules ;). No but seriously; its Isolation. We really like the hard-caps on each VPS, and the guarantee that your neighbor won’t take your VPS down.

HostingFu: Currently VPS Media uses WHMCS for client management and HyperVM for VPS management, however I heard that you have plans to roll your own customized panel. Can you tell us a bit more about it? What technology/programming language will you be using?

Carlos: Ah! our new system it’s still in the works, and currently testing the new features as I type ;). Lets make it a surprise.

HostingFu: How do you plan to build VPS Media in the next 12 months? What do you envision it to be?

Carlos: We envision VPS Media as a great resource for Designers & Developers who want full control. Who want to be able to load their own kernel modules, who want to customize their setup, and who want a real system. Also, as part of a big community in which developers and designers alike, come together and participate helping each other.

Thanks to Carlos for answering all the questions (and kept the part that I, a software engineer, am interested in, a surprise :)

Conclusion

VPS Media is still a very young company that is growing its customer base, and is still going through a lot of changes. I am generally happy with the service they have provided, and would not hesitate to recommend them if you are looking for an unmanaged Xen VPS server around US East Coast.

However they also face a lot of stiff competitions in their targeted market (developers looking for unmanaged Xen server). Around East Coast there are also Linode and Panix and many others. All the best to Carlos and the team, that they will be able to build something unique.

Crucial Paradigm is a web hosting/web design company in Sydney, Australia that have both Australian and US based operations (just realised that their office is around 400 metres to where I work). They have been in business for 5 years now, and early this year they launched their virtual dedicated servers service, providing fully managed Xen VPS hosted at the Equinix data centre in Sydney (which is on the same road where I live!)

They have recently re-launched their Xen VPS service. Instead of being fully managed, Crucial Paradigm is now offering them unmanaged with full-management option at extra A$50/month. That has driven down the cost of their base packages. The VPS I am testing now, which is the cheapest on their plan page, has

• 256MB memory + 256MB swap partition
• 25GB RAID-10 storage
• 10GB data transfer/month

All that for AUD$29/month, which makes it one of the cheapest virtual dedicated server plans in Australia. I know 10GB of data transfer per month is nothing and some sites can zap through that in a day. AUD$3.50 per excessive GB of transfer is not particular cheap either — but then again this is Australia we are talking about, unfortunately.

Signing Up and Deployment

You can sign up a Xen VPS account from the plan page, which is integrated with their WHMCS hosting billing system (why oh why do you want a domain name for unmanaged VPS service?!) Anyway, signing up is a breeze after putting in example.com. I think all sign ups are manually provisioned so there is no instant VPS deployment (although I think WHMCS does have HyperVM provisioning module). My VPS was deployed within an hour anyway as Crucial Paradigm has staff on support 24/7. Pretty good so far.

Note that at the point of signing up, you can choose from a list of available Linux distributions — CentOS 4/5, Debian 3/4, Ubuntu 7.10/8.04, Gentoo 2006 and Fedora Core 6. You can also choose a control panel, some with extra cost — cPanel, DirectAdmin and LxAdmin. More about this later — but for a start I chose a Debian 4 template.

CPU & I/O Performance

According to Crucial Paradigm’s VPS page, they have pretty impressive physical hardware running the dom0. At the time of writing they are deploying new VPS onto hardware with the following spec:

• 2 x Xeon Quad Core 5410 (8 x 2.33Ghz, 24MB Cache)
• 32GB DDR2 FB-DIMM RAM
• 8+ Enterprise Drives RAID-10
• Dual Redundant Power Supplies

While the physical server has access to 8 CPU cores, the low-end 256MB VPS I’ve got has only access to one according to /proc/cpuinfo. I think higher plans might have access to more cores. It is not a big issue for me as the bottleneck of most websites is usually not on CPU (if the scripts and database have been tuned properly). Disk I/O performance is usually more critical.

I used methods detailed in this LinuxInsignt article to measure the linear read and seek performance of the hard disk (tested at 9:30pm on Sunday):

# hdparm -tT /dev/sda1

/dev/sda1:
 Timing cached reads:   23060 MB in  2.00 seconds = 11555.72 MB/sec
 Timing buffered disk reads:  948 MB in  3.01 seconds = 315.39 MB/sec

# seeker /dev/sda1
Seeker v2.0, 2007-01-15, http://www.linuxinsight.com/how_fast_is_your_disk.html
Benchmarking /dev/sda1 [25000MB], wait 30 seconds..............................
Results: 141 seeks/second, 7.05 ms random access time

Result is pretty good. However I do personally find “performance testing” under the VPS environment is a bit less than useful because it all depends on how much resource your neighbours are using at the same time. Take disk read test for example, the result is different everytime, although it is usually between 200MB/sec to 315MB/sec.

I am happy to say that Crucial Paradigm is offering a very solid virtual dedicated server.

Network Connectivity

Crucial Paradigm’s Xen VPS is connected to a 100mbps port with pretty good connectivity. First of all it peers with PIPE networks which means pretty good speed for many broadband users in NSW. I also get quick downloads from inter-state mirrors. For example 3.5MB/sec from mirror.aarnet.edu.au in Queensland.

Best of all, it has very sweet latency for NSW users. 20ms average from my ADSL-connected home to my testing VPS, consider the first hoop is already 18 milliseconds!

Revolutionary?

There are now over 30+ hosting companies providing virtual private/dedicated servers in Australia. The question is — what sets Crucial Paradigm apart?

• Xen VPS. You’ll find most Australian VPS providers go with Virtuozzo-based service, and those that use Xen are few, although Xen itself is free (whereas Virtuozzo costs). I have nothing against Parallels but I do personally prefer Xen as it is more dedicated-like.

  Check my previous article on Xen vs. OpenVZ on why I prefer one rather than the other.

• Affordable entry cost. What attracts me first to VPS is its price tag — it costs way less than a full brown dedicated server and yet it feels and behaves like one. Looking at the list of providers in Australia, not many of them can offer you a VPS for less than AUD$30 per month, so that Aussie Linux geeks can have a “cheap root” somewhere Down Under.

  Crucial Paradigm, GPLHost, and Labyrinth Data are the only ones I can find so far. Interestingly they are all Xen providers. Crucial Paradigm and Labyrinth Data’s low-end plans are pretty much on-par. CP’s one has more memory and offers plans scale all the way to 4GB RAM, but LD’s has more disk space, free WAIX traffic, etc.

• HyperVM VPS Management. I won’t say HyperVM is revolutionary, but it is indeed unique amongst Australian VPS providers. If you check out the VPS hosting offers section at WHT, you will find around 1/2 of the new providers are using HyperVM and OpenVZ. The reason is simple — you don’t have to code your own VPS management suite and at 50 cents/VPS/month it is probably the most economical solution for new providers.

  HyperVM is actually not too bad. I do prefer custom panels from SliceHost and Linode and found HyperVM too cluttered (icons everywhere giving you too much information). However it is powerful and it gives customers everything to shoot themselves in the foot. You can manage multiple nodes from a single installation.

  Why there’s no other Australian providers offering HyperVM-based solution is really beyond me. At least Crucial Paradigm is taking the lead here.

Another point related to HyperVM is LxAdmin, LxLabs’ light weight web hosting control panel that is free with HyperVM-powered VPS, that I have reviewed here. LxAdmin has evolved since I last reviewed it last June. Instead of 100 free domains you now only get to host 40 domains, but it has also added a lot more functionality that makes it a more complete end-to-end web hosting package. It even integrates with WHMCS that Crucial Paradigm says is “free” on their plan page.

Why am I bringing LxAdmin up? In fact after a few hours my Debian VPS was up, I wiped it clean and installed CentOS 5 + LxAdmin template. The LxAdmin template has everything you need to get web hosting up and running — web server, DNS server, mail server, IMAP server, control panel, etc. Basically you can start hosting your sites or your clients’ sites straight away — why needs cPanel and DirectAdmin? You might swap out Lighttpd for Apache though for compatibility, and 256MB RAM is more than sufficient to run Apache + LxAdmin.

This is again something I have been wondering, as due to abundant budget VPS over there in US, LxAdmin/Host-In-A-Box has raised in popularity due to its cost and memory usage. No one in Australia seems to care about it but I am glad that Crucial Paradigm brought it in as the low-cost solution to run your own small hosting shop.

Conclusion

After playing around with a testing VPS for a few days, I can say that Crucial Paradigm has launched some very solid Xen VPS packages. While VPS providers are getting crowded in Australia, Crucial Paradigm managed to differentiate itself from the competitions by providing a low entry-level price tag plus a powerful server management control panel (HyperVM). By offering a free web hosting control panel (LxAdmin) and a free billing system (WHMCS), I think this VPS will be very appealing to small hosting shops or web design/publishing shops.

Will I host with Crucial Paradigm? Maybe, but definitely not at this stage. Bandwidth in Australia is still way too expensive. Unless you are a big bandwidth buyer, you usually have to pay $3-$4 per gigabyte transferred. Crucial Paradigm also lacks a community feeling on its site — where do their customers hang out?! I guess most individual/small business customers won’t care, but as a developer I do look out for signs of forums and active blogs to see how lively the community is (which usually reflects how good the service is).

All the best to Aaron and team.

Web24 — Virtuozzo VPS in Australia

Back in early this year I talked about writing a review on Web24.com.au, which I used to replace a VPS I got from GPLHost, which I terminated last December as I was using too much bandwidth (and was too cheap to pay :) The VPS I got from Web24 was their Silver package, a Virtuzzo Linux VPS running Ubuntu Linux, with 384MB guaranteed memory, over 6+GB of privvmpages (burstable memory), 50GB/month data transfer, and was located in Fujitsu data centre in Melbourne (see their profile on VPSAU). All these for under AUD$50/month — very affordable for my little sites.

Great Network, Great Uptime, Great Customer Service

Other than the Web24-wide connectivity issue back in February, their network has been rock solid. I am getting around 35ms from my home ADSL2+ to my VPS, and have used 30-40GB a month without issue.

My VPS there also had great uptime. It was up for the entire period — almost 120 days without rebooting. It has certainly demonstrated the stability of Virtuozzo. Hardware wise Web24 was using a 2.4Ghz Xeon with 4x physical core and 8GB of physical memory. However, cpulimit of my VPS has been set to 160% so I can only use around 960Mhz of each CPU core (and 4 cores at the same time), which is not bad at all. Web applications are rarely constrained by CPU power anyway.

Customer service wise — top notch, all the way from pre-sale to support. I have sent in a few support tickets and all of them have been answered in timely manner. GPLHost still has a better service though in my opinion because they are not only solving your problems, reporting back their progress, but also following up after the issues have been resolved.

Then There was Performance Issue

The main reason I am moving away from Web24 is performance, or lack of it. As I have said previously that the server CPUs have more than enough power, but my website over there has suffered much degraded performance throughout February and March, due to large amount of I/O wait.

First of all, what am I hosting there? A small-medium sized Drupal-based community site with around 5,000-7,000 visits per day, 800+k page views and 7.5 million hits per month. Lots of customisation and optimisation, and lots of fragment caching. Lighty on the front-end with 5x PHP+xcache/FastCGI processes (which is more than enough). MySQL 5 with a fat key buffer and query_cache_size to optimise on read requests. I am also trying to stay way below the memory limitation just to be on the safe side. In fact I usually use no more than 200MB under privvmpages. And I thought I was safe.

Not really.

Unfortunately I am not able to reproduce all the communications between Web24 support and myself (as I have a bad habit of deleting old emails), but most support issues I raised were performance related.

Swapping Swapping and Constantly Swapping

20 December 2007 evening, my VPS suddenly stopped responding. Yes it still responded to ICMP pings. Yes you can still connect to port 80, but PHP/FastCGI backend won’t load. SSH in took forever, and once I was in I checked the loadavg it went skyrocket. Type in free -m and I was shocked to discover that all the swap space has been exhausted. free, buffers and cached are all on single digit. Note that on Virtuozzo 3 UBC this is from meminfo of the host server, and has nothing to do with my own VPS. The host server was breathlessly swapping to cope with memory shortage. I cannot remember how long it lasted but it did resolve itself at the end, probably with OOM and some process slaughtering. It wasn’t pretty.

So in the middle of my panicking, I fired a support request. It was escalated and a few hours later someone responded saying it was a CPU issue rather than memory issue, and was caused by someone else’s VPS getting compromised. The swap space on the host server then got increased from 2GB to 12GB. I was trying to argue that it WAS a memory/swapping issue, and my processes cannot get enough CPU time because everyone was busy paging in and out! Nope. Web24 denied about it. The problem was with the CPU they said, which was definitely not what vmstat told me.

I/O Wait Issue Got Worse

Performance irregularity continued, and got worse around February and March. During peak hours (which for my site is 9am-11pm AEST), the 15-minute load average might go up to 4-5 for a period of 5-10 minutes every now and then, and my site would completely stall. 15-minute load average doesn’t drop back down below 0.5 during peak hours, and simple commands line ls feels sluggish — a sure sign of I/O wait issue. Swap usage on the host server has been ranging from 4GB-6GB (out of 12GB total swap space), and sometimes some of my key processes got swapped out even though I am only using half the guaranteed memory.

So I posted this question at WebHostingTalk (without mentioning who the provider was), and asked for opinions.

The question I would like to ask is — is it possible to pin-point why the VPS is slow? My apps are obviously not CPU bounded, and as iostat is not really working in Virtuozzo I cannot tell whether it is IO-bounded inside my VPS either. Is it possible to find out whether it is due to excessive paging on the physical server, so I can go to my provider and say, “hey you are overselling and you should not pack that many VE into a physical server”.

And the responses I’ve got:

… when a VPS node is using ANY swap at all that’s your first warning sign that the server is being pushed to its extremes.
– seankoons at Zone.net

My take: It’s vastly oversold…
– TheWiseOne at TekTonic

If 5GB of SWAP is being used, I’d say they are overselling…
– devonblzx at Reseller3k

Well. I guess I got te message. Despite claims that “We do not overcommit on our VPS infrastructure”, looks like Web24 has just jammed too many people onto that host node. Or maybe my hypothesis with UBC-based VPS hosting is true — by providing 6+GB of burstable memory, you are basically inviting everyone to use as much as they want, and OOM won’t kick in until swap space has been exhausted — which would be probably too late. In that case they might not be overselling if they calculated with the amount of guaranteed memory, but poor UBC settings can still make you look bad. Not that I am going to be hosted with another UBC-based VPS provider anyway.

Moved to Linode

As traffic to my website has also been slowly growing, I was facing two choices — move up to the next plan, or move elsewhere (again). With the performance issues at Web24, I do not think moving up to the 512MB guaranteed plan will do any wonder. AUD$50/month is all I am willing to pay anyway, which means I won’t be able to find any virtual servers in Australia that will fit in my budget. With over 85% of traffic coming from Down Under, it makes sense to also host my site in Australia. But when budget becomes a constraint — well I guess I can live with a bit of latency.

As I still kept a Xen VPS at Linode after my review in January (yes, I actually became a customer), it is at Fremont CA which is “close enough” to Australia for me, I decided to make a migration from Web24 to Linode at the end of March.

So one evening I changed the DNS TTL to 15 minutes, put up a maintenance notice on my website, copy all the files across (which was less than 200MB), set everything up at Linode, make sure everything works, and update the DNS records to let them propagate. The next morning — traffic as usual, and everything “just works”. You do feel the lag typing inside a SSH session, but you cannot actually tell much difference with page serving. I guess the poor performance at Web24 sort of cancels out the benefit of low latency.

My community site has now been running for a month now at Linode and the performance has been great. In April it has been 8,000+ visitors/day and used around 62GB of data. Linode Platform Manager shows 4% of average CPU utilisation last month, and my Linode 360 uses little swap and always has 200+MB of (free + buffers + cached). 15 minute loadavg rarely goes above 0.1… 200GB/month of data transfer means it will probably last me a while.

Verdict

A few concluding points:

• Don’t always believe what the support says. Gather your own evidence. Even a vmstat dump can point out roughly where the bottleneck is.

• Xen >> Virtuozzo. Yes I know Matt is advertising here, but personally I still much prefer Xen than Virtuozzo. At least you get a virtualised block device where iostat can tell you how much I/O you are doing.

• When you have high load on a Virtuozzo/OpenVZ VPS, adding more memory have no value by itself (despite many have suggested this at WHT). If you are CPU-bound (not likely) — check whether there’s a cpulimit on your VPS and ask your provider to remove it. If you are IO-bound (very likely), then you can go and get more memory PLUS implement aggressive caching throughout the stack so hopefully there is less load on the database, if you are the trouble maker.

  If you are not the one causing the problem, or maybe optimisation is not your cup of tea, then prepare to jump ship.

• Oversold VPS exists.

It is basically preaching the same message that DreamHost blogged about almost 2 years ago. None of these web hosting review sites are reliable because top spots can usually be bought with monthly subscription fee. Either that or a hefty affiliation payout. Basically they should have renamed the top 10 hosting companies to the Top 10 Web Hosting Companies that Pay Me The Most. Funny that the site is hosted on GoDaddy, which quite often appears on those “top web hosting site” lists as well.

What About Paid Reviews?

I guess the same issue goes to many web hosting related articles/blog site as well, where sometimes it can be difficult to tell whether a review on a blog is genuine or not. Any money changed hands for this particular post? Are the links full of affiliation IDs? I was reading Daw Web Hosting Blog and Dimitar has pointed out an increasing number of paid-review blogs on web hosting, and those paid reviews can usually mislead consumers over purchasing decisions, as many people search for “<company name> review” thinking they got postive reviews from genuine customers.

I am not saying I am sinless in this department. I did a paid review a year ago on a UK hosting company when I tried out ReviewMe (although I can’t see any sale pitch in that article). As Dimitar said, I think doing those paid reviews might hurt the credibility of this site in the long run. I have since cancelled my ReviewMe account, and turned down many review requests — most are generic oversellers that offer nothing unique anyway.

Any Genuine Reviews Out There?

With paid listing and paid reviews littering the net, I am afraid to say that it can be almost impossible to find genuine reviews. Most website owners don’t write about them when things are smooth sailing so all the positive reviews ended up coming from the paid review sites.

One alternative is to search through Web Hosting Talk forums (where all the disgruntled customers whinged about their web hosting companies). Again, lots of negative reviews and sometimes I doubt how genuine some of the positive reviews are over there.

I guess there is no “The Best Host” where everyone raves about unconditionally. Do a bit of research making sure there are not too many negative reviews. Give them a try without committing too much up front. Make sure you do not purchase domain from the same place where you purchase hosting accounts. And this is important — do daily backups so you can move over to the next host if anything does go wrong…

1. Provisioning and Setting Up
2. CPU — Severely Limited
3. Support
4. Conclusion

Last week I needed a small VPS to test a new project so I went to Web Hosting Talk VPS offers forum to look for something really cheap, and found this deal at ServerWays.com. ServerWays is a new company, with its domain registered in December 2007. It is also a one-man shop according to this discussion thread. However it does not really matter to me as (1) it is only less than $5 per month (2) it is for a throw away project. So I clicked on the Order button for their lowest spec’ed VPS 111.

This is what you get for $4.90/month, billed monthly and no set up fee.

• 100MB guaranteed memory
• 10GB storage
• 100GB/month transfer
• OpenVZ + HyperVM

Provisioning and Setting Up

I submitted an order at around 22:30 -700 and got an notification pretty much straight away. It was set up at 1:40 -700 so 3 hours turn around setting up new accounts at odd hour of the day — not too bad. I ordered a Debian 4 Etch minimum, and then installed a basic PHP stack for some quick test.

• MySQL 5.0.32 — default set up with query cache and innode DB turned off
• PHP 5.2.0 — 2x FastCGI slaves
• Lighttpd 1.4.13

A few observations while setting up this low end box.

• Network is Cogent, which explains why it is so cheap.

• Throughput of network is actually pretty good. I can download big Debian packages at around 500+ Kbps from US mirrors. Latency is not too bad either as the server is located in Portland Oregon. 190ms from my Sydney ADSL connection.

• There is no “burstable” memory. 25,600 privvmpages — exactly 100MB that can be allocated by all processes. However, a simple stack like above + Exim4 + cron + syslog-ng + xinetd comes down to around only 35MB used.

• Physical server is an old Pentium D 3.0Ghz. It is dual core, but your VPS only has access to only one of its CPU cores.

• Most user_beancounters values are very reasonable. I guess these were set up by HyperVM.

• meminfo has been virtualised so there is no way to find out the memory utilisation on the physical node.

It is all good so far. Until I started doing some test.

CPU — Severely Limited

So I have pretty much everything needed to serve up dynamic PHP content. I then checked out the latest WordPress from subversion, and started to install a new blog.

And it was frustratingly slow!!

An out-of-box WordPress front page having only one single “Hello World” post will take around 5 seconds to render (compare to around 0.3 second on my SliceHost VPS, and 0.8 second on my DreamHost account). Everytime you click on a link, you have to wait for 5 seconds for the page to be generated, and you can see the load went straight up with php-cgi as the most CPU-consuming process. However, there is something wrong –

It is only using less than 4% of CPU time!!

Basically throughout the 5 seconds it tried to generate the pages, PHP interpretor process exhausted all the CPU time this VPS is allowed to have, which is less than 4% of a single out-dated NetBurst core! It can be easily verified — just create a simple busy loop program in C, run it, and watch it consuming all the CPU resources you got — no more than 4% of the CPU time!

There might be a few possible explanations:

• ServerWays has an overcrowded server where everyone is fighting for the CPU time
• My VPS has been throttled with limited CPU time

To me it is more of later than former as the maximum CPU time my processes can get seems to be fixed at always less than 4%. Looking at /proc/cpuinfo reveals that the CPU has indeed been capped at 114Mhz. It appears that cpulimit has been used so my “burstable CPU” can never exceed 4% of a CPU core!

That seems to be defeating the purpose of virtualisation, where you get burstable CPU time to improve CPU utilisation, doesn’t it? So I opened a support ticket for clarification.

Support

It is also a great opportunity to test the support response. ServerWays uses WHM Complete as billing/support portal so I posted the following support ticket at 018:30 -700:

Hi ServerWays,

I have recently acquired a VPS 111 (great price, btw) to do some testing, and noted that it seems to be *limited* to 100Mhz, instead of guaranteed with 100Mhz, i.e. cpulimit instead of cpuunit in OpenVZ VE parameters. It is a bit frustrating seeing even a process with a spinning loop can use no more than 4% of CPU. With a testing WordPress install it is still taking around 5 seconds to generate each page because CPU utilisation has been throttled. Is that intentional with this plan?

Response came back 6 hours later. I won’t post the verbatim response here, but basically:

• Yes it is intentional, setting provided by HyperVM.
• You need to upgrade plans to get more CPU time.

Hmm. All right. Even if I upgraded to their VPS 555 plan with 500MB privvmpages, 50GB storage and 500GB monthly transfer for only $12.90, the CPU will still be limited at 500Mhz. The WordPress page generation time might be reduced by one fifth to 1 second per page, it is still far more than what I have expected.

Conclusion

There is no doubt ServerWays.com provides cheap low-end VPS. $4.90/month is even cheaper than many shared hosting accounts. 100MB burstable memory is plenty with a lighttpd/MySQL/PHP stack to run low traffic sites. And you have root to install anything you want.

However, its CPU limit has drastically reduced its usefulness. Forget about using it to run any dynamic site unless you are training your users the virtue of patience. For me, it is $4.90 wasted.

1. Update September 2008
2. Expecting UML, Got Xen
3. Linode Platform Manager
   1. Initialisation and Booting up
   2. Disk Images
4. Configuration Profiles
5. Finnix Recovery Distro
6. Linode Shell
7. Linode Community
8. But, What About Reviewing the Actual VPS?!
9. Conclusion

When I first went shopping around for VPS hosting back in early 2006, Linode was high on my list. There were only a few Linux VPS providers back then but I knew Linode has been in the VPS/VDS hosting business for years, way before the flood of HyperVM/OpenVZ hosts that I have observed in the recent months. However I did not go with them in the end (but went with Unixshell instead) because (1) they were a bit more expensive (2) I heard User Mode Linux (virtualisation used by Linode) was not as fast as Xen.

Fast forward 2 years.

A few days past Christmas, Thomas from Linode offered me a VPS account to review, and I emailed back asking for a Linode 360 in their Fremont California rack. 2 hours later my account has been set up, and I was already inside their Linode Platform Manager configuring up my VPS. So far so good.

Expecting UML, Got Xen

When Thomas contacted me about reviewing their VPS, I was expecting to get a UML node but instead, Linode has set me up on one of their new Xen boxes, which they are currently beta-testing. As I have no experience with UML I cannot make any claim on how it works, but according to what I have read they are similar to Xen where each guest OS gets a dedicated-server-like environment (swap partitions, actual RAM allocation, full-blown kernel, etc). However,

• Xen is generally faster.
• Xen is para-virtualisation with dedicated resources, where a UML is just a process in userland that can potentially be paged out.
• Xen gets SMP support, which is very useful with today’s multi-core CPUs.
• Xen is on the news everywhere.

UML does have a few advantages.

• More mature and has good track record
• UML token Bucket IO Limiter developed by Linode that prevents any VPS from overtaking the entire IO bandwidth on the host node.

I am perfectly fine with trying out Xen, especially when Linode has developed their Platform Manager control panel to work on both Xen and UML.

Linode Platform Manager

One thing that really stands out from Linode is their Linode Platform Manager, which is a very well designed and flexible control panel for Linode end users to configure and deploy their new virtual servers. You can read about all its features here. While it is flexible, it can also be confusing sometimes, especially when you came from other VPS providers where you are expecting just a SSH prompt after you submit your credit card details.

Initialisation and Booting up

There is a excellent video on the features page on how you can set up your Linode VPS for the first time. Basically after you logged into the Platform Manager, you need to

1. Choose between one of free data centres if it is your first time logging in (Atlanta, Dallas or Fremont).

2. Go to Distro Wizard to configure your new VPS.

   

   1. Select from a wide range of Linux distributions such as Arch Linux, CentOS, Debian, Fedora Core, Gentoo, Mandrake, OpenSUSE, Slackware and Ubuntu.

      Note that they are all bare-bone distributions and they do not have ready-to-run LAMP or RoR distributions like SliceHost or VPSLink. You are responsible for getting all the software installed yourself as Linode provides strictly unmanaged service.

   2. Select disk image size (for your main partition). It is defaulted to your plan’s size – swap size.

   3. Select swap partition size.

   4. Choose your root password.

   When you click on “Create Profile”, it creates (1) a configuration profile (2) a main disk image initialised with the Linux distro of your choice (3) another disk image containing the swap partition if you choose to have one. It might take 1-2 minutes depending on the size of your distribution.

3. Go back to Dashboard, and hopefully all the profile creation and image initialisation jobs have been completed. All your configuration profiles and disk images will be listed here, with the “Boot” button next to a bootable profile.

   

   Click on “Boot”, a job will be dispatched to the physical server hosting your VPS, which should be executed in no time. Now your new VPS is up and running, and only at this point we can SSH into it. You can find out the public IP address under the Remote Access tab.

Most VPS users will just stop here, focus on building their shinning new VPS and never visit the Platform Manager again. However, Platform Manager is far more powerful than just initialising your disk images and rebooting your servers. This review does it no justice if we just stop here.

Disk Images

If your VPS uses OpenVZ, Virtuozzo or Linux-VServer, then it’s too bad — you do not have a choice on how your partition is configured. There is no choice on multiple partitions, no choice on the file system to use and how file system is formatted (for example, how many number of kilobytes per inode). If you are using Xen, then your VPS is fully capable of doing that, but you might not have an option from your VPS provider.

On Linode, you can

• Create multiple disk images as long as the total size does not exceed your allocated size.

• Choose the disk image size when they are created.

• Choose the default file system on these images. The default choices are ext2 and ext3, but I think you can create them as raw images and then format them yourself from your own VPS. These are the file systems supported by my Xen VPS kernel:

  # cat /proc/filesystems | grep -v nodev
          reiserfs
          ext3
          ext2
          cramfs
          minix
          msdos
          vfat
          iso9660
          romfs
          fuseblk
          udf
          jfs
          xfs
  

To create extra disk images, click on “Create a new Disk Image” from dashboard.

Unfortunately there does not seem to be a way to prepare your root partition with anything but ext3. Not that it is going to be an issue as ext3 has been proved to be a rock solid file system, but depending on your applications, ResierFS or XFS might be more suitable for the job.

The advantage of choosing ext2/3 is that you will be able to resize the disk image from inside the Linode Platform Manager, if you happen to upgrade your plan or buy more space later on.

Configuration Profiles

After creating disk images, you can now change your configuration profile to attach them to specific devices when your VPS boots up. However, there is much more to configuration profiles than just mapping disk images to devices.

It sets up all sorts of boot parameters for your Linux VPS, including:

• Kernel to boot from (choosing different 2.6 kernels and I think you can even choose 2.4 kernels for UML)

• Memory allocated to VPS (although I can’t see why you’ll set anything less than maximum)

• Change run-level (default, single user or set init=/bin/bash). Useful when the VPS is really screwed up.

• Root device and whether it is mounted r/o or r/w upon booting up.

• Some misc helper operations to prepare the disk images before booting.

How useful are these configuration options? Probably not much when your VPS is up 24×7 serving content. However I can see them being extremely useful when you are trying to recover from a corrupted file system, or you like to play around with multiple Linux distributions.

Finnix Recovery Distro

On the topic of recovering an unbootable Linux box, I used to have to hunt down the LiveCD, attach a monitor, reconfigure the BIOS so it reboots into the CD-ROM. With the number of hard disk failures these days in data centres, I wonder how often an average NOC has to perform these operations to get their clients’ cheap dedicated server back to life.

On Linode, booting into a LiveCD is a breeze. Follow these instructions to create a new recovery profile that boots from Finnix, a small rescue distribution.

Great feature, but something you wish there is never a need to use it :)

Linode Shell

Another unique feature is their Linode Shell, where you get out-of-band console access to your VPS, which is useful when you cannot login to your server from network (for example, bad iptables rule blocks out SSH access). Wait! There are many Xen and OpenVZ providers that also give you console access! But Lish is a bit different.

• Your console access runs inside GNU screen, and you can detach from screen to reveal Lish command prompt.

• You can reboot/shutdown your VPS from Lish. You can also choose different configuration profile to boot from.

• You can check jobs queued and performed on your VPS.

It is like a mini-Linode Platform Manager for those who stuck on those old Wyse-60 terminals :)

Linode Community

Like SliceHost that I am currently using, Linode also has a great list of community resources — forums, IRC chat, Wiki, blog aggregator (planet) and “pastbin“.

An active community is a sign that the hosting provider is doing a good job. Not only it establishes two way communication between the provider and its users, an active community also shows it has enough “fans” willing to participate and help each other out.

Obviously Linode has done something right here…

But, What About Reviewing the Actual VPS?!

Yes, I have realised that the focus of this review has been on the Linode Platform Manager, as it is what’s unique about Linode. What about its Xen VPS? It is just like all the other well-performing Xen VPS. My server can burst to a quad-core Xeon L5335 “Clovertown” at 2Ghz, disk IO has been very fast (around 2.3GB/sec if I read from disk cache — some serious setup there), and network connectivity has been great (>2MB/sec downloading from Amazon S3 and 185ms ping from my home in Sydney Australia). However it is difficult to perform some real test without putting a live site with moderate traffic on it over a longer period of time.

I guess my point is — running stable and fast Xen VPS on server-grade hardware is no longer rocket science, and that is probably one of the reasons why Linode came back to beta-test Xen again. What really sets Linode apart is its Platform Manager, its active community and its years of experience in the Linux VPS industry.

And the price? Remember that I was turned off from their price 2 years months ago. Now you can get their base package, Linode 360, for USD$19.95/month, which gives you:

• 360MB RAM
• 10GB space
• 200GB/month data transfer

Great price indeed.

Conclusion

I am quite happy with the Xen VPS Linode has provided, and I think I know where I will get my next VPS from. ’nuff said :)