96MB Low End VPS Review Part IX – FST Servers
Joe from FST Servers left a comment on one of my previous reviews stating that they would like me to write a review about them. I have never heard about this company before – no reviews, no history, but the fact that they are using Citrix XenServer is something that makes them a really worthy one to be reviewed for.
General Information and Prices
There is really limited information I can find about FST Servers, through the domain checks, their domain is registered to Brendan Fusco in Chicago Illinois and it seems to be where their servers are based as well according to this post on LowEndBox.com. As per this post on Webhostingtalk, they have decent processing power on the hardware nodes, as well as completely dedicated resources with nothing “burstable”.
Prices offered are definitely in the low end of the budget spectrum, especially for XenServer platform, with the lowest priced plan starting at $4 per month (with further discount if you pay annually), which is definitely the cheapest I have seen with Chicago servers using such platform. Their additional IP addresses are also relatively cheap at $1/month per IP address. Furthermore, because of the XenServer platform, you are essentially allowed to load any OS, including Windows 2003 and 2008 (although I would really doubt if 128MB of RAM is capable enough to support a Windows installation).
As the review was requested by FST Servers, the test box was set up without going through the regular sign up process and therefore can not comment on the time taken for the server to set up.
Custom Control Panel
The FST Control Panel is the name for their custom control panel, and is definitely one of the highlights of the review. For most of the XenServer providers, there is no control panel and you practically have to submit a ticket for everything from reloading the OS to recovery console access if clients are unable to access the VPS via the shell.
After login through HTTPs, here is how the control panel looks like:
As you can see, everything is on the same screen and is very straight-forward, which is definitely something that will be appealing to many clients.
On the main screen, the VPS could be reset, stopped and started with a click of a button, emergency recovery console is also available. Furthermore, adding reverse DNS can be done with a click of mouse as well.
Contact Support is pretty straight forward as well with four departments for technical support, sales, billing and other.
When clicking on the VPS, there are further configuration options that are available to the VPS:
From here, you can upgrade VM to a higher plan, as well as reloading OS, changing root password and hostname. It is interesting to note that in order to install CentOS, VM-2 is required (Windows as well, but that is hardly a surprise), since I am not aware that CentOS can’t run on 128MB on VPS.
The upgrade VM option does not allow upgrading according a certain feature of the VPS (i.e., upgrading only the RAM or the hard drive space), instead, you will need to upgrade to the next level of VM plan, making it impossible to create a customized VPS plan, which is perhaps the biggest distinction I can see between FST’s VPSes and those of the trending “cloud” hosts.
For their VM-0 VPS, the following OS templates are available:
which basically including the latest versions of Debian and Ubuntu. There is also a PBX in a Flash template, which would be useful for those interested in using the VPS box as a private phone exchange system.
When clicking on the second tab, creating VPS, which is essentially their order page, shows up:
It seems that new VPS could be created “within a few minutes” after the payment is submitted, which seems to hint that the provision of VPS is instant. Unfortunately, from what I can see, at this point, credit card seems to be the only payment method, which is perhaps not a very great piece of news to Paypal lovers.
FST Servers also has a detailed documentation about their control panel, which can be found here.
Overall, the custom control panel is pretty good, providing users a very compact but yet powerful interface to control the VPS, however, if CentOS could be permitted for VM-0 or VM-1 packages (at least a minimal template) and more payment gateways could be used, it would be really great.
Tests on the VPS
The test VPS that I have received is their lowest priced plan (VM-0), which include 128MB of RAM, 7GB of hard drive and 500GB per month of bandwidth. When the OS was first loaded up, only 8MB of RAM was used, which is impressive:
Linux 96mb 2.6.32-5-686-bigmem #1 SMP Wed May 18 07:33:52 UTC 2011 i686 GNU/Linux df -h Filesystem Size Used Avail Use% Mounted on /dev/mapper/216-root 6.5G 496M 5.7G 8% / tmpfs 61M 0 61M 0% /lib/init/rw udev 50M 64K 50M 1% /dev tmpfs 61M 0 61M 0% /dev/shm /dev/xvda1 228M 15M 202M 7% /boot free -m total used free shared buffers cached Mem: 121 59 62 0 2 47 -/+ buffers/cache: 8 112 Swap: 239 0 239
There is only one CPU core offered though, and is locked to 2.0GHz, which limits the processing power of the VPS:
cat /proc/cpuinfo processor : 0 vendor_id : AuthenticAMD cpu family : 16 model : 9 model name : AMD Opteron(tm) Processor 6128 stepping : 1 cpu MHz : 1999.998 cache size : 512 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 de tsc msr pae cx8 cmov pat clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt 3dnowext 3dnow constant_tsc up nonstop_tsc amd_dcm pni popcnt hypervisor cmp_legacy extapic cr8_legacy abm sse4a misalignsse 3dnowprefetch nodeid_msr bogomips : 3999.99 clflush size : 64 cache_alignment : 64 address sizes : 48 bits physical, 48 bits virtual power management: ts ttp tm stc 100mhzsteps hwpstate
Meminfo does not really offer anything fancy:
cat /proc/meminfo MemTotal: 124352 kB MemFree: 91564 kB Buffers: 3284 kB Cached: 19436 kB SwapCached: 156 kB Active: 23288 kB Inactive: 2560 kB Active(anon): 1360 kB Inactive(anon): 1836 kB Active(file): 21928 kB Inactive(file): 724 kB Unevictable: 0 kB Mlocked: 0 kB HighTotal: 0 kB HighFree: 0 kB LowTotal: 124352 kB LowFree: 91564 kB SwapTotal: 245752 kB SwapFree: 245596 kB Dirty: 0 kB Writeback: 0 kB AnonPages: 2988 kB Mapped: 3828 kB Shmem: 68 kB Slab: 4796 kB SReclaimable: 2756 kB SUnreclaim: 2040 kB KernelStack: 376 kB PageTables: 340 kB NFS_Unstable: 0 kB Bounce: 0 kB WritebackTmp: 0 kB CommitLimit: 307928 kB Committed_AS: 34284 kB VmallocTotal: 729080 kB VmallocUsed: 776 kB VmallocChunk: 728224 kB HardwareCorrupted: 0 kB HugePages_Total: 0 HugePages_Free: 0 HugePages_Rsvd: 0 HugePages_Surp: 0 Hugepagesize: 2048 kB DirectMap4k: 131072 kB DirectMap2M: 0 kB
VMStat shows the CPU is pretty light as well:
vmstat procs -----------memory---------- ---swap-- -----io---- -system-- ----cpu---- r b swpd free buff cache si so bi bo in cs us sy id wa 0 0 0 30468 3100 81252 0 0 1 1 10 2 0 0 100 0
Inode is set not exactly at a low range, but is somewhat reasonable:
df -i Filesystem Inodes IUsed IFree IUse% Mounted on /dev/mapper/216-root 428240 15375 412865 4% / tmpfs 15544 4 15540 1% /lib/init/rw udev 12677 457 12220 4% /dev tmpfs 15544 1 15543 1% /dev/shm /dev/xvda1 124496 222 124274 1% /boot
Even with a full LNMP stack installed, only 26MB of RAM is used, which is really good:
free -m total used free shared buffers cached Mem: 121 112 8 0 4 82 -/+ buffers/cache: 26 95 Swap: 239 5 234
And the output of top with the full LNMP stack:
top - 19:12:48 up 4 days, 23:55, 1 user, load average: 0.00, 0.00, 0.00 Tasks: 56 total, 1 running, 55 sleeping, 0 stopped, 0 zombie Cpu(s): 0.0%us, 0.0%sy, 0.0%ni,100.0%id, 0.0%wa, 0.0%hi, 0.0%si, 0.0%st Mem: 124352k total, 115368k used, 8984k free, 4244k buffers Swap: 245752k total, 6056k used, 239696k free, 84264k cached PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND 8028 www 20 0 14508 9776 724 S 0.0 7.9 0:00.45 nginx 8018 root 20 0 24204 3756 1088 S 0.0 3.0 0:57.48 php-cgi 8020 www 20 0 24204 3380 748 S 0.0 2.7 0:00.00 php-cgi 8019 www 20 0 24204 3372 748 S 0.0 2.7 0:00.00 php-cgi 8023 www 20 0 24204 3364 748 S 0.0 2.7 0:00.00 php-cgi 8024 www 20 0 24204 3352 748 S 0.0 2.7 0:00.00 php-cgi 8021 www 20 0 24204 3344 748 S 0.0 2.7 0:00.00 php-cgi 10712 root 20 0 8464 2968 2348 S 0.0 2.4 0:00.27 sshd 21604 mysql 20 0 35100 2240 1136 S 0.0 1.8 0:00.02 mysqld 10716 root 20 0 4496 1724 1388 S 0.0 1.4 0:00.03 bash 10852 root 20 0 2336 1104 880 R 0.0 0.9 0:00.01 top 529 root 20 0 27408 1040 788 S 0.0 0.8 0:00.34 rsyslogd 8026 root 20 0 4696 612 248 S 0.0 0.5 0:00.00 nginx 575 root 20 0 5496 512 396 S 0.0 0.4 0:00.45 sshd 1 root 20 0 2036 468 440 S 0.0 0.4 0:04.16 init 566 root 20 0 3788 432 384 S 0.0 0.3 0:00.46 cron 592 root 20 0 1708 412 408 S 0.0 0.3 0:00.00 getty
However, the surprises did not stop there, when I did a disk I/O test, I was literally astonished by the I/O speed:
dd if=/dev/zero of=test bs=64k count=16k conv=fdatasync 16384+0 records in 16384+0 records out 1073741824 bytes (1.1 GB) copied, 9.60772 s, 112 MB/s
And I did it one more time just to make sure I was not just being lucky:
dd if=/dev/zero of=test bs=64k count=16k conv=fdatasync 16384+0 records in 16384+0 records out 1073741824 bytes (1.1 GB) copied, 9.61134 s, 112 MB/s
This is definitely one of the best, if not the best, disk I/O that I have witnessed.
I have done quite a bit of research about their VPS offerings, however one thing that I could not find is their port speed. Nonetheless, the wget results using Cachefly seems to indicate that they are on Gbit port:
wget http://cachefly.cachefly.net/100mb.test --2011-06-17 13:27:44-- http://cachefly.cachefly.net/100mb.test Resolving cachefly.cachefly.net... 205.234.175.175 Connecting to cachefly.cachefly.net|205.234.175.175|:80... connected. HTTP request sent, awaiting response... 200 OK Length: 104857600 (100M) [application/octet-stream] Saving to: â100mb.testâ 100%[======================================>] 104,857,600 23.9M/s in 4.8s 2011-06-17 13:27:49 (20.9 MB/s) - â100mb.testâ
Upload speed from the test VPS, although not surprisingly impressive, are definitely pretty good:
First one is the BuyVM VPS from Fremont, CA:
wget 216.231.132.67/100mb.test --2011-06-17 05:15:08-- http://216.231.132.67/100mb.test Connecting to 216.231.132.67:80... connected. HTTP request sent, awaiting response... 200 OK Length: 104857600 (100M) [application/octet-stream] Saving to: `100mb.test' 100%[======================================>] 104,857,600 1.49M/s in 80s 2011-06-17 05:16:28 (1.25 MB/s) - `100mb.test' saved [104857600/104857600]
Next is Nix Communication VPS from Montreal, Canada, which is perhaps the closest to Chicago among the tests I did:
wget 216.231.132.67/100mb.test --2011-06-17 05:18:32-- http://216.231.132.67/100mb.test Connecting to 216.231.132.67:80... connected. HTTP request sent, awaiting response... 200 OK Length: 104857600 (100M) [application/octet-stream] Saving to: `100mb.test' 100%[======================================>] 104,857,600 6.87M/s in 17s 2011-06-17 05:18:49 (5.93 MB/s) - `100mb.test' saved [104857600/104857600]
And finally Quickweb VPS from London, UK, which, surprisingly, was not all that lagging behind:
wget 216.231.132.67/100mb.test --2011-06-17 05:19:38-- http://216.231.132.67/100mb.test Connecting to 216.231.132.67:80... connected. HTTP request sent, awaiting response... 200 OK Length: 104857600 (100M) [application/octet-stream] Saving to: `100mb.test' 100%[======================================>] 104,857,600 5.53M/s in 18s 2011-06-17 05:19:56 (5.62 MB/s) - `100mb.test' saved [104857600/104857600]
Finally the UnixBench results and the GeekBench results, which unfortunately is not that remarkable, perhaps due to the fact that only single core is available:
# # # # # # # ##### ###### # # #### # # # # ## # # # # # # # ## # # # # # # # # # # # ## ##### ##### # # # # ###### # # # # # # ## # # # # # # # # # # # # ## # # # # # # # ## # # # # #### # # # # # ##### ###### # # #### # # Version 5.1.3 Based on the Byte Magazine Unix Benchmark Multi-CPU version Version 5 revisions by Ian Smith, Sunnyvale, CA, USA January 13, 2011 johantheghost at yahoo period com 1 x Dhrystone 2 using register variables 1 2 3 4 5 6 7 8 9 10 1 x Double-Precision Whetstone 1 2 3 4 5 6 7 8 9 10 1 x Execl Throughput 1 2 3 1 x File Copy 1024 bufsize 2000 maxblocks 1 2 3 1 x File Copy 256 bufsize 500 maxblocks 1 2 3 1 x File Copy 4096 bufsize 8000 maxblocks 1 2 3 1 x Pipe Throughput 1 2 3 4 5 6 7 8 9 10 1 x Pipe-based Context Switching 1 2 3 4 5 6 7 8 9 10 1 x Process Creation 1 2 3 1 x System Call Overhead 1 2 3 4 5 6 7 8 9 10 1 x Shell Scripts (1 concurrent) 1 2 3 1 x Shell Scripts (8 concurrent) 1 2 3 ======================================================================== BYTE UNIX Benchmarks (Version 5.1.3) System: 96mb: GNU/Linux OS: GNU/Linux -- 2.6.32-5-686-bigmem -- #1 SMP Wed May 18 07:33:52 UTC 2011 Machine: i686 (unknown) Language: en_US.utf8 (charmap="UTF-8", collate="UTF-8") CPU 0: AMD Opteron(tm) Processor 6128 (4000.0 bogomips) Hyper-Threading, MMX, AMD MMX, Physical Address Ext, SYSCALL/SYSRET 19:25:13 up 5 days, 8 min, 1 user, load average: 0.16, 0.03, 0.01; runlevel 2 ------------------------------------------------------------------------ Benchmark Run: Mon Jun 20 2011 19:25:13 - 19:53:09 1 CPU in system; running 1 parallel copy of tests Dhrystone 2 using register variables 8842764.1 lps (10.0 s, 7 samples) Double-Precision Whetstone 1526.5 MWIPS (10.0 s, 7 samples) Execl Throughput 1623.3 lps (29.9 s, 2 samples) File Copy 1024 bufsize 2000 maxblocks 299978.0 KBps (30.0 s, 2 samples) File Copy 256 bufsize 500 maxblocks 85266.5 KBps (30.0 s, 2 samples) File Copy 4096 bufsize 8000 maxblocks 601062.9 KBps (30.0 s, 2 samples) Pipe Throughput 565178.6 lps (10.0 s, 7 samples) Pipe-based Context Switching 61785.9 lps (10.0 s, 7 samples) Process Creation 2761.1 lps (30.0 s, 2 samples) Shell Scripts (1 concurrent) 2700.8 lpm (60.0 s, 2 samples) Shell Scripts (8 concurrent) 348.6 lpm (60.1 s, 2 samples) System Call Overhead 674854.4 lps (10.0 s, 7 samples) System Benchmarks Index Values BASELINE RESULT INDEX Dhrystone 2 using register variables 116700.0 8842764.1 757.7 Double-Precision Whetstone 55.0 1526.5 277.6 Execl Throughput 43.0 1623.3 377.5 File Copy 1024 bufsize 2000 maxblocks 3960.0 299978.0 757.5 File Copy 256 bufsize 500 maxblocks 1655.0 85266.5 515.2 File Copy 4096 bufsize 8000 maxblocks 5800.0 601062.9 1036.3 Pipe Throughput 12440.0 565178.6 454.3 Pipe-based Context Switching 4000.0 61785.9 154.5 Process Creation 126.0 2761.1 219.1 Shell Scripts (1 concurrent) 42.4 2700.8 637.0 Shell Scripts (8 concurrent) 6.0 348.6 581.0 System Call Overhead 15000.0 674854.4 449.9 ======== System Benchmarks Index Score 456.6
GeekBench results is above average, which suggested really good performance otherwise:
Geekbench 2.1.13 : http://www.primatelabs.ca/geekbench/ Geekbench is in tryout mode. Your copy of Geekbench is in tryout mode. If you like Geekbench, please purchase it so we can continue to make Geekbench awesome. You can purchase Geekbench online at: http://store.primatelabs.ca/ If you've already purchased Geekbench, enter your license information with the following command line: dist/Geekbench21-Linux/geekbench_x86_32 -r System Information Platform: Linux x86 (32-bit) Compiler: GCC 4.1.2 20070925 (Red Hat 4.1.2-33) Operating System: Linux 2.6.32-5-686-bigmem i686 Model: Linux PC (AMD Opteron(tm) Processor 6128) Motherboard: Unknown Motherboard Processor: AMD Opteron(tm) Processor 6128 Processor ID: AuthenticAMD Family 16 Model 9 Stepping 1 Logical Processors: 1 Physical Processors: 1 Processor Frequency: 2.00 GHz L1 Instruction Cache: 64.0 KB L1 Data Cache: 64.0 KB L2 Cache: 512 KB L3 Cache: 10.00 MB Bus Frequency: 0.00 Hz Memory: 121 MB Memory Type: N/A SIMD: 1 BIOS: N/A Processor Model: AMD Opteron(tm) Processor 6128 Processor Cores: 1 Integer Blowfish single-threaded scalar 1443 ||||| multi-threaded scalar 1553 |||||| Text Compress single-threaded scalar 1541 |||||| multi-threaded scalar 1493 ||||| Text Decompress single-threaded scalar 1539 |||||| multi-threaded scalar 1578 |||||| Image Compress single-threaded scalar 1253 ||||| multi-threaded scalar 1229 |||| Image Decompress single-threaded scalar 1009 |||| multi-threaded scalar 1035 |||| Lua single-threaded scalar 2174 |||||||| multi-threaded scalar 2214 |||||||| Floating Point Mandelbrot single-threaded scalar 1416 ||||| multi-threaded scalar 1438 ||||| Dot Product single-threaded scalar 2056 |||||||| multi-threaded scalar 2178 |||||||| single-threaded vector 1929 ||||||| multi-threaded vector 2220 |||||||| LU Decomposition single-threaded scalar 1698 |||||| multi-threaded scalar 1730 |||||| Primality Test single-threaded scalar 1942 ||||||| multi-threaded scalar 1556 |||||| Sharpen Image single-threaded scalar 5219 |||||||||||||||||||| multi-threaded scalar 5305 ||||||||||||||||||||| Blur Image single-threaded scalar 5694 |||||||||||||||||||||| multi-threaded scalar 5718 |||||||||||||||||||||| Memory Read Sequential single-threaded scalar 1915 ||||||| Write Sequential single-threaded scalar 2836 ||||||||||| Stdlib Allocate single-threaded scalar 2122 |||||||| Stdlib Write single-threaded scalar 1235 |||| Stdlib Copy single-threaded scalar 1778 ||||||| Stream Stream Copy single-threaded scalar 1963 ||||||| single-threaded vector 2193 |||||||| Stream Scale single-threaded scalar 1984 ||||||| single-threaded vector 2125 |||||||| Stream Add single-threaded scalar 1790 ||||||| single-threaded vector 1993 ||||||| Stream Triad single-threaded scalar 1929 ||||||| single-threaded vector 1479 ||||| Integer Score: 1505 |||||| Floating Point Score: 2864 ||||||||||| Memory Score: 1977 ||||||| Stream Score: 1932 ||||||| Overall Geekbench Score: 2117 ||||||||
In conclusion, other than the processing power, this VPS box offers impressive disk I/O and network speed, however, for $4/month, there is really not a lot more you can ask for.
Support and Customer Services
I have to confess that I had a hard time to send a ticket in to see how well FST Servers responds, because basically everything was in the control panel. However, when I finally did, the question was responded in 13 minutes (question submitted at 5:03PM, response received at 5:16PM) with a very detailed description and the link to the documentation page containing relevant information.
Conclusion
Low prices, compact but useful custom control panel, fast disk I/O and network speed, and fast ticket response are a few great things that worth giving this box a try. Granted it is not the most powerful box that ever listed on 96MB review and adding a little more power to the processors would perhaps be better (and I am having a hard time comprehending why CentOS could not run on a 128MB box), but for merely 4USD per month, this service definitely worth every single cent that is paid for.
Nice Review,
It is nice to see some providers with a custom control panel this seems to be getting more popular these days.
Note on the CentOS only being available on the 256mb+ packages I would assume this is because YUM has a hard time with anything less that 256mb.
That said with enough SWAP which you seem to have according to ‘top’ it usually manages ok.
As you said for $4 it does pretty well by the look of it, although budget providers seem to be getting a hard press these days the fact that an effort has been made to create their own CP shows a good amount of effort has been made with the back end infrastructure services so although I have never heard of them either they are probably going to be a good company 🙂 +1
@Anthony: I do think it is a great idea to create a custom control panel, especially for Xen server, which I am not sure if SolusVM does support it fully yet. Therefore, it definitely make sense for them to implement their own custom control panel instead of receiving 20 tickets a day on rebooting/reloading OS. CentOS seems to be a memory hogger from what I can see, so may be restricting it to higher plans does make sense. Good company you have as well and I sincerely hope all the great VPS providers will be around for a long long time 🙂
Pingback: FST Servers – $4/Month 512MB Xen VPS in Chicago – Low End Box