96MB Low End VPS Review Part 54 – Prometeus KVM
Flipping through the emails in my inbox, the first email from Prometeus to me was sent on January 28th of this year, when they were first getting into the low end VPS market and I have replied with a self-imposed deadline of early April. I have almost missed the deadline by another four months and meanwhile, Prometeus has evolved from one of the brand names that few low end VPS community members know to one of the most popular brands, earning the second place in the Top Providers Q2 2012 poll on LowEndTalk. Obviously, part of their success attribute to the great personality that the owner of Prometeus has, including the time where he actually helped some VPS-provider-wanna-be to start his business. Having said that, however, are their servers really solid? I have decided to get one and take a look at it myself.
Disclaimer: I have received two free VPS from Prometeus, however they are unrelated to this particular post and for this particular test VPS, I have bought and paid for it as a regular customer without notifying Prometeus prior to this review being released.
General Information and Background
2012 may not be a good year for many people, myself included as I was having considerable amount of troubles in various aspects of my life, both career-wise and personal-wise, however, it was a good year for Prometeus as they have just celebrated their 15th birthday, making them pretty much one of the oldest VPS hosting providers that I am aware of. However, the best part of this is they have actually offered quite a few deals that has to do with the magic number of 15, and as per their advertisement on WHT, here is what you get for less than 15 USD per year:
As far as I can see, this is probably one of the cheapest KVM plans out there, even cheaper than one of the most famous provider, BuyVM’s 128MB OpenVZ plan. However, do note that the hard drive size is a lot larger in the case of BuyVM and the bandwidth amount is more generous as well.
The sign up system is pretty straight-forward and Prometeus has made use of the cool slider feature in the new WHMCS system:
Prometeus has also have a special low end VPS group for some even cheaper and smaller OpenVZ VPS, such as some of the plans listed below:
![image_thumb[2]](http://www.96mb.com/wp-content/uploads/2012/08/image_thumb21.png "image_thumb[2]")
and the following as well:
![image_thumb[4]](http://www.96mb.com/wp-content/uploads/2012/08/image_thumb4.png "image_thumb[4]")
and in particular, a Windows low end plan, which is not the cheapest plan I have seen, particularly knowing that the hard drive size might be a bit too small, but is quite reasonably priced:
![image_thumb[7]](http://www.96mb.com/wp-content/uploads/2012/08/image_thumb7.png "image_thumb[7]")
As you can see, the KVM 1 plan is actually not listed on the slider and you can choose to view in either USD or EUR. Furthermore, there is very detailed information on each of the plans, and you do not actually need to click on Order Now to see the details.
There are not really a lot of options that you can choose from during the sign up process, which obviously makes the clients sign up much faster, but however could be a lost of business opportunities to them as many people, myself included, like to pick some additional options during the sign up process.
The first option you can choose from is the OS selector:
As you can see, there are very few options that you can choose from, basically just the latest version of CentOS, Debian and Ubuntu. Granted that this is probably the most frequently used systems among the frequent users in this community, it could be nice if they could include a few other system installer, including the FreeBSD, which is one of the reasons for many people chose to sign up for a KVM-based system.
Besides that, the only thing you can choose from is the additional bandwidth, and interestingly, the only option you have is to add an additional 2TB, which cost 5 Euros per month:
Service activation was not instant for me, however was pretty quick, I received my Paypal payment confirmation at 11:21pm and the new server welcome email arrived at 11:58pm, which, considering it is probably 5 or 6am in Europe, is pretty impressive.
The welcome email looks something like this:
First note that you will have to install the OS yourself, as the initial system has no OS installed. Also, Protemteus was kind enough to point out that you will need additional RAM if you want to get CentOS 6 running.
I was actually quite impressed by the ability of Prometeus to have DHCP enabled for the server. Although this feature may seem to be rather minor, I have lost count of the number of times which I was really annoyed during OS installation that I have to manually type in the strings of IP addresses for my own server IP, the server gateway and DNS server IP, and the worst part with Debian installer is that you won’t know if you have typed the correct IP addresses until you have come to the stage of configuring apt, which is a lot later during the process. Therefore, although it is a relatively minor feature, I actually really liked it.
The other nice feature that they have is IPv6 is enabled by default to your VPS, which is different from many providers that you will have to put in a ticket to ask for it. Granted it may be a waste of their IPv6 address space since probably 10% of those VPS users will actively make use of the IPv6 address they are assigned to, however it was nice that the feature is automated and you can actually check the IPv6 address that you are assigned to via link without having to go through the trouble of logging into SolusVM panel.
Prometeus, like most of the VPS providers, use SolusVM to control their VPS, and the control panel is using HTTPS protocol on port 5656. Once logged in, a pretty standard SolusVM interface was shown:
As you can see, there is no central backup feature and the IPv6 addresses are actually not shown here. Furthermore, the IPv4 addresses do not have instant rDNS:
In the settings menu, the usual KVM VPS settings can be seen:
The only settings that I have not seen before is the VNC Keymap settings, however for that particular setting, only “Default” is available.
For Boot order, you can choose to use Hard disk or CDROM first or just boot from one of them:
For network cards, you can choose between RealTek, Intel and Virtio cards:
Finally, for the disk driver, you can choose between Virtio and ide drivers:
There are a few ISO available:
As you can see, besides the pretty standard Linux operating system installer, there are a few not-so-common templates such as SME Server 8 and Tiny CorePlus.
Test on the VPS
The KVM VPS that I have bought for testing has 128MB of RAM, 6GB of hard drive and 256GB of bandwidth. The test server is located in Milan, Italy as per their advertisement on WHT and I have installed Debian 6 32-bit OS for testing purposes. Note that initially, you will have do your own installation as you were given a blank server. However I am sure for everyone who is reading this blog right now, installing Debian is not a big issue.
When the system installation was first completed, about 14MB of RAM was used. Note that obviously this depends on the packages that you have chosen to install during the installation. For me, installing the standard system utilities and SSH server seems to be sufficient. In fact, you can choose to not install the SSH server and install Dropbear directly from VNC later on if you need SSH, or simply installing and optimizing memory usage later on:
free -m
total used free shared buffers cached
Mem: 121 51 70 0 2 34
-/+ buffers/cache: 14 107
Swap: 236 0 236
Top showing all the processes running, amazingly a common package found in the standard OpenVZ Debian 6 template, Apache2, was not installed by default in this case:
top - 21:53:53 up 25 min, 2 users, load average: 0.00, 0.00, 0.00
Tasks: 58 total, 1 running, 57 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: 124880k total, 52968k used, 71912k free, 2312k buffers
Swap: 242680k total, 0k used, 242680k free, 35396k cached
PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
1368 root 20 0 8264 2864 2344 S 0.5 2.3 0:00.07 sshd
1340 root 20 0 5516 2820 1468 S 0.0 2.3 0:00.20 bash
1371 root 20 0 5508 2816 1468 S 0.0 2.3 0:00.15 bash
932 root 20 0 27452 1536 1032 S 0.0 1.2 0:00.01 rsyslogd
1291 root 20 0 2564 1304 1024 S 0.0 1.0 0:00.05 login
1388 root 20 0 2336 1124 904 R 0.0 0.9 0:00.01 top
1338 root 20 0 5500 968 580 S 0.0 0.8 0:00.00 sshd
347 root 16 -4 2420 964 404 S 0.0 0.8 0:00.03 udevd
1273 Debian-e 20 0 6520 928 612 S 0.0 0.7 0:00.00 exim4
790 statd 20 0 1940 788 656 S 0.0 0.6 0:00.00 rpc.statd
460 root 18 -2 2388 768 332 S 0.0 0.6 0:00.00 udevd
1011 root 20 0 3788 764 600 S 0.0 0.6 0:00.00 cron
470 root 18 -2 2388 760 328 S 0.0 0.6 0:00.00 udevd
1 root 20 0 2036 712 616 S 0.0 0.6 0:00.72 init
996 root 20 0 1708 584 480 S 0.0 0.5 0:00.00 acpid
1293 root 20 0 1712 572 492 S 0.0 0.5 0:00.00 getty
1294 root 20 0 1712 572 492 S 0.0 0.5 0:00.00 getty
And htop output:
About 638MB of hard drive space was used after the initial installation:
df -h Filesystem Size Used Avail Use% Mounted on /dev/vda1 5.7G 638M 4.8G 12% / tmpfs 61M 0 61M 0% /lib/init/rw udev 57M 112K 57M 1% /dev tmpfs 61M 0 61M 0% /dev/shm
The Inodes were a bit towards the low end, obviously due to the small hard drive size:
df -i Filesystem Inodes IUsed IFree IUse% Mounted on /dev/vda1 378256 25288 352968 7% / tmpfs 15610 5 15605 1% /lib/init/rw udev 14498 496 14002 4% /dev tmpfs 15610 1 15609 1% /dev/shm
After the full LNMP stack was loaded, the RAM usage went to 31MB:
free -m
total used free shared buffers cached
Mem: 121 117 4 0 5 80
-/+ buffers/cache: 31 90
Swap: 236 6 230
Top showing the processes running:
top - 10:30:53 up 1:02, 2 users, load average: 0.08, 0.23, 0.44
Tasks: 68 total, 1 running, 67 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: 124880k total, 121792k used, 3088k free, 5892k buffers
Swap: 242680k total, 6392k used, 236288k free, 83684k cached
PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
2940 www 20 0 15004 10m 432 S 0.0 8.9 0:00.01 nginx
2921 mysql 20 0 33948 4636 2040 S 0.0 3.7 0:00.01 mysqld
2931 root 20 0 24124 4548 1408 S 0.0 3.6 0:00.03 php-cgi
2932 www 20 0 24124 4156 1016 S 0.0 3.3 0:00.00 php-cgi
2933 www 20 0 24124 4156 1016 S 0.0 3.3 0:00.00 php-cgi
2934 www 20 0 24124 4156 1016 S 0.0 3.3 0:00.00 php-cgi
2935 www 20 0 24124 4156 1016 S 0.0 3.3 0:00.00 php-cgi
2936 www 20 0 24124 4156 1016 S 0.0 3.3 0:00.00 php-cgi
1371 root 20 0 5508 1160 852 S 0.0 0.9 0:00.17 bash
2974 root 20 0 2336 1132 904 R 0.0 0.9 0:00.00 top
932 root 20 0 27452 988 760 S 0.0 0.8 0:00.02 rsyslogd
1368 root 20 0 8404 800 668 S 0.0 0.6 0:01.82 sshd
2939 root 20 0 4792 720 272 S 0.0 0.6 0:00.00 nginx
2819 root 20 0 1752 556 472 S 0.0 0.4 0:00.00 mysqld_safe
1 root 20 0 2036 504 480 S 0.0 0.4 0:00.73 init
1291 root 20 0 2564 500 500 S 0.0 0.4 0:00.05 login
347 root 16 -4 2420 492 308 S 0.0 0.4 0:00.03 udevd
And htop output:
Close to 2GB of hard drive space was used:
df -h Filesystem Size Used Avail Use% Mounted on /dev/vda1 5.7G 1.9G 3.6G 35% / tmpfs 61M 0 61M 0% /lib/init/rw udev 57M 112K 57M 1% /dev tmpfs 61M 0 61M 0% /dev/shm
And with one third of hard drive space used, about one fifth of Inodes were used as well:
df -i Filesystem Inodes IUsed IFree IUse% Mounted on /dev/vda1 378256 73969 304287 20% / tmpfs 15610 5 15605 1% /lib/init/rw udev 14498 496 14002 4% /dev tmpfs 15610 1 15609 1% /dev/shm
Uptime shows the VPS is rarely used and is pretty stable:
uptime 04:59:43 up 53 days, 19:31, 2 users, load average: 0.00, 0.00, 0.00
VMStats shows similar information:
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 52484 4440 51424 0 0 33 5 16 13 0 0 99 0
There is only one CPU core for this VPS, and the clock speed is 2.0GHz:
cat /proc/cpuinfo processor : 0 vendor_id : GenuineIntel cpu family : 6 model : 13 model name : QEMU Virtual CPU version (cpu64-rhel6) stepping : 3 cpu MHz : 1999.999 cache size : 4096 KB fdiv_bug : no hlt_bug : no f00f_bug : no coma_bug : no fpu : yes fpu_exception : yes cpuid level : 4 wp : yes flags : fpu de pse tsc msr pae mce cx8 apic mtrr pge mca cmov pse36 clflush mmx fxsr sse sse2 syscall nx lm up pni cx16 hypervisor lahf_lm bogomips : 3999.99 clflush size : 64 cache_alignment : 64 address sizes : 46 bits physical, 48 bits virtual power management:
The meminfo output is pretty standard for a KVM-based machine:
cat /proc/meminfo MemTotal: 124880 kB MemFree: 3176 kB Buffers: 19292 kB Cached: 70260 kB SwapCached: 1772 kB Active: 32624 kB Inactive: 77288 kB Active(anon): 7548 kB Inactive(anon): 12836 kB Active(file): 25076 kB Inactive(file): 64452 kB Unevictable: 0 kB Mlocked: 0 kB HighTotal: 0 kB HighFree: 0 kB LowTotal: 124880 kB LowFree: 3176 kB SwapTotal: 242680 kB SwapFree: 237420 kB Dirty: 36 kB Writeback: 0 kB AnonPages: 18844 kB Mapped: 4232 kB Shmem: 12 kB Slab: 8900 kB SReclaimable: 5512 kB SUnreclaim: 3388 kB KernelStack: 544 kB PageTables: 664 kB NFS_Unstable: 0 kB Bounce: 0 kB WritebackTmp: 0 kB CommitLimit: 305120 kB Committed_AS: 79448 kB VmallocTotal: 897028 kB VmallocUsed: 5700 kB VmallocChunk: 883628 kB HardwareCorrupted: 0 kB HugePages_Total: 0 HugePages_Free: 0 HugePages_Rsvd: 0 HugePages_Surp: 0 Hugepagesize: 4096 kB DirectMap4k: 12276 kB DirectMap4M: 118784 kB
And time sync:
time sync
0.03s real 0.00s user 0.00s system
For the little hard drive space provided, I assume there was some really high-end setup with the RAID arrays and therefore should have a pretty good disk I/O speed, I am glad that this VPS did not fail me:
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, 6.01159 s, 179 MB/s
If this is not fast enough for you, the second dd output has actually went above 200MB/s:
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, 5.23702 s, 205 MB/s
And the third time the VPS produced even better output:
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, 4.62384 s, 232 MB/s
IOPing is not all 0.1ms, but is nontheless really stable with no surprises:
ioping -c 10 . 4096 bytes from . (ext3 /dev/disk/by-uuid/4d39c3e3-d142-4ec4-a762-a5a5da549afc): request=1 time=0.1 ms 4096 bytes from . (ext3 /dev/disk/by-uuid/4d39c3e3-d142-4ec4-a762-a5a5da549afc): request=2 time=0.2 ms 4096 bytes from . (ext3 /dev/disk/by-uuid/4d39c3e3-d142-4ec4-a762-a5a5da549afc): request=3 time=0.2 ms 4096 bytes from . (ext3 /dev/disk/by-uuid/4d39c3e3-d142-4ec4-a762-a5a5da549afc): request=4 time=0.2 ms 4096 bytes from . (ext3 /dev/disk/by-uuid/4d39c3e3-d142-4ec4-a762-a5a5da549afc): request=5 time=0.2 ms 4096 bytes from . (ext3 /dev/disk/by-uuid/4d39c3e3-d142-4ec4-a762-a5a5da549afc): request=6 time=0.2 ms 4096 bytes from . (ext3 /dev/disk/by-uuid/4d39c3e3-d142-4ec4-a762-a5a5da549afc): request=7 time=0.2 ms 4096 bytes from . (ext3 /dev/disk/by-uuid/4d39c3e3-d142-4ec4-a762-a5a5da549afc): request=8 time=0.3 ms 4096 bytes from . (ext3 /dev/disk/by-uuid/4d39c3e3-d142-4ec4-a762-a5a5da549afc): request=9 time=0.3 ms 4096 bytes from . (ext3 /dev/disk/by-uuid/4d39c3e3-d142-4ec4-a762-a5a5da549afc): request=10 time=0.2 ms --- . (ext3 /dev/disk/by-uuid/4d39c3e3-d142-4ec4-a762-a5a5da549afc) ioping statistics --- 10 requests completed in 9004.0 ms, 5061 iops, 19.8 mb/s min/avg/max/mdev = 0.1/0.2/0.3/0.0 ms
Testing again showed the same:
ioping -c 10 . 4096 bytes from . (ext3 /dev/disk/by-uuid/4d39c3e3-d142-4ec4-a762-a5a5da549afc): request=1 time=0.1 ms 4096 bytes from . (ext3 /dev/disk/by-uuid/4d39c3e3-d142-4ec4-a762-a5a5da549afc): request=2 time=0.3 ms 4096 bytes from . (ext3 /dev/disk/by-uuid/4d39c3e3-d142-4ec4-a762-a5a5da549afc): request=3 time=0.2 ms 4096 bytes from . (ext3 /dev/disk/by-uuid/4d39c3e3-d142-4ec4-a762-a5a5da549afc): request=4 time=0.2 ms 4096 bytes from . (ext3 /dev/disk/by-uuid/4d39c3e3-d142-4ec4-a762-a5a5da549afc): request=5 time=0.2 ms 4096 bytes from . (ext3 /dev/disk/by-uuid/4d39c3e3-d142-4ec4-a762-a5a5da549afc): request=6 time=0.2 ms 4096 bytes from . (ext3 /dev/disk/by-uuid/4d39c3e3-d142-4ec4-a762-a5a5da549afc): request=7 time=0.2 ms 4096 bytes from . (ext3 /dev/disk/by-uuid/4d39c3e3-d142-4ec4-a762-a5a5da549afc): request=8 time=0.2 ms 4096 bytes from . (ext3 /dev/disk/by-uuid/4d39c3e3-d142-4ec4-a762-a5a5da549afc): request=9 time=0.2 ms 4096 bytes from . (ext3 /dev/disk/by-uuid/4d39c3e3-d142-4ec4-a762-a5a5da549afc): request=10 time=0.2 ms --- . (ext3 /dev/disk/by-uuid/4d39c3e3-d142-4ec4-a762-a5a5da549afc) ioping statistics --- 10 requests completed in 9003.6 ms, 4990 iops, 19.5 mb/s min/avg/max/mdev = 0.1/0.2/0.3/0.0 ms
Download speed is OK, but not exactly impressive:
wget cachefly.cachefly.net/100mb.test -O /dev/null --2012-08-06 23:52:23-- 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: `/dev/null' 100%[======================================>] 104,857,600 6.39M/s in 16s 2012-08-06 23:52:39 (6.31 MB/s) - `/dev/null' saved [104857600/104857600]
And the second time is actually a little worse than the first one:
wget cachefly.cachefly.net/100mb.test -O /dev/null --2012-08-07 09:15:29-- 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: `/dev/null' 100%[======================================>] 104,857,600 5.78M/s in 18s 2012-08-07 09:15:46 (5.67 MB/s) - `/dev/null' saved [104857600/104857600]
Upload speed, on the other end, was unfortunately on the slow end of the spectrum.
First is with my Quickweb VPS in Chicago, IL:
wget 37.247.xxx.xxx/100mb.test -O /dev/null --2012-08-06 11:54:43-- http://37.247.xxx.xxx/100mb.test Connecting to 37.247.xxx.xxx:80... connected. HTTP request sent, awaiting response... 200 OK Length: 104857600 (100M) [application/octet-stream] Saving to: `/dev/null' 100%[======================================>] 104,857,600 1.00M/s in 1m 43s 2012-08-06 11:56:26 (994 KB/s) - `/dev/null' saved [104857600/104857600]
As you can see, the speed was hitting close to 1MB/s. I have actually did another round of test later on just to make sure it is the “usual” speed for the VPS and it turned out to be the case:
wget 37.247.xxx.xxx/100mb.test -O /dev/null --2012-08-06 18:03:02-- http://37.247.xxx.xxx/100mb.test Connecting to 37.247.xxx.xxx:80... connected. HTTP request sent, awaiting response... 200 OK Length: 104857600 (100M) [application/octet-stream] Saving to: `/dev/null' 100%[======================================>] 104,857,600 1009K/s in 1m 43s 2012-08-06 18:04:46 (991 KB/s) - `/dev/null' saved [104857600/104857600]
Next, my Quickweb VPS in Los Angeles, CA, showed the slowest speed among the three test VPS that I have.
wget 37.247.xxx.xxx/100mb.test -O /dev/null --2012-08-06 11:57:11-- http://37.247.xxx.xxx/100mb.test Connecting to 37.247.xxx.xxx:80... connected. HTTP request sent, awaiting response... 200 OK Length: 104857600 (100M) [application/octet-stream] Saving to: `/dev/null' 100%[======================================>] 104,857,600 175K/s in 14m 55s 2012-08-06 12:12:07 (114 KB/s) - `/dev/null' saved [104857600/104857600]
Testing again showed a much better result, but is nonetheless pretty slow:
wget 37.247.xxx.xxx/100mb.test -O /dev/null --2012-08-06 21:22:22-- http://37.247.xxx.xxx/100mb.test Connecting to 37.247.xxx.xxx:80... connected. HTTP request sent, awaiting response... 200 OK Length: 104857600 (100M) [application/octet-stream] Saving to: `/dev/null' 100%[======================================>] 104,857,600 471K/s in 2m 56s 2012-08-06 21:25:19 (582 KB/s) - `/dev/null' saved [104857600/104857600]
Finally, the upload speed from my XenVZ test VPS in Maidenhead, UK, is a lot better as it is geographically close:
wget 37.247.xxx.xxx/100mb.test -O /dev/null --2012-08-07 00:10:20-- http://37.247.xxx.xxx/100mb.test Connecting to 37.247.xxx.xxx:80... connected. HTTP request sent, awaiting response... 200 OK Length: 104857600 (100M) [application/octet-stream] Saving to: `/dev/null' 100%[======================================>] 104,857,600 1.61M/s in 51s 2012-08-07 00:11:11 (1.96 MB/s) - `/dev/null' saved [104857600/104857600]
Testing again showed pretty consistent speed:
wget 37.247.xxx.xxx/100mb.test -O /dev/null --2012-08-07 09:26:40-- http://37.247.xxx.xxx/100mb.test Connecting to 37.247.xxx.xxx:80... connected. HTTP request sent, awaiting response... 200 OK Length: 104857600 (100M) [application/octet-stream] Saving to: `/dev/null' 100%[======================================>] 104,857,600 1.77M/s in 48s 2012-08-07 09:27:28 (2.07 MB/s) - `/dev/null' saved [104857600/104857600]
Therefore, approximately you get 1MB/s of upload speed along the East coast, 0.5MB/s along west coast and 2MB/s in Europe, which is not exactly too impressive. Having said that though, sometimes changing the network card could make a difference to the speed in the KVM VPS.
UPDATE: Salvatore, the owner of Prometeus, has emailed me proposing some configuration changes to boost the network speed for this VPS and here:
1. Use Virtio network card as the others are emulated Intel and Realtek (note that you will need to reboot the VPS after this).
2. Edit /etc/sysctl.conf and add the following:
net.core.rmem_max=16777216 net.core.wmem_max=16777216 net.ipv4.tcp_rmem=4096 87380 16777216 net.ipv4.tcp_wmem=4096 65536 16777216
And run:
sysctl -p
From what I can see, the download speed is much better this way:
The first Cachefly test was done at night in EST time zone:
wget cachefly.cachefly.net/100mb.test -O /dev/null --2012-08-08 22:49:18-- 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: `/dev/null' 100%[======================================>] 104,857,600 21.2M/s in 6.8s 2012-08-08 22:49:25 (14.7 MB/s) - `/dev/null' saved [104857600/104857600]
And then when I repeat the exercise this morning, the result is even better:
wget cachefly.cachefly.net/100mb.test -O /dev/null --2012-08-09 07:38:51-- 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: `/dev/null' 100%[======================================>] 104,857,600 30.9M/s in 3.5s 2012-08-09 07:38:54 (28.7 MB/s) - `/dev/null' saved [104857600/104857600]
However the upload speed does not seem to show a lot difference, I started off again with my Quickweb VPS in Chicago, IL:
wget 37.247.xxx.xxx/100mb.test -O /dev/null --2012-08-08 22:58:56-- http://37.247.xxx.xxx/100mb.test Connecting to 37.247.xxx.xxx:80... connected. HTTP request sent, awaiting response... 200 OK Length: 104857600 (100M) [application/octet-stream] Saving to: `/dev/null' 100%[======================================>] 104,857,600 1001K/s in 1m 44s 2012-08-08 23:00:40 (984 KB/s) - `/dev/null' saved [104857600/104857600]
I did the test again this morning just to make sure it does not have anything to do with the time, and the result is actually slightly worse than last night:
wget 37.247.xxx.xxx/100mb.test -O /dev/null --2012-08-09 19:39:31-- http://37.247.xxx.xxx/100mb.test Connecting to 37.247.xxx.xxx:80... connected. HTTP request sent, awaiting response... 200 OK Length: 104857600 (100M) [application/octet-stream] Saving to: `/dev/null' 100%[======================================>] 104,857,600 945K/s in 1m 48s 2012-08-09 19:41:19 (950 KB/s) - `/dev/null' saved [104857600/104857600]
Similar results as before for the Quickweb VPS I have in Los Angeles, CA:
wget 37.247.xxx.xxx/100mb.test -O /dev/null --2012-08-08 23:11:32-- http://37.247.xxx.xxx/100mb.test Connecting to 37.247.xxx.xxx:80... connected. HTTP request sent, awaiting response... 200 OK Length: 104857600 (100M) [application/octet-stream] Saving to: `/dev/null' 100%[======================================>] 104,857,600 1.38M/s in 2m 54s 2012-08-08 23:14:26 (589 KB/s) - `/dev/null' saved [104857600/104857600]
And finally the XenVZ VPS in Maidenhead, UK:
wget 37.247.xxx.xxx/100mb.test -O /dev/null --2012-08-09 11:44:24-- http://37.247.xxx.xxx/100mb.test Connecting to 37.247.xxx.xxx:80... connected. HTTP request sent, awaiting response... 200 OK Length: 104857600 (100M) [application/octet-stream] Saving to: `/dev/null' 100%[======================================>] 104,857,600 2.10M/s in 48s 2012-08-09 11:45:12 (2.10 MB/s) - `/dev/null' saved [104857600/104857600]
As you can see, this fix is a huge boost for the download speed and therefore if your server is primarily for downloading data, it is definitely worthwhile to try it. However for upload, the gain is not that significant.
Finally, benchmarking time. As this VPS only has a single CPU core, I was not expecting really spectacular performance, and realistically, nothing above 1000 points in the Unixbench, and the actual test result seems to echo that:
# # # # # # # ##### ###### # # #### # #
# # ## # # # # # # # ## # # # # #
# # # # # # ## ##### ##### # # # # ######
# # # # # # ## # # # # # # # # #
# # # ## # # # # # # # ## # # # #
#### # # # # # ##### ###### # # #### # #
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: xxxxxx: GNU/Linux
OS: GNU/Linux -- 2.6.32-5-686 -- #1 SMP Sun May 6 04:01:19 UTC 2012
Machine: i686 (unknown)
Language: en_US.utf8 (charmap="ANSI_X3.4-1968", collate="ANSI_X3.4-1968")
CPU 0: QEMU Virtual CPU version (cpu64-rhel6) (4000.0 bogomips)
x86-64, MMX, Physical Address Ext, SYSCALL/SYSRET
10:34:57 up 1:06, 2 users, load average: 0.00, 0.09, 0.33; runlevel 2
------------------------------------------------------------------------
Benchmark Run: Wed Jun 13 2012 10:34:57 - 11:02:50
1 CPU in system; running 1 parallel copy of tests
Dhrystone 2 using register variables 12659574.8 lps (10.0 s, 7 samples)
Double-Precision Whetstone 1961.5 MWIPS (10.2 s, 7 samples)
Execl Throughput 4254.1 lps (30.0 s, 2 samples)
File Copy 1024 bufsize 2000 maxblocks 553621.2 KBps (30.0 s, 2 samples)
File Copy 256 bufsize 500 maxblocks 146388.5 KBps (30.0 s, 2 samples)
File Copy 4096 bufsize 8000 maxblocks 1248771.0 KBps (30.0 s, 2 samples)
Pipe Throughput 872091.2 lps (10.0 s, 7 samples)
Pipe-based Context Switching 234744.2 lps (10.0 s, 7 samples)
Process Creation 15513.0 lps (30.0 s, 2 samples)
Shell Scripts (1 concurrent) 5421.2 lpm (60.0 s, 2 samples)
Shell Scripts (8 concurrent) 693.2 lpm (60.0 s, 2 samples)
System Call Overhead 746137.6 lps (10.0 s, 7 samples)
System Benchmarks Index Values BASELINE RESULT INDEX
Dhrystone 2 using register variables 116700.0 12659574.8 1084.8
Double-Precision Whetstone 55.0 1961.5 356.6
Execl Throughput 43.0 4254.1 989.3
File Copy 1024 bufsize 2000 maxblocks 3960.0 553621.2 1398.0
File Copy 256 bufsize 500 maxblocks 1655.0 146388.5 884.5
File Copy 4096 bufsize 8000 maxblocks 5800.0 1248771.0 2153.1
Pipe Throughput 12440.0 872091.2 701.0
Pipe-based Context Switching 4000.0 234744.2 586.9
Process Creation 126.0 15513.0 1231.2
Shell Scripts (1 concurrent) 42.4 5421.2 1278.6
Shell Scripts (8 concurrent) 6.0 693.2 1155.3
System Call Overhead 15000.0 746137.6 497.4
========
System Benchmarks Index Score 922.4
Second time actually showed slightly worse results:
# # # # # # # ##### ###### # # #### # #
# # ## # # # # # # # ## # # # # #
# # # # # # ## ##### ##### # # # # ######
# # # # # # ## # # # # # # # # #
# # # ## # # # # # # # ## # # # #
#### # # # # # ##### ###### # # #### # #
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: XXXXX: GNU/Linux
OS: GNU/Linux -- 2.6.32-5-686 -- #1 SMP Sun May 6 04:01:19 UTC 2012
Machine: i686 (unknown)
Language: en_US.utf8 (charmap="ANSI_X3.4-1968", collate="ANSI_X3.4-1968")
CPU 0: QEMU Virtual CPU version (cpu64-rhel6) (4000.0 bogomips)
x86-64, MMX, Physical Address Ext, SYSCALL/SYSRET
00:27:50 up 54 days, 14:59, 2 users, load average: 0.00, 0.00, 0.00; runlevel 2
------------------------------------------------------------------------
Benchmark Run: Tue Aug 07 2012 00:27:50 - 00:55:41
1 CPU in system; running 1 parallel copy of tests
Dhrystone 2 using register variables 12248369.0 lps (10.0 s, 7 samples)
Double-Precision Whetstone 1947.0 MWIPS (10.1 s, 7 samples)
Execl Throughput 3880.5 lps (30.0 s, 2 samples)
File Copy 1024 bufsize 2000 maxblocks 500175.3 KBps (30.0 s, 2 samples)
File Copy 256 bufsize 500 maxblocks 133138.5 KBps (30.0 s, 2 samples)
File Copy 4096 bufsize 8000 maxblocks 1120725.1 KBps (30.0 s, 2 samples)
Pipe Throughput 859722.1 lps (10.0 s, 7 samples)
Pipe-based Context Switching 220215.7 lps (10.0 s, 7 samples)
Process Creation 13816.7 lps (30.0 s, 2 samples)
Shell Scripts (1 concurrent) 5060.8 lpm (60.0 s, 2 samples)
Shell Scripts (8 concurrent) 650.0 lpm (60.0 s, 2 samples)
System Call Overhead 730563.6 lps (10.0 s, 7 samples)
System Benchmarks Index Values BASELINE RESULT INDEX
Dhrystone 2 using register variables 116700.0 12248369.0 1049.6
Double-Precision Whetstone 55.0 1947.0 354.0
Execl Throughput 43.0 3880.5 902.5
File Copy 1024 bufsize 2000 maxblocks 3960.0 500175.3 1263.1
File Copy 256 bufsize 500 maxblocks 1655.0 133138.5 804.5
File Copy 4096 bufsize 8000 maxblocks 5800.0 1120725.1 1932.3
Pipe Throughput 12440.0 859722.1 691.1
Pipe-based Context Switching 4000.0 220215.7 550.5
Process Creation 126.0 13816.7 1096.6
Shell Scripts (1 concurrent) 42.4 5060.8 1193.6
Shell Scripts (8 concurrent) 6.0 650.0 1083.3
System Call Overhead 15000.0 730563.6 487.0
========
System Benchmarks Index Score 864.0
For Geekbench though, the score definitely looks a lot better, I was able to get more than 3200 points on both of the Geekbench runs:
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 i686
Model: Linux PC (QEMU Virtual CPU version (cpu64-rhel6))
Motherboard: Unknown Motherboard
Processor: QEMU Virtual CPU version (cpu64-rhel6)
Processor ID: GenuineIntel Family 6 Model 13 Stepping 3
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: 0.00 B
Bus Frequency: 0.00 Hz
Memory: 122 MB
Memory Type: N/A
SIMD: 1
BIOS: N/A
Processor Model: QEMU Virtual CPU version (cpu64-rhel6)
Processor Cores: 1
Integer
Blowfish
single-threaded scalar 1291 |||||
multi-threaded scalar 1384 |||||
Text Compress
single-threaded scalar 1738 ||||||
multi-threaded scalar 1690 ||||||
Text Decompress
single-threaded scalar 1870 |||||||
multi-threaded scalar 1911 |||||||
Image Compress
single-threaded scalar 1458 |||||
multi-threaded scalar 1432 |||||
Image Decompress
single-threaded scalar 1445 |||||
multi-threaded scalar 1490 |||||
Lua
single-threaded scalar 2429 |||||||||
multi-threaded scalar 2419 |||||||||
Floating Point
Mandelbrot
single-threaded scalar 1661 ||||||
multi-threaded scalar 1686 ||||||
Dot Product
single-threaded scalar 2734 ||||||||||
multi-threaded scalar 2899 |||||||||||
single-threaded vector 3226 ||||||||||||
multi-threaded vector 3747 ||||||||||||||
LU Decomposition
single-threaded scalar 2103 ||||||||
multi-threaded scalar 2145 ||||||||
Primality Test
single-threaded scalar 2925 |||||||||||
multi-threaded scalar 2341 |||||||||
Sharpen Image
single-threaded scalar 6566 ||||||||||||||||||||||||||
multi-threaded scalar 6457 |||||||||||||||||||||||||
Blur Image
single-threaded scalar 5222 ||||||||||||||||||||
multi-threaded scalar 5254 |||||||||||||||||||||
Memory
Read Sequential
single-threaded scalar 3938 |||||||||||||||
Write Sequential
single-threaded scalar 7255 |||||||||||||||||||||||||||||
Stdlib Allocate
single-threaded scalar 2894 |||||||||||
Stdlib Write
single-threaded scalar 2480 |||||||||
Stdlib Copy
single-threaded scalar 4932 |||||||||||||||||||
Stream
Stream Copy
single-threaded scalar 4682 ||||||||||||||||||
single-threaded vector 7695 ||||||||||||||||||||||||||||||
Stream Scale
single-threaded scalar 4771 |||||||||||||||||||
single-threaded vector 7355 |||||||||||||||||||||||||||||
Stream Add
single-threaded scalar 5684 ||||||||||||||||||||||
single-threaded vector 7617 ||||||||||||||||||||||||||||||
Stream Triad
single-threaded scalar 6076 ||||||||||||||||||||||||
single-threaded vector 5808 |||||||||||||||||||||||
Integer Score: 1713 ||||||
Floating Point Score: 3497 |||||||||||||
Memory Score: 4299 |||||||||||||||||
Stream Score: 6211 ||||||||||||||||||||||||
Overall Geekbench Score: 3304 |||||||||||||
Trying again showed slgithly worse results but is nontheless pretty decent:
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 i686
Model: Linux PC (QEMU Virtual CPU version (cpu64-rhel6))
Motherboard: Unknown Motherboard
Processor: QEMU Virtual CPU version (cpu64-rhel6)
Processor ID: GenuineIntel Family 6 Model 13 Stepping 3
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: 0.00 B
Bus Frequency: 0.00 Hz
Memory: 122 MB
Memory Type: N/A
SIMD: 1
BIOS: N/A
Processor Model: QEMU Virtual CPU version (cpu64-rhel6)
Processor Cores: 1
Integer
Blowfish
single-threaded scalar 1264 |||||
multi-threaded scalar 1383 |||||
Text Compress
single-threaded scalar 1733 ||||||
multi-threaded scalar 1677 ||||||
Text Decompress
single-threaded scalar 1815 |||||||
multi-threaded scalar 1869 |||||||
Image Compress
single-threaded scalar 1451 |||||
multi-threaded scalar 1424 |||||
Image Decompress
single-threaded scalar 1414 |||||
multi-threaded scalar 1459 |||||
Lua
single-threaded scalar 2408 |||||||||
multi-threaded scalar 2412 |||||||||
Floating Point
Mandelbrot
single-threaded scalar 1660 ||||||
multi-threaded scalar 1688 ||||||
Dot Product
single-threaded scalar 2716 ||||||||||
multi-threaded scalar 2881 |||||||||||
single-threaded vector 3136 ||||||||||||
multi-threaded vector 3625 ||||||||||||||
LU Decomposition
single-threaded scalar 2053 ||||||||
multi-threaded scalar 2063 ||||||||
Primality Test
single-threaded scalar 2903 |||||||||||
multi-threaded scalar 2351 |||||||||
Sharpen Image
single-threaded scalar 6653 ||||||||||||||||||||||||||
multi-threaded scalar 6720 ||||||||||||||||||||||||||
Blur Image
single-threaded scalar 5209 ||||||||||||||||||||
multi-threaded scalar 5220 ||||||||||||||||||||
Memory
Read Sequential
single-threaded scalar 4264 |||||||||||||||||
Write Sequential
single-threaded scalar 7724 ||||||||||||||||||||||||||||||
Stdlib Allocate
single-threaded scalar 3053 ||||||||||||
Stdlib Write
single-threaded scalar 2485 |||||||||
Stdlib Copy
single-threaded scalar 4930 |||||||||||||||||||
Stream
Stream Copy
single-threaded scalar 4272 |||||||||||||||||
single-threaded vector 6731 ||||||||||||||||||||||||||
Stream Scale
single-threaded scalar 4414 |||||||||||||||||
single-threaded vector 5691 ||||||||||||||||||||||
Stream Add
single-threaded scalar 5003 ||||||||||||||||||||
single-threaded vector 7183 ||||||||||||||||||||||||||||
Stream Triad
single-threaded scalar 6200 ||||||||||||||||||||||||
single-threaded vector 5750 |||||||||||||||||||||||
Integer Score: 1692 ||||||
Floating Point Score: 3491 |||||||||||||
Memory Score: 4491 |||||||||||||||||
Stream Score: 5655 ||||||||||||||||||||||
Overall Geekbench Score: 3277 |||||||||||||
Customer Service and Support
I have dealt with the customer service quite a few times and Salvatore is a very pleasant person to speak with, although the ticket response time might be a little longer than some of the top providers in the LEB market (for example, one of my technical tickets was submitted at 11:47pm and was not responded until 2:16am, which, considering it is dawn in Europe, is actually pretty reasonable, sales ticket takes a bit longer though, a ticket was sent in at 9:56pm was not responded until 1:41am the next day) , his personality has definitely earned him a few points in the low end community. Furthermore, instead of getting a canned response like “please hold on while I look into this”, I have yet to respond to any ticket again other than confirming that I got what I need.
Conclusion
Overall, Prometeus has left with me a pretty good impression. Granted, their network does not seems to serve the North American market well (particularly the west coast), however the VPS is pretty stable and the overall performance, particularly disk I/O, is quite impressive for such a cheap VPS. Furthermore, although not the fastest ticket responses that I have seen, the pleasant personality of the owner, together with the no-nonsense support, has definitely made it a good candidate for a VPS that does not require heavy data transmission.
Congratulations for the great post, and thank you for spending money just to give good reviews for the community.
I have 4 VPS with prometeus, 2 OVZ and 2 KVM. And I think your review is on point, and I share exactly the same comment. I love prometeus, but there are still some areas that can be improved =)
How about reviewing Prometeus OVZ products?
@jcaleb: will definitely include that in the pipeline!
Pingback: [EU]Prometeus Autumn-Winter collection, OpenVZ SSD VPS with a huge 60% OFF « « Daily VPS OffersDaily VPS Offers
Pingback: [EU]Prometeus, XEN, OPENVZ. KVM SSD PLANS! 35% RECURRING, 30GB SSD/1GB RAM @ €9.75/MO | whatisvps.info
Pingback: Prometeus Bundle Offer – 512MB OpenVZ + 128MB OpenVZ for €35.90/yr in Milan, Italy – Low End Box
Pingback: 96MB Low End VPS Community Member Review Part 1 - Prometeus OpenVZ SSD - 96MB.com
Pingback: [EU]Prometeus, 15% TO 40% DISCOUNT AND A FREE MYSQL OFFLOAD ACCOUNT FOR LIFE! | whatisvps.info
Pingback: [EU]Prometeus, 15% TO 40% DISCOUNT! KVM XEN AND OPENVZ, CHOOSE YOUR VPS TODAY!!! | whatisvps.info
Pingback: Prometeus - 15% OFF KVM 2GB RAM, 50GB SAS DISK, 4TB BW - $16.15/mo - Italy - Node Deals | Node Deals