For the last eight months I’ve had a server at So You Start (SYS) which is a division of OVH, and is meant to be an entry level service. This harware server (not virtual, hardware) costs me about $60 a month, has a eight cores (hyperthreading), 32GB of ECC ram, a 250MB Internet connection, and 2x2TB disks. Also it’s in Canada, which, as a Canadian, suits me great.
I have recently given up this server, but only because I wasn’t using it, not because it didn’t work for me. In fact quite the opposite–it was working great. It was a great deal, real hardware (albeit a bit old), and it was up pretty much the whole darn time. I wish I could have kept it because it was a good deal, especially considering it’s probably 2/3 of the price most people (at least in Canada) pay for their cell phone.
I really don’t have any complaints. Most of the below notes are minor issues or positives.
Always seem to be sold out of all the different models of hardware they have (I think this is because they are very popular). You’d almost have to write a script to check the site. This is the opposite of the OVH “business” site on which most of the servers are available within 120 seconds.
There is no KVM attached. This is pretty hair-raising if when updating the server. It seems they have a KVM service, which is $30 for 24 hours, and they say they will attach it within two hours. This would make it pretty hard to run a “real” production service off So You Start, but I don’t think that’s the point.
If you need any help from So You Start, you probably won’t get it in any timely fashion. But, I’m just guessing based on forum posts and reputation–I never asked for help from them so can’t say from experience.
This is my own fault, but I should have got an SSD based server. Spinning disks are so slow.
Billing: They could not do recurring credit card payments. You could pay for one or several months in a row, but not recurring. I always felt like if I forgot to pay, just once, my server would be deleted and given to someone else. I’d have preferred to have recurring CC payments, but you could do multi-months.
FTP backup: They do provide 100GB of FTP backup which I used. I gpg-zipped my backup files and uploaded them to the FTP server. If I was running a production service I’d find a way to go off-site with those backups, but it’s great that So You Start provides 100Gb of backup.
No 2-factor authentication to the admin panel, so if your login was compromised then your server is owned. I’m not even sure if the higher end OVH service provides this, not that they should differentiate on that feature. Every service should 2-factor their logins.
ECC memory: my server was ECC, but many of the So You Start models are not. On the range page they do incidate which servers are ECC and the rest are not. This is something to pay attention to if you want to avoid memory errors cropping up. As of this writing it seems like none of the models with SSDs are ECC.
root# dmidecode --type 16
# dmidecode 2.12
SMBIOS 2.7 present.
Handle 0x0028, DMI type 16, 23 bytes
Physical Memory Array
Location: System Board Or Motherboard
Use: System Memory
Error Correction Type: Single-bit ECC
Maximum Capacity: 32 GB
Error Information Handle: Not Provided
Number Of Devices: 4
root# dmidecode --type 17
# dmidecode 2.12
SMBIOS 2.7 present.
Handle 0x002B, DMI type 17, 34 bytes
Memory Device
Array Handle: 0x0028
Error Information Handle: Not Provided
Total Width: 72 bits
Data Width: 64 bits
Size: 8192 MB
Form Factor: DIMM
Set: None
Locator: DIMM_A2
Bank Locator: BANK 0
Type: DDR3
Type Detail: Synchronous
Speed: 1600 MHz
Manufacturer: Samsung
Serial Number: 378CA173
Asset Tag: 9876543210
Part Number: M391B1G73QH0-YK0
Rank: 2
Configured Clock Speed: 1600 MHz
SNIP!
I had this server for 8 months. It had ~18 minutes of downtime, about 15 of which were my own fault. Otherwise, I had pingdom monitoring it for the last four months, and it had 7 small outages, most less than a few seconds, and none that I recieved an email notice from Pingdom about.
I believe this system is the E3-SAT-3 model, though it wasn’t called that when I first purchased the service. Also I believe mine is actually ECC memory, though the E3-SAT-3 model isn’t supposed to be.
For CPUs it has eight cores: four physical, four hyperthreading. I’m just showing one core below.
processor : 7
vendor_id : GenuineIntel
cpu family : 6
model : 58
model name : Intel(R) Xeon(R) CPU E3-1245 V2 @ 3.40GHz
stepping : 9
microcode : 0x15
cpu MHz : 3401.000
cache size : 8192 KB
physical id : 0
siblings : 8
core id : 3
cpu cores : 4
apicid : 7
initial apicid : 7
fpu : yes
fpu_exception : yes
cpuid level : 13
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 ht tm pbe syscall nx rdtscp lm constant_tsc arch_perfmon pebs bts rep_good nopl xtopology nonstop_tsc aperfmperf eagerfpu pni pclmulqdq dtes64 monitor ds_cpl vmx smx est tm2 ssse3 cx16 xtpr pdcm pcid sse4_1 sse4_2 x2apic popcnt tsc_deadline_timer aes xsave avx f16c rdrand lahf_lm ida arat epb xsaveopt pln pts dtherm tpr_shadow vnmi flexpriority ept vpid fsgsbase smep erms
bogomips : 6784.57
clflush size : 64
cache_alignment : 64
address sizes : 36 bits physical, 48 bits virtual
power management:
For memory:
MemTotal: 32902800 kB MemFree: 474852 kB Buffers: 179312 kB Cached: 29534464 kB SwapCached: 37392 kB Active: 18247988 kB Inactive: 12369164 kB Active(anon): 551720 kB Inactive(anon): 499556 kB Active(file): 17696268 kB Inactive(file): 11869608 kB Unevictable: 65104 kB Mlocked: 65104 kB SwapTotal: 5118972 kB SwapFree: 4963700 kB Dirty: 200 kB Writeback: 0 kB AnonPages: 931436 kB Mapped: 125728 kB Shmem: 141504 kB Slab: 1149056 kB SReclaimable: 985140 kB SUnreclaim: 163916 kB KernelStack: 4608 kB PageTables: 32776 kB NFS_Unstable: 0 kB Bounce: 0 kB WritebackTmp: 0 kB CommitLimit: 21570372 kB Committed_AS: 7632684 kB VmallocTotal: 34359738367 kB VmallocUsed: 522704 kB VmallocChunk: 34358922952 kB HardwareCorrupted: 0 kB AnonHugePages: 397312 kB HugePages_Total: 0 HugePages_Free: 0 HugePages_Rsvd: 0 HugePages_Surp: 0 Hugepagesize: 2048 kB DirectMap4k: 166884 kB DirectMap2M: 33341440 kB </pre>## Disk IO Pretty much looks like a 2x2TB SATA mirror. Bleh. But again, expected. I usually take disk IO testing pretty seriously, but I just chucked this test together in a couple minutes, so take it as you will.# cat *.fio [random-read] rw=randread size=128m directory=/tmp/fio-testing/data [random-write1mb] rw=randwrite size=128m directory=/tmp/fio-testing/data direct=1 bs=1m [random-write] rw=randwrite size=128m directory=/tmp/fio-testing/data direct=1Random read:Random writes 4k blocksize.# fio randread.fio random-read: (g=0): rw=randread, bs=4K-4K/4K-4K/4K-4K, ioengine=sync, iodepth=1 fio-2.1.3 Starting 1 process random-read: Laying out IO file(s) (1 file(s) / 128MB) SNIP! random-read: (groupid=0, jobs=1): err= 0: pid=10023: Mon Nov 3 21:22:14 2014 read : io=131072KB, bw=1009.3KB/s, iops=252, runt=129874msec clat (usec): min=52, max=236152, avg=3955.39, stdev=4034.99 lat (usec): min=52, max=236152, avg=3955.79, stdev=4035.00 clat percentiles (usec): | 1.00th=[ 55], 5.00th=[ 77], 10.00th=[ 114], 20.00th=[ 119], | 30.00th=[ 126], 40.00th=[ 2416], 50.00th=[ 3696], 60.00th=[ 4960], | 70.00th=[ 6240], 80.00th=[ 7520], 90.00th=[ 8768], 95.00th=[ 9408], | 99.00th=[10176], 99.50th=[10304], 99.90th=[18560], 99.95th=[31616], | 99.99th=[96768] bw (KB /s): min= 506, max= 4221, per=98.48%, avg=993.71, stdev=325.57 lat (usec) : 100=6.09%, 250=27.71%, 500=0.05%, 750=0.04%, 1000=0.11% lat (msec) : 2=2.93%, 4=15.37%, 10=45.83%, 20=1.77%, 50=0.06% lat (msec) : 100=0.02%, 250=0.01% cpu : usr=0.28%, sys=0.98%, ctx=33070, majf=0, minf=28 IO depths : 1=100.0%, 2=0.0%, 4=0.0%, 8=0.0%, 16=0.0%, 32=0.0%, >=64=0.0% submit : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%, >=64=0.0% complete : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%, >=64=0.0% issued : total=r=32768/w=0/d=0, short=r=0/w=0/d=0 Run status group 0 (all jobs): READ: io=131072KB, aggrb=1009KB/s, minb=1009KB/s, maxb=1009KB/s, mint=129874msec, maxt=129874msec Disk stats (read/write): sda: ios=32491/327, merge=0/601, ticks=128180/16464, in_queue=144612, util=98.77%Rand writes 1mb blocksize:# fio randwrite.fio random-write: (g=0): rw=randwrite, bs=4K-4K/4K-4K/4K-4K, ioengine=sync, iodepth=1 fio-2.1.3 Starting 1 process random-write: Laying out IO file(s) (1 file(s) / 128MB) Jobs: 1 (f=1): [w] [100.0% done] [0KB/472KB/0KB /s] [0/118/0 iops] [eta 00m:00s] random-write: (groupid=0, jobs=1): err= 0: pid=11028: Mon Nov 3 21:45:39 2014 write: io=131072KB, bw=484491B/s, iops=118, runt=277028msec clat (msec): min=1, max=227, avg= 8.45, stdev= 6.98 lat (msec): min=1, max=227, avg= 8.45, stdev= 6.98 clat percentiles (msec): | 1.00th=[ 3], 5.00th=[ 4], 10.00th=[ 5], 20.00th=[ 6], | 30.00th=[ 7], 40.00th=[ 8], 50.00th=[ 8], 60.00th=[ 9], | 70.00th=[ 9], 80.00th=[ 10], 90.00th=[ 10], 95.00th=[ 11], | 99.00th=[ 40], 99.50th=[ 43], 99.90th=[ 74], 99.95th=[ 95], | 99.99th=[ 147] bw (KB /s): min= 184, max= 591, per=100.00%, avg=473.57, stdev=48.38 lat (msec) : 2=0.28%, 4=7.89%, 10=85.24%, 20=2.71%, 50=3.70% lat (msec) : 100=0.15%, 250=0.03% cpu : usr=0.15%, sys=0.73%, ctx=33273, majf=0, minf=27 IO depths : 1=100.0%, 2=0.0%, 4=0.0%, 8=0.0%, 16=0.0%, 32=0.0%, >=64=0.0% submit : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%, >=64=0.0% complete : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%, >=64=0.0% issued : total=r=0/w=32768/d=0, short=r=0/w=0/d=0 Run status group 0 (all jobs): WRITE: io=131072KB, aggrb=473KB/s, minb=473KB/s, maxb=473KB/s, mint=277028msec, maxt=277028msec Disk stats (read/write): sda: ios=0/33913, merge=0/5241, ticks=0/329844, in_queue=329792, util=99.17%## Bandwidth Using the speedtest-cli, which is very handy. Only found out about it today. Closest speedtest server:# fio randwrite1mb.fio random-write1mb: (g=0): rw=randwrite, bs=1M-1M/1M-1M/1M-1M, ioengine=sync, iodepth=1 fio-2.1.3 Starting 1 process random-write1mb: Laying out IO file(s) (1 file(s) / 128MB) Jobs: 1 (f=1) random-write1mb: (groupid=0, jobs=1): err= 0: pid=11513: Mon Nov 3 21:50:09 2014 write: io=131072KB, bw=75199KB/s, iops=73, runt= 1743msec clat (msec): min=8, max=26, avg=13.56, stdev= 2.34 lat (msec): min=8, max=26, avg=13.61, stdev= 2.33 clat percentiles (usec): | 1.00th=[ 8256], 5.00th=[ 9792], 10.00th=[10688], 20.00th=[11712], | 30.00th=[12352], 40.00th=[13120], 50.00th=[13760], 60.00th=[14400], | 70.00th=[14784], 80.00th=[15040], 90.00th=[15808], 95.00th=[16192], | 99.00th=[19840], 99.50th=[26240], 99.90th=[26240], 99.95th=[26240], | 99.99th=[26240] bw (KB /s): min=73142, max=76749, per=99.63%, avg=74922.67, stdev=1803.93 lat (msec) : 10=6.25%, 20=92.97%, 50=0.78% cpu : usr=0.40%, sys=0.69%, ctx=135, majf=0, minf=27 IO depths : 1=100.0%, 2=0.0%, 4=0.0%, 8=0.0%, 16=0.0%, 32=0.0%, >=64=0.0% submit : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%, >=64=0.0% complete : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%, >=64=0.0% issued : total=r=0/w=128/d=0, short=r=0/w=0/d=0 Run status group 0 (all jobs): WRITE: io=131072KB, aggrb=75199KB/s, minb=75199KB/s, maxb=75199KB/s, mint=1743msec, maxt=1743msec Disk stats (read/write): sda: ios=0/241, merge=0/0, ticks=0/2876, in_queue=2900, util=92.62%# speedtest Retrieving speedtest.net configuration... Retrieving speedtest.net server list... Testing from OVH Hosting ()... Selecting best server based on latency... Hosted by 3Men@Work (Montreal, QC) [2.38 km]: 39.06 ms Testing download speed........................................ Download: 358.12 Mbits/s Testing upload speed.................................................. Upload: 58.36 Mbits/s </code> </pre> New York City! # speedtest --server 5029 Retrieving speedtest.net configuration... Retrieving speedtest.net server list... Testing from OVH Hosting ()... Hosted by AT&T (New York City, NY) [533.43 km]: 45.257 ms Testing download speed........................................ Download: 384.99 Mbits/s Testing upload speed.................................................. Upload: 18.05 Mbits/s </code> </pre> My hometown ISP, Shaw. Good old monopolies. # speedtest --list | grep -i "edmonton" 4242) Shaw Communications (Edmonton, AB, Canada) [2972.64 km] 3050) Telus (Edmonton, AB, Canada) [2972.64 km] 1051) Tera-byte Dot Com Inc (Edmonton, AB, Canada) [2972.64 km] root@vurt01:~# speedtest --server 4242 Retrieving speedtest.net configuration... Retrieving speedtest.net server list... Testing from OVH Hosting ()... Hosted by Shaw Communications (Edmonton, AB) [2972.64 km]: 126.014 ms Testing download speed........................................ Download: 118.54 Mbits/s Testing upload speed.................................................. Upload: 38.73 Mbits/s </code> </pre> Toronto: # speedtest --server 3575 Retrieving speedtest.net configuration... Retrieving speedtest.net server list... Testing from OVH Hosting ()... Hosted by TELUS (Toronto, ON) [504.80 km]: 65.678 ms Testing download speed........................................ Download: 126.21 Mbits/s Testing upload speed.................................................. Upload: 45.24 Mbits/s </code> </pre> ## Conclusion This server worked great for me. I had no problems. However, if I was thinking of hosting a production server, I would probably go with the OVH "business" servers which are more money, but have additional features like full-time KVM and a virtual network that can be setup between servers. Also I would like SSDs and ECC memory, a combination So You Start doesn't seem to provide. But, given my positive experience at So You Start, I would certainly give it a try at OVH, especially because they are in Quebec!