Update Vultr

Matt

Owner
Going well so far. Site is actually faster according to the debug stats than it was when running on Linode and DigitalOcean.

upload_2015-2-9_13-17-16.png

Code:
# siege https://mattwservices.co.uk -c 1000 -d 30 -t 1M
** SIEGE 2.70
** Preparing 1000 concurrent users for battle.
The server is now under siege...
Lifting the server siege...      done.
Transactions:                   3479 hits
Availability:                 100.00 %
Elapsed time:                  59.94 secs
Data transferred:              42.34 MB
Response time:                  0.47 secs
Transaction rate:              58.04 trans/sec
Throughput:                     0.71 MB/sec
Concurrency:                   27.07
Successful transactions:        3479
Failed transactions:               0
Longest transaction:            5.46
Shortest transaction:           0.10
 

Matt

Owner
I tested the site on their London facility. The page generation increased from 0.05 on the forum home to ~0.17/0.2, so almost 4 times slower than in their France site.
 

Matt

Owner
I think this is due to the increase latency between the VMs

France
Code:
[root@nginx public]# ping 10.99.0.11        
PING 10.99.0.11 (10.99.0.11) 56(84) bytes of data.
64 bytes from 10.99.0.11: icmp_seq=1 ttl=64 time=0.450 ms
64 bytes from 10.99.0.11: icmp_seq=2 ttl=64 time=0.375 ms
64 bytes from 10.99.0.11: icmp_seq=3 ttl=64 time=0.375 ms
64 bytes from 10.99.0.11: icmp_seq=4 ttl=64 time=0.361 ms
64 bytes from 10.99.0.11: icmp_seq=5 ttl=64 time=0.442 ms
64 bytes from 10.99.0.11: icmp_seq=6 ttl=64 time=0.435 ms
64 bytes from 10.99.0.11: icmp_seq=7 ttl=64 time=0.453 ms
^C
--- 10.99.0.11 ping statistics ---
7 packets transmitted, 7 received, 0% packet loss, time 6295ms
rtt min/avg/max/mdev = 0.361/0.413/0.453/0.037 ms
[root@nginx public]#
London
Code:
[root@nginx ~]# ping 10.99.0.11
PING 10.99.0.11 (10.99.0.11) 56(84) bytes of data.
64 bytes from 10.99.0.11: icmp_seq=1 ttl=64 time=0.647 ms
64 bytes from 10.99.0.11: icmp_seq=2 ttl=64 time=0.700 ms
64 bytes from 10.99.0.11: icmp_seq=3 ttl=64 time=0.691 ms
64 bytes from 10.99.0.11: icmp_seq=4 ttl=64 time=0.668 ms
64 bytes from 10.99.0.11: icmp_seq=5 ttl=64 time=0.766 ms
64 bytes from 10.99.0.11: icmp_seq=6 ttl=64 time=0.614 ms
64 bytes from 10.99.0.11: icmp_seq=7 ttl=64 time=0.861 ms
64 bytes from 10.99.0.11: icmp_seq=8 ttl=64 time=0.715 ms
64 bytes from 10.99.0.11: icmp_seq=9 ttl=64 time=0.620 ms
 

Matt

Owner
I've just moved my final customer OFF Vultr after nothing but problems. They randomly reload their host nodes without telling you, they have broken their own private networking (and changed how you need to configure it....without telling people), so enough is enough.
 

eva2000

Customer
I've just moved my final customer OFF Vultr after nothing but problems. They randomly reload their host nodes without telling you, they have broken their own private networking (and changed how you need to configure it....without telling people), so enough is enough.
haven't had that happen but Vultr has lowered their cpu specs from 3.4+Ghz to 2.4Ghz and lower disk performing configs too. Everytime i have tested London Vultr it's been slower performing too - currently have a London Vultr location for my new centminmod.com cluster I am building across several geographical locations including Los Angeles, Seattle, Dallas, New York, New Jersey, London, Sydney and Singapore.

So Matt who are you using for London VPS now ?
 

eva2000

Customer
I see $20/month bit beyond my budget seeing as each geographic region I will have my own load balanced cluster too :)
 

Matt

Owner
This is currently on a £36 a month VPS from them (my dev server). I'm going to message you some info about the specific issues with Vultr.
 

eva2000

Customer
This is currently on a £36 a month VPS from them (my dev server). I'm going to message you some info about the specific issues with Vultr.
cheers Matt
eva2000 any good tut for creating clusters? I am a total noob when it comes to these things :D
there's so many ways of doing this so really comes down to googling and practise - load balancing can be done from nginx or haproxy itself, dns level need geodns support like with dnsmadeeasy or route53 and mysql from replication, galera cluster, mysql cluster, tungsten replicator replication etc.

i am using private git backed repository to keep configuration management easier to deploy my site to new VPS cluster nodes - takes me 10-20mins to install Centmin Mod .08 beta03 latest LEMP stack and 5-10min to setup site nginx/spdy ssl and deploy another git repository for site files. Centmin Mod .08 beta03 has new nv command line to create new domains and sites so you set a up new domain with just one line
Code:
/usr/bin/nv -d newdomain.com -s y -u ftpusername
so to install centmin mod + new site domain newdomain.com in 2 command lines
Code:
curl -sL http://centminmod.com/betainstaller.sh | bash;
/usr/bin/nv -d newdomain.com -s y -u ftpusername
 
Last edited:
Top