Server migration

Could be a cookie problem geoffw - notwithstanding your avatar's predilection for such.

Problems with being logged in when you shouldn't be, or not being logged in when you should ... are usually cookie related.

Getting "logged off" when clicking New Posts is a frequent occurrence- I get that several times a day. However, if I click on Quick Links and then Today's Posts I'm logged in again- then clicking on New Posts again shows me as logged in.

It behaves as if I have one session time (I'm not sure, perhaps 6 hours) applying to New Posts and another much longer session time- some days- applying to the forum as a whole.

I don't know if this is restricted to iPhone. I do 98% of my forum browsing from my phone.
 
This appears to be an issue with Chrome. I've switched back to Safari and I don't have the same problems.

Some of the features of Chrome have now been incorporated into Safari so the difference is not as big- those features just work differently.
 
Was it always hosted in Japan?
Definitely quicker.

It's been in Japan for a while now.

The changes which have happened recently (first the web/application server migration and then last night the database server) was to new server infrastructure which has just been deployed which uses SSD drives - has quite a significant performance boost for IO intensive processes. Also the new servers have faster CPUs, which helps too.
 
i thought to myself wow this site loads quick tyhese days then found out why.
all websites should be hosted on ssd's
 
Who pays for all this Sim?

For the hosting itself - I pay for the infrastructure (I run other websites as well), and Jan and Ian contribute towards the costs of keeping Somersoft going out of their own pocket.

There's quite a bit involved to keep a busy site like this operational:

- web/application server (PHP / nginx)
- 2x database servers (primary write + secondary read, MySQL)
- mail server (inbound mail for support queries, Postfix + Dovecot)
- search server (ElasticSearch - coming soon!)
- cache server (Redis)
- mail deliverability service (Mandrill)
- anti-spam service (Akismet)
- backup storage service (Amazon S3)
- CDN service (Amazon Cloudfront)
- DNS management service (dnz.io)
- help desk system for support emails (HelpSpot)
- system monitoring service (Longview)
- log and event monitoring service (HipChat)
- system uptime monitoring service (Pingdom)
- IP address geolocation service (MaxMind)
- source code hosting service (Bitbucket)
- software deployment service (DeployHQ)
- chatroom service (Mibbit)

Some of these services are shared with other sites I host as well, which helps offset the costs.

It is all vastly different from the tiny website we first started with on shared hosting. But then, the traffic (and spam!) is vastly higher than it was back then too!
 
For the hosting itself - I pay for the infrastructure (I run other websites as well), and Jan and Ian contribute towards the costs of keeping Somersoft going out of their own pocket.

There's quite a bit involved to keep a busy site like this operational:

- web/application server (PHP / nginx)
- 2x database servers (primary write + secondary read, MySQL)
- mail server (inbound mail for support queries, Postfix + Dovecot)
- search server (ElasticSearch - coming soon!)
- cache server (Redis)
- mail deliverability service (Mandrill)
- anti-spam service (Akismet)
- backup storage service (Amazon S3)
- CDN service (Amazon Cloudfront)
- DNS management service (dnz.io)
- help desk system for support emails (HelpSpot)
- system monitoring service (Longview)
- log and event monitoring service (HipChat)
- system uptime monitoring service (Pingdom)
- IP address geolocation service (MaxMind)
- source code hosting service (Bitbucket)
- software deployment service (DeployHQ)
- chatroom service (Mibbit)

Some of these services are shared with other sites I host as well, which helps offset the costs.

It is all vastly different from the tiny website we first started with on shared hosting. But then, the traffic (and spam!) is vastly higher than it was back then too!

Interesting setup.
Utilize a bunch of those services myself :)
 
There's quite a bit involved to keep a busy site like this operational:

- web/application server (PHP / nginx)
- 2x database servers (primary write + secondary read, MySQL)
- mail server (inbound mail for support queries, Postfix + Dovecot)
- search server (ElasticSearch - coming soon!)
- cache server (Redis)
- mail deliverability service (Mandrill)
- anti-spam service (Akismet)
- backup storage service (Amazon S3)
- CDN service (Amazon Cloudfront)
- DNS management service (dnz.io)
- help desk system for support emails (HelpSpot)
- system monitoring service (Longview)
- log and event monitoring service (HipChat)
- system uptime monitoring service (Pingdom)
- IP address geolocation service (MaxMind)
- source code hosting service (Bitbucket)
- software deployment service (DeployHQ)
- chatroom service (Mibbit)

Wow - that's quite a list of servers and services. I certainly second jackbak's thanks to you, the other moderators and Jan and Ian. Interesting to know all the hard work that goes on behind the scenes.

Cheers,
GreenGoblin
 
Back
Top