Separate DB server

[Pushpendra Singh Chauhan]

Hi,

I want to separate the DB server. Is there a way to change the database location in Admin Panel to point to the SQL server?
or where should i change the SQL Server Path?

[Martin A.]

The database details are stored in the file conf_global.php which is located in your installation root folder.

[Lucas James]

5 hours ago, Pushpendra Singh Chauhan said:

I want to separate the DB server

FYI: You might experience some lag/latency issue if your DB server is not hosted on the same network.

[ASTRAPI]

@Pushpendra Singh Chauhan

Don't forget to disable at your my.cnf file the skip name resolve:

#skip-name-resolve

then restart Mysql...

It will be great if users will be able to set that kind of options from the admin panel or if not possible to download a file and replace the existing one as they already do with constants.php.

Also for Elastic search shards for example...

[bfarber]

It isn't the software's responsibility to generate a my.cnf or Elasticsearch configuration file. There is no one size fits all for this sort of thing, and MOST of our self-hosted clients are on shared servers where they can't control these things either.

If you are going to host yourself, you're responsibility for the configuration of the hosting. 😉 If you want things more automated, use our Community In The Cloud services.

[All Astronauts]

44 minutes ago, bfarber said:

MOST of our self-hosted clients are on shared servers

I don't want you to spill out internal numbers or anything but would you care to throw a percentage on that? Personally I've yet to directly work (support) a shared server since... maybe 4.2?

Also, frightening - unless you are counting VPS numbers in that shared category. You're not though, right?

[ASTRAPI]

Quote

There is no one size fits all for this sort of thing, and MOST of our self-hosted clients are on shared servers where they can't control these things either.

On that case of shared hosting it's the same when they want to update the constants file... Anyway...

Not sure why they are selecting a shared hosting and not get at least a vps for 5$ or 10$ per month that will offer a lot more performance but anyway.

It was just a proposal 🙂

 

[Lucas James]

7 hours ago, ASTRAPI said:

Not sure why they are selecting a shared hosting and not get at least a vps

Because majority of them are simply unaware of a Shared hosting vs VPS vs Dedicated server – and they tend to adopt the cheaper/easy solution.

[Rhett]

9 hours ago, All Astronauts said:

I don't want you to spill out internal numbers or anything but would you care to throw a percentage on that? Personally I've yet to directly work (support) a shared server since... maybe 4.2?

Also, frightening - unless you are counting VPS numbers in that shared category. You're not though, right?

A VPS is still sharing resources, they may not market it that way, but it is. You may have allocations of Disk i/o, Ram and CPU, most all other resources are still shared, which can still impact performance.  There are also many many people on typical shared hosting as well, most hosting providers do use some form of limitations on the number of processes, Disk i/o, inodes etc to help control the impact on other users.  

On 12/18/2019 at 4:24 PM, ASTRAPI said:

@Pushpendra Singh Chauhan

Don't forget to disable at your my.cnf file the skip name resolve:

#skip-name-resolve

then restart Mysql...

It will be great if users will be able to set that kind of options from the admin panel or if not possible to download a file and replace the existing one as they already do with constants.php.

Also for Elastic search shards for example...

You don't need to do that as long as you use the IP, which is recommenced as it's better to skip the overhead of a lookup/resolve. 

[ASTRAPI]

Separate database server will give you concurrency more than response times/throughput which would be slower than if you had database server on the same server.

[The Old Man]

Blimey, I thought the days of hosting IPS on less than a VPS were long gone since v4 was released?

Just for fun, I always wonder what kind of quality VPS can you get for $5-10 a month? I'm picturing an old *Kellogg's cardboard cereal box with some spaghetti network cables sticking out the back.

 

*Other cereals are available.

[jair101]

3 hours ago, The Old Man said:

Just for fun, I always wonder what kind of quality VPS can you get for $5-10 a month? I'm picturing an old *Kellogg's cardboard cereal box with some spaghetti network cables sticking out the back.

 

In this price range there are many providers that scrape the barrel and some are outright scams. 

But there are reasonable offers as well. Not everyone needs 99.999% SLA and massive resources 🙂 For $10 you can get 2xCPU cores and 2GB ram with about 50GB SSD storage from a reputable company. Pretty solid for small to light-medium community. Only caveat is that it is unmanaged, you need to be competent enough for the day to day administration and have solid backup strategy in case the company or the hardware go sideways. 

[ASTRAPI]

As it is not allowed i can't post here a ton of reputable providers for a 5$ starting price and of course KVM based !

For me the 10$ plans are recommended for anyone that needs a starting Vps plan...

2GB ram
2 cores 3+ Ghz clock
50GB space Ssd based

Dedicated starting servers (very reputable provider) are very cheap also like:

CPU: 1x Intel® Xeon® E3 1220 v2
RAM: 16 GB DDR3 ECC
Storage: 2x 1 TB SATA
Bandwidth: Basic Unmetered 1 Gbit/sec

35 Euro only - (it is the normal price and not a special offer)

[Adlago]

I am disappointed with the host providers.
Is there a host provider guaranteeing a first byte below 200ms?
The IPS  suite works very well - but I have yet to find a suitable first-byte host host provider.
VPN is not an offer to me. I also don't have time to deal with servers.
Now I pay enough for a good server - but the support exhales- it plays the first byte in the 90 to 2000 ms range and since my pay date is coming, now I'm terrorizing them ...
Why is this happening - it's easy to make offers, but who and how protects the client?

[ASTRAPI]

The time to first byte refers to how long it takes for visitors to receive the first byte of data after requesting your URL. The time to start render is the point at which the user’s browser actually begins displaying content. The former is largely dependt on your server setup, but the latter depends more on how your CSS is structured.

That said, browsers won’t start rendering until they receive data, so a slow time to first byte will obviously push back your time to start render. Therefore, you should prioritize resolving any potential issues with your server before you can reap the full benefits of optimizing your CSS performance.

How CSS Performance Affects Time to Start Render

Before a browser can start laying out a webpage’s content, it needs instructions in the form of HTML and CSS. Thus, rendering cannot begin until all external stylesheets have been downloaded and processed. The more round trips this requires, the longer visitors have to wait.

Using external CSS involves making one or more HTTP requests, so your goal should be to minimize the number of required requests as much as possible. For example, putting your plugin, banner, and layout link styles into a single .css file can significantly speed up your time to first render.

 

TTFB factors

• Geographically relative source and target distance = how close test/visitor is to your server
• DNS, server connection & response time
• MySQL or database backend performance/settings
• Server hardware

but TTFB is not the only thing that you must focus as it is just the starting point. You should care more about initial render time.

For example if you get a 100ms TTFB and the initial render time is 3 seconds then the result is not what you may expected.

For dynamic web sites is harder to get a very low TTFB but with static pages is very easy especially if you use Cloudflare in front of your web server as you will always hit there cache.

 

 

[Adlago]

In this case, I refer to the waiting time as part of the first byte.
This is disturbing - I have 70% waiting time in the 90-180 ms interval, and 30% with a strange variable over a 200 to 2000 ms wide range ....
Here my settings have no effect but only a response to the server.
My host plan is expensive and I expect the host company to fulfill contract parameters - it doesn't work ...

[ASTRAPI]

Quote

In this case, I refer to the waiting time as part of the first byte.

The waiting time is related to:

TTFB factors

• Geographically relative source and target distance = how close test/visitor is to your server
• DNS, server connection & response time
• MySQL or database backend performance/settings
• Server hardware

Quote

My host plan is expensive and I expect the host company to fulfill contract parameters

Expensive is not something that will guaranty low TTFB. If they promise you that you will get a low TTFB because you will get a good server, that will never happen as i explain above the factors.

Examples of why it can't be done:

For example if you are hosting your nameservers/DNS: You can't compare that by using a reputable DNS provider like:

If for example your server is at Europe let's say Italy and your visitors in the US you can't get a TTFB at 200ms by hosting yourself your DNS.

Optimization of your Mysql is very important if you are testing a dynamic web site.

Another factor is your ssl certificate. Most users use Let's Encrypt RSA 2048bit  (that is great) but they have for all visitors the same certificate.

But not all visitors use some outdated Android devices and most of them can support more modern and faster certificates like an ECDSA 256bit.

So what i do is having both and let Nginx check which one to serve on each user. Using a new version of openssl also helps. Compiling by using a newer than the default GCC will help also.

There are a ton of little things that can save a few ms left and right but again for me the most important is to have the page rendering finish as soon as possible.

Of course having an expensive server it helps as you will probably get a cpu with some extra instructions set specific for ssl.

There is a lot more than getting the most expensive server...

I think we are a bit out of the main topic 🙂