[deanhills]

Last night I was able to upgrade Debian 8 to Debian 10 from the command line, using a tutorial of Mashiro's:
https://post4vps.com/Thread-Easy-way-to-...-Debian-10

To say the speed increased, is understating it.  Debian 8 performance was OK and normal fast.  But after the reboot, the performance of my VPS was like lightning in comparison. How come?

So what was there in Debian 10 that made the VPS perform so much faster (and possibly more efficiently) than Debian 8 did.  Again, Debian 8 was doing OK.  Not as fast as Debian 9 that I was using yesterday.  But Debian 10 for me is a total surprise.  A nice one!  

I should check out tutorials for Debian as well.  Like it's not that difficult to use Debian so far coming from CentOS as all I do is replace yum with apt-get, and Google has an answer for everything.  But would be nice for me to have a closer look with a quality how-to tutorial.

Over the last few days of tutorials I just come to experience first hand how many iffy and rubbish tutorials there are out there under really good and what seems to be reliable names. Like maybe the tutorials are mostly used as a cheap means of advertising the business. They are not that complete and can put one on the wrong path very easily, particularly if one takes them seriously.

[sagher]

@deanhills : i have so question.
Can i use this command on my VPS-9 ?
Is this a soft upgradation command which remains all my custom programs and settings, or convert into fresh copy with zero updates.
my basic Running programs are in
/etc
/var
/usr
is this all directories still remains or all OS back to fresh?

[tryp4vps]

Hello may I know when you say VPS speed increases, what kind of speed parameter(s) you are referring to?

I mean network speed, disk speed, or what else?

Normally a fresh install of an OS should run faster than the previous one you have used for some period of time.

[Mashiro]

Debian 8 uses the Linux Kernel 3.16.* (according to https://wiki.debian.org/DebianJessie). Debian 9 has Linux Kernel 4.9 LTS (according to https://wiki.debian.org/DebianStretch). Debian 10 is running Linux Kernel 4.19 (according to https://wiki.debian.org/DebianBuster).

With such kernel jumps there are a lot of improvements to await. Especially from 3.16 to 4.9 there were a lot of changes. With such kernel updates a lot of important things are being either fixed or improved. We're talking about things like improvements to the network stack, disk management, RAM management, file system and etc. Main goal is also to improve the performance of the system overall, fix security holes and other goals like introducing new kernel level features.

The list is huge:
- https://kernelnewbies.org/Linux_3.16
- https://kernelnewbies.org/Linux_4.9
- https://kernelnewbies.org/Linux_4.19


Another important point are the actual software packages that come with the Debian OS. They also get a lot of updates as with every new Debian version they usually skip a lot of versions (sometimes less and other times more) between them. Improvements to the software that is running can also have very positive effects. They applications might be way more optimized and run much better overall.


Might be that you are experiencing this. Maybe. The possibility would be you're in a lucky position / moment and the server and network has a lot more power freely available than usual.

If you can reproduce this performance gain with every Debian 8 to 10 upgrade process at anytime it is most likely that you profit from the huge updates that came with the new kernel version and software packages in the newer Debian versions.


Another possible option: it's related to the new kernel version but it's actually some kernel magic related to network stack, caching and such. That's too deep to dive in though. I cannot even explain it if I wanted right now. I've read about it somewhere before though.

[deanhills]

Thanks @Mashiro  I get it now.  Like with myVestaCP of dpeca's, his panel script must have been upgraded too - everything involved with Debian 10 and that has been written for Debian 10 must be on a higher level.

At any rate, it's been a powerful lesson to upgrade Debian 8 to 10 from the command line.  I've never done this before.  I never expected it to be as simple either.  Or with such a meaningful difference in speed and performance.

@deanhills : i have so question.
Can i use this command on my VPS-9 ?
Is this a soft upgradation command which remains all my custom programs and settings, or convert into fresh copy with zero updates.
my basic Running programs are in
/etc
/var
/usr
is this all directories still remains or all OS back to fresh?

@sagher I'd be careful if I were you as you have content on your VPS.  I did the upgrade without any content on my VPS.  Maybe we should ask @Mashiro's advice what the effect can be if you have content on your VPS.

Hello may I know when you say VPS speed increases, what kind of speed parameter(s) you are referring to?

I mean network speed, disk speed, or what else?

Normally a fresh install of an OS should run faster than the previous one you have used for some period of time.

I didn't do a technical analysis of the speed before I started the upgrade.  So I can't really compare stats other than my experience with loading VestaCP on Debian 8 and running a WordPress installation on it.  The speed was not slow when I downloaded themes and plugins and uploaded pages and posts.  But with the same installation on Debian 9 and Debian 10, the plugins uploaded noticeably faster.   In absence of real corroborating data, the speed could easily be due to different factors of course as Mashiro also suggested.

[Mashiro]

@deanhills : i have so question.
Can i use this command on my VPS-9 ?
Is this a soft upgradation command which remains all my custom programs and settings, or convert into fresh copy with zero updates.
my basic Running programs are in
/etc
/var
/usr
is this all directories still remains or all OS back to fresh?

@sagher I'd be careful if I were you as you have content on your VPS.  I did the upgrade without any content on my VPS.  Maybe we should ask @Mashiro's advice what the effect can be if you have content on your VPS.

To answer what I have already answered:

Warning: Debian might be generally a very stable OS but distribution upgrades are something that can be rated as dangerous and should be handled with care. I don't recommend to dist upgrade a long running server with applications and other things already running on it. Even if you decide to do so take backups! Things could go wrong and the server could be rendered unusable! Therefore I would like to explain that I'm not responsible for any damage that might happen to any server when performing a distribution upgrade. I recommend only to do this on a fresh installed Debian 9 VPS where you have nothing to lose!

(https://post4vps.com/Thread-Easy-way-to-...8#pid32428)

@hamed @sagher

Have you both ignored the red text that is placed at almost the top of the tutorial post?

It answers your questions pretty much in terms of if you should do this or not and when you should do it or not.

About the question by @hamed: During the upgrade the packages are upgraded and the kernel. Of course partially configurations are replaced with newer versions that come with the new software. However for everything you changed or custom configured you will be asked mostly if you wish to replace the configuration or keep the original. And for the most it will NOT replace any configuration during an upgrade. So nothing is mixed up. You are still just upgrading the packages to the level the newer Debian version.

(https://post4vps.com/Thread-Easy-way-to-...3#pid40753)


Nothing changed since I posted these statements the last time. It's all still the same.

Everything and nothing can happen... you do it at your own risk. My clear recommendation is DON'T DO IT on a production server. Especially with your level of knowledge and experience it is highly NOT recommended to perform this at all.

[deanhills]

@Mashiro  Thank you so much for the awesome tutorial about how to set up Debian and Nginx for publishing a WP site.

So tonight I went through the upgrade process again.  And had success for a second time from a clean install of Debian 8.  In anticipation of trying out your tutorial, I wanted to have a VPS with up to date Debian 10 on it, and I think this is the closest I'll be able to get to that for now.

In the meanwhile, while I had to reinstall the OS anyway, I took speed tests and benchmark tests of VPS 9 Dallas with the two different OS Installations CentOS 7 and Debian 8, as well as Debian 10 upgraded version after reboot.  I know there can still be variances, but what has me curious is that the speed tests with the different OS installations are taken from varying locations and not the same one every time.  Here are the results of my tests:


(1) Centos 7: Speed Test



(2) Debian 8: Speed Test



(3) Debian 10: Speed Test



---------------------------------------------------------------------------

(1) CentOS 7: Benchmark Test
Benchmark started on Wed May 19 15:28:49 EDT 2021
Full benchmark log: /root/bench.log

System Info
-----------
Processor       : Intel® Xeon® CPU E5-2670 v2 @ 2.50GHz
CPU Cores       : 2 @ 2499.998 MHz
Memory          : 7821 MiB
Swap            : 31 MiB
Uptime          : 5 min,

OS              : CentOS Linux 7 (Core)
Arch            : x86_64 (64 Bit)
Kernel          : 3.10.0-1160.25.1.el7.x86_64
Hostname        : kvm-Post4VPS-Dallas


Speedtest (IPv4 only)
---------------------
Your public IPv4 is 23.**.**.**

Location                Provider        Speed
--------                --------        -----
CDN                     Cachefly        205MB/s

Atlanta, GA, US         @sohamb03       79.5MB/s
Dallas, TX, US          Softlayer       101MB/s
San Jose, CA, US        Softlayer       29.9MB/s
Washington, DC, US      Leaseweb        58.0MB/s

Sao Paulo, Brazil       Softlayer       2.33MB/s

Singapore               Softlayer       7.28MB/s
Taiwan                  Hinet           13.9MB/s
Tokyo, Japan            Linode          13.9MB/s

Nuremberg, Germany      Hetzner         13.9MB/s
Rotterdam, Netherlands  id3.net         61.2MB/s
Haarlem, Netherlands    Leaseweb        160MB/s
Milan, Italy            Softlayer       6.24MB/s

Sydney, AU              Future Hosting  350KB/s


Buffered Sequential Write Speed
-------------------------------
I/O (1st run)   : 343 MB/s
I/O (2nd run)   : 575 MB/s
I/O (3rd run)   : 731 MB/s
Average I/O     : 549.667 MB/s

---------------------------------------------------------


(2) Debian 8: Benchmark Test
Benchmark started on Wed May 19 15:50:28 EDT 2021
Full benchmark log: /root/bench.log

System Info
-----------
Processor       : Intel® Xeon® CPU E5-2670 v2 @ 2.50GHz
CPU Cores       : 2 @ 2499.998 MHz
Memory          : 8003 MiB
Swap            : buffers/cache: MiB
Uptime          : 11 min,

OS              : Debian GNU/Linux 8 (jessie)
Arch            : x86_64 (64 Bit)
Kernel          : 3.16.0-4-amd64
Hostname        : kvm-Post4VPS-Dallas


Speedtest (IPv4 only)
---------------------
Your public IPv4 is 23.**.**.**

Location                Provider        Speed
--------                --------        -----
CDN                     Cachefly        267MB/s

Atlanta, GA, US         @sohamb03       76.6MB/s
Dallas, TX, US          Softlayer       105MB/s
San Jose, CA, US        Softlayer       37.4MB/s
Washington, DC, US      Leaseweb        56.1MB/s

Sao Paulo, Brazil       Softlayer       2.20MB/s

Singapore               Softlayer       8.31MB/s
Taiwan                  Hinet           326KB/s
Tokyo, Japan            Linode          13.3MB/s

Nuremberg, Germany      Hetzner         14.8MB/s
Rotterdam, Netherlands  id3.net         58.1MB/s
Haarlem, Netherlands    Leaseweb        178MB/s
Milan, Italy            Softlayer       6.87MB/s

Sydney, AU              Future Hosting  351KB/s


Buffered Sequential Write Speed
-------------------------------
I/O (1st run)   : 307 MB/s
I/O (2nd run)   : 651 MB/s
I/O (3rd run)   : 751 MB/s
Average I/O     : 569.667 MB/s

-------------------------------------------------------------------

(3) Debian 10 Benchmark Test
Benchmark started on Wed 19 May 2021 04:49:22 PM EDT
Full benchmark log: /root/bench.log

System Info
-----------
Processor       : Intel® Xeon® CPU E5-2670 v2 @ 2.50GHz
CPU Cores       : 2 @ 2499.998 MHz
Memory          : 7978 MiB
Swap            : 31 MiB
Uptime          : 6 min,

OS              : Debian GNU/Linux 10 (buster)
Arch            : x86_64 (64 Bit)
Kernel          : 4.19.0-16-amd64
Hostname        : kvm-Post4VPS-Dallas


Speedtest (IPv4 only)
---------------------
Your public IPv4 is 23.**.**.**

Location                Provider        Speed
--------                --------        -----
CDN                     Cachefly        94.0MB/s

Atlanta, GA, US         @sohamb03       70.9MB/s
Dallas, TX, US          Softlayer       88.1MB/s
San Jose, CA, US        Softlayer       35.7MB/s
Washington, DC, US      Leaseweb        52.4MB/s

Sao Paulo, Brazil       Softlayer       6.56MB/s

Singapore               Softlayer       6.71MB/s
Taiwan                  Hinet           326KB/s
Tokyo, Japan            Linode          14.2MB/s

Nuremberg, Germany      Hetzner         14.5MB/s
Rotterdam, Netherlands  id3.net         62.8MB/s
Haarlem, Netherlands    Leaseweb        126MB/s
Milan, Italy            Softlayer       6.16MB/s

Sydney, AU              Future Hosting  350KB/s


Buffered Sequential Write Speed
-------------------------------
I/O (1st run)   : 344 MB/s
I/O (2nd run)   : 599 MB/s
I/O (3rd run)   : 817 MB/s
Average I/O     : 586.667 MB/s


OK.  I'm now ready to start with @Mashiro's tutorial for installing a WP site from Debian and Nginx.

[fChk]

Last night I was able to upgrade Debian 8 to Debian 10 from the command line, using a tutorial of Mashiro's:
https://post4vps.com/Thread-Easy-way-to-...-Debian-10

To say the speed increased, is understating it.  Debian 8 performance was OK and normal fast.  But after the reboot, the performance of my VPS was like lightning in comparison.  How come?

There isn't much input here to objectively reason with!.. But yes, if everything was equal application-wise, there is the Kernel upgrade that takes place from a major distribution upgrade to another [1].

Based on @Mashiro's input in Post#4, I would say that the jump from a 3.16.x (in Debian 8) to a 4.19.x kernel (in Debian 10) should not go unnoticed while everything else being equal.

I should check out tutorials for Debian as well.  Like it's not that difficult to use Debian so far coming from CentOS as all I do is replace yum with apt-get, and Google has an answer for everything.  But would be nice for me to have a closer look with a quality how-to tutorial.

@deanhills
You should know that except for the package manager and the commands that go with it (plus few advanced assumptions) you're still using the BASH SHELL, thus at the CLI-level everything  should still work the same... That's the beauty of GNU/LINUX systems!

Over the last few days of tutorials I just come to experience first hand how many iffy and rubbish tutorials there are out there under really good and what seems to be reliable names.  Like maybe the tutorials are mostly used as a cheap means of advertising the business. They are not that complete and can put one on the wrong path very easily, particularly if one takes them seriously.

Agreed!.. 75% of those tutorials are just copy/paste stuff anyway. For the other 25%, you should be aware of their assumptions before using them, which generally require some advanced knowledge or at least an advanced familiarity with the system you're working on.

Hence why blindly following a tutorial rarely works!

------------------
[1]- Well I just saw the benchmarks in the last post of yours, I may comment on that next week.

[deanhills]

There isn't much input here to objectively reason with!.. But yes, if everything was equal application-wise, there is the Kernel upgrade that takes place from a major distribution upgrade to another [1].

Based on @Mashiro's input in Post#4, I would say that the jump from a 3.16.x (in Debian 8) to a 4.19.x kernel (in Debian 10) should not go unnoticed while everything else being equal.

Thanks for the feedback.

Hence why blindly following a tutorial rarely works!

Unless it's by a reliable source like Mashiro.    I'm almost convinced that he puts himself completely in the shoes of who is going to work through the tutorial.  And when he tests his tutorial (which he always does as they work each and every time) he also perfects all of the steps with the user of the tutorial in mind.

But yes, I've worked through MANY tutorials on the Web for step by step LAMP + WP from places like rosehosting.com and equivalent, and I think those are tutorials that have been written as "bait" to get traffic and to subscribe to their managed accounts services.  Like more often than not the tutorials are part of SEO.  So inevitably they miss out on finer details like step by step for creating directory trees, config files etc. I'm glad that I stumbled through them, even though it was a great waste of time, as I then got to appreciate the gaps in their tutorials when I worked through Mashiro's tutorials in the end. It was a good learning curve overall.

------------------

[1]- Well I just saw the benchmarks in the last post of yours, I may comment on that next week.

Thanks.  

[Sn1F3rt]

But yes, I've worked through MANY tutorials on the Web for step by step LAMP + WP from places like rosehosting.com and equivalent, and I think those are tutorials that have been written as "bait" to get traffic and to subscribe to their managed accounts services.  Like more often than not the tutorials are part of SEO.  So inevitably they miss out on finer details like step by step for creating directory trees, config files etc.  I'm glad that I stumbled through them, even though it was a great waste of time, as I then got to appreciate the gaps in their tutorials when I worked through Mashiro's tutorials in the end.  It was a good learning curve overall.

Well, not all tutorials are baited just for traffic. It's just that you should take a look at the command and analyze if you should actually execute it, like @fChk said that is not to follow a tutorial blindly. DigitalOcean has some awesome tutorials, and I find confidence in pretty much every tutorial posted on their website. 

However, I keep a track of what I'm executing. Often in DigitalOcean and Vultr tutorials, there are some extra performance-related steps, for example, DO has a tutorial on SWAP configuration for CentOS and towards the end, there are a few extra commands on swapiness and stuff that are optional. What I'm trying to say is that we shouldn't end up over executing commands than the required bit.