HOSTNAME fixes ROX slowness problem - SOLVED

[tazoc]

I am appealing to TazOC for some insights that could be applied in the 32bit community (or ALL PUPs, in general).

The only change to Personalize Settings in Lighthouse 64-511 Update 1 (related to hostname) is the same one that Shinobar has incorporated at line 1199 of /usr/sbin/countrywizard.qs in firstrun-1.8.4:

Code: Select all

   hostname $HOSTNAME
   echo -n $HOSTNAME > /etc/hostname
   echo "127.0.0.1 localhost $HOSTNAME" > /tmp/hosts
   grep -vw 'localhost'  /etc/hosts >> /tmp/hosts
   [ -s /tmp/hosts ] && mv -f /tmp/hosts /etc/hosts

I appreciate Shinobar's fine work on this, as well as his and others efforts to improve NLS support. 01Micko's method to ensure a unique hostname seems like a good idea as well. Combined with Personalize Settings, the user is free, if so desired, to change the numerical suffix (or the entire hostname) to something easy to remember.

I also changed a line in /usr/sbin/ipinfo from

Code: Select all

	# tab 1 - interfaces
	var01=`echo Hostname: "$HOSTNAME"`

to

Code: Select all

	# tab 1 - interfaces
	var01=`echo Hostname: "$(hostname)"`

so that it reflects any change to hostname without having to reboot.
-TazOC

[greengeek]

gcmartin - well done in bringing clarity to this.

If I may defend rcrsn51's comments - maybe he feels as I do, that Puppy has allowed "free-thinkers" to compute the way they want to, without adhering to external "standards".

Windows has its own approach to "standards" - it is a quicksand.
Mac imposes it's own standards - it opposes individuality.
Puppy sets me free - I choose what I want to do. (Thanks Barry!)

I understand why the Hostname issue must be adressed, but I also understand why some feel that they do not want to see Puppy become subservient to other peoples rules.

You have said "I have never subscribed to the "every man (application) for himself" in systems development"

I have a different view.

Independence is what made puppy useful for me.

Please do not shackle Puppy to those who invented TCP/IP and other network "standards". Such people work for defence forces and governments.

They understand centralised control, and subjugation of the individual.

Ok, lets improve any faults within Puppy, but without having to force Puppy users to conform.

[gcmartin]

Thanks @greengeek for your comments.

I agree with you about the freedom to create without shackles. This thread is NOT about shackling anything.

Its about your Puppy Linux system operation and its about your LAN operation.

As I remember, @Rcrsn51 (whom I regard highly for the many many contributions he makes. Whenever he says something I listen. ) did NOT come out with concerns about Puppy freedom, his concern was about whether there is harm in leaving the problem built into Puppy. Further scrutiny shows that there is harm when hostname fields do not agree. This thread's contributions have resolved that concern for all of us.

Linux is a system. Its not a single program. Its applications and services, that run, expect a certain set of standards for it to integrate and to operation politely.

This thread's discussion and its solution is about just that. Not being shackled as you infer, but about providing a standard set of services to allow your PC to operate politely; internally and externally. This gives us the freedom to do all kinds of things while the underlying system insures proper operations for us.

Further as is noted by Shinobar's and by Tazoc's efforts, it provides you, the user, the ability to freely select in such a way to maintain stability and polite executions internally and externally.

Please understand this this is and was NOT a personal crusade. It was about bringing awareness to resolving a problem.

You now will be able to have your system experience the benefits provided by the contributions you find from this thread. I do NOT take credit for this solution. The contributors are my heroes for their recognitions and addresses they provide us.

Timeline:

1. 01Micko recognize this anomaly several years ago and started work but never revisited it
2. Shinobar added hostname to FirstRUN, last year to make it easy and conscious
3. TaZOC was the first implementation that successfully managed all internal Puppy fields updating via Personalized Settings wihtout desktop shutdown
4. 01Micko revisited his work and gave us the OTB boot-time hostname update to handle the system fields so that when the Live media boots to desktop, all fields would agree and LAN behavior would not conflict
5. Barry embraced 01Micko's work in WOOF because it makes sense to address this.

Thus, now every distro maker can benefit from everyone's contributions here. And further, every user of Puppy distros will benefit from everyone's contribution. Not one of these contributors are doing anything except addressing proper system behavior.

Also, because of Puppy's design, if you are not happy with this work by the contributors, anyone may, either use Shinobar's FirstRUN and reset all of your Puppy PC names to the same (in violation of what every instructor and LAN administrator on the planet teaches) or you may modify your Puppy to pull out the harmless additions everyone has contributed for you.

So, I hope you understand and see that this thread's contribution is NOT about conformity. But, it is about correcting an overlooked "fault".

The intent here is to make a great distro. That means all the necessary tests to be called a great OS is in its behavior on the desktop, within the LAN and within the WAN.

This thread is a very very small step to go along with ALL of the small steps as Puppy matures. And it is maturing very, very rapidly! I have written before about some of the great contributions made over the past 12-18 months in this community. These have been game changers that are part of the maturation we are witnessing. It's making newbie use as well as season veteran use, a breeze for the well thought thru solutions provided for all of us. Heck we can now run 32bit Puppy flavors on PCs utilizing up to 64GB of RAM, today, on a LiveCD. This was unheard of a year ago. Puppy is maturing in ways and at rates that other mainstream distros cannot.

Question for anyone
Please point out how the hostname resolutions instituted, have a negative impact on a user's creativity? Has this thread given the appearance that creativity is in peril?

Acknowledgements
I don't think I have done this until now, but I want to thank everyone who has responded to this thread. Collectively, you have made this problem understood and resolved. Thanks EVERYONE!

Hope his helps

[gcmartin]

There are 2 elements to this problem. First, what happens when ANY PUP is booted for first-tiime from LiveCD or Netboot (pfix=ram). Next, a tool to assist changing the hostname after the system reaches desktop.

Solved by Barry's incorporation into WOOF for base ISOs built. The base allows all PUP LIveCD/ISO to boot with a unique, out-of-the-box (OTB) LAN hostname with all internal Puppy subsystems files consistent with that host's name.

The following distros ALSO employs Shinobar's Personlaize Seettings (FirstRUN) in such a way that when anyone changes the hostname field in Personalize Settings, all proper Puppy files and fields are correspondingly updated as wel without ANY user needing to "restart the desktop" or "logout at Ctrl-Alt-F1 prompt" or "save-session+reboot"l:

• TaZOC has solved this in LightHouse64
• 01Micko has solved this in SLACKO

Hope this helps all distro builders.

[amigo]

BTW, I don't think anyone ever explained why the hostname would make rox slow. It's because rox can transparently handle NFS-mounts (and maybe other network mounts) -FWIW.