- This topic has 658 replies, 44 voices, and was last updated Apr 3-3:29 pm by Brian Masinick.
-
Posts
-
Brian Masinick
As far as init variations, our runit is better than ever because the runit services are finally being implemented through commands entirely managed by runit. Even Version 23.2 had a custom implementation that I believe that @Xecure helped develop, using sysVinit services operated through the runit service manager; now unless I am mistaken all services are 100% runit service manager procedures and you can see the different feedback when you run them. It’s matured very well.
I like some stuff with dinit but we still have a bit more work to do to get it right. For example, date/time clock management does not seem to be 100% implemented, or if it is, it’s not working properly; I’ve had to use our manual tools many times to correct the time on my system when I move back and forth between systems that handle time management differently. I’m being very picky about all of this because I’m trying to help test and get the things I am familiar with perfect. I sincerely hope that others are doing the same with applications they use and are familiar with. Collectively that’s how we as a community get this stuff working with excellence!
Keep up the great work! Let’s help move things along so when our excellent leader @anticapitalista returns he knows what needs updating and what’s good to go!
As for our other contributors, whether it’s development, testing, user interface improvements, or visual design, each element is guided by those with interests and contributions; it all matters and I appreciate everyone’s role!
Even those of us who don’t know much about computers can help by letting us know which things are understandable to use and which are difficult or confusing. It IS possible to keep things efficient, yet design and implement them well so that they work the way the community benefits from them, so all of us, even novices, can play a role to allow us to retain efficiency while honing and perfecting what we know how to do best.
--
Brian Masinick
Alternate "Search B":
This search pageandfree
Unfortunately, this issue that had not affected my antiX-26-rc1 systems till now, has just appeared on Vivobook (my previous posts, earlier today, were about Vaio, not about Vivobook). I can’t boot the legacy (5.10) kernel with persistence on my installed system anymore.
I reinstalled the legacy kernel and now it’s OK.
It seems that the same issue affected also the modern (6.6) kernel, but only on 66 (while with legacy kernel it affects all init programs).
I also reinstalled 66:
sudo apt purge 66* oblibs
and
sudo apt install 66 66-service-boot 66-service-cron 66-service-rc-local-user 66-tools s6-66-init-antix 66-service-acpid 66-service-boot-user 66-service-dbus 66-service-scandir 66-service-alsa 66-service-connmand 66-service-dhcpcd 66-service-seatd s6-66-antix-live-user
but, unfortunately, this issue remained.So, on Vivobook now, both kernels work properly with all init programs, except for 66.
I turned my two frugal (static) installations (one for each kernel) to 66 (boot parameter “init=s6-66”) and both worked properly for a while, but after some reboots, it crashed for the modern kernel and I turned that frugal installation back to runit.antiX utilities for processing PDF files (posted by ChPol)
Brian Masinick
@ProwlerGr, do you have any insights or suggestions on how to establish the correct intended timezone details for dinit (similar to your suggestions for 66)? I’ve been unable to locate the dinit tool for establishing a time service.
I like our graphical /usr/local/bin/set_time-and_date.sh tool for being able to manually alter the time, date, and timezone, but this is primarily for moving around. I’m looking for an accurate, reproducible, hopefully service oriented method to set and maintain the correct timezone parameters for dinit. As you may recall, I prefer setting
TZ=America/New_York HARDWARECLOCK=localtime--
Brian Masinick
Alternate "Search B":
This search pageBrian Masinick
@andfree based on some testing I did a while ago with an old Lenovo Thinkpad X201 (which by the way, despite it’s age, is an excellent system) it did show some shortcomings. In particular, programs that utilize a lot of “Rust” code – that’s a relatively new systems programming language, and even the Linux kernel is starting to have some Rust code in addition to the mountains of classic C code – this tends to cause PROBLEMS on old systems like this. As good as Lenovo laptops have been, the old generation does not work very well with this new Rust code at all. While our systems don’t have a lot of the very newest cutting edge stuff, based on what I’ve seen with this wonderful old system, it DOES now have its limitations and it runs best on old software. For example, Version 4.4 Linux kernels run GREAT on this system. While I CAN run some V6 stuff, the older the kernel is, the less likely it is to encounter Rust system code or application code. On newer systems that can handle it, Rust has good security compared to old stuff; on the other hand it consumes resources and generates a lot of heat, something an old box like the Thinkpad X201 does not take to well at all, so my guess is that these older models you have are equally affected by anything resembling “new” software, so keep that in mind.
The Vivobook may tolerate some of it, the Vaio (I think that was the other one in question) probably struggles a lot with this newer stuff; my guess is that one or both of these boxes are better off with 23.2 unless you can manage to configure a super lean setup with the oldest supported kernel you can find.
--
Brian Masinick
Alternate "Search B":
This search pageBrian Masinick
@ProwlerGr, do you have any insights or suggestions on how to establish the correct intended timezone details for dinit (similar to your suggestions for 66)? I’ve been unable to locate the dinit tool for establishing a time service.
I like our graphical /usr/local/bin/set_time-and_date.sh tool for being able to manually alter the time, date, and timezone, but this is primarily for moving around. I’m looking for an accurate, reproducible, hopefully service oriented method to set and maintain the correct timezone parameters for dinit. As you may recall, I prefer setting
TZ=America/New_York HARDWARECLOCK=localtimeBy the way, it’s clear to me why dinit time is not working the way I’d like: it sets the time according to UTC and except for the manual methods, there’s no way to alter it, but if I just run the hwclock program and specify –utc when I’m running dinit, it works; of course the “other” clocks are off, so I have to go the other way for my other setups, which is why I’d like a solution to resolve this other than /usr/local/bin/set_time-and_date.sh, which is too interactive for the automatic operation I prefer.
I do have a couple of ways to resolve this myself; I’ve indicated a couple of them, but chances are I won’t be the only one encountering this; if someone sets up a local time configuration they will get the same behavior. At least I know what’s there and what’s missing. If I could find the resource that is setting up the UTC configuration I could change that — and if CHANGING that could be added to the dinit setup that would be awesome and ideal.
--
Brian Masinick
Alternate "Search B":
This search page(…) even the Linux kernel is starting to have some Rust code in addition to the mountains of classic C code – this tends to cause PROBLEMS on old systems (…) On newer systems that can handle it, Rust has good security compared to old stuff; on the other hand it consumes resources and generates a lot of heat (…) so my guess is that these older models you have are equally affected by anything resembling “new” software (…) are better off with 23.2 unless you can manage to configure a super lean setup with the oldest supported kernel you can find.
Thanks for all this info and your advice.
The configuration of the leanest possible system that covers all my needs with minimal resource consumption and heat generation is my constant pursuit for all the models I use, more or less old.
For years I tried to use older kernels for this purpose, living under the constant stress of impending EOL. This changed only after I used kernel 5.10, which was long supported and worked well with all those boxes. Later, I did a comparison with kernel 6.6, but I’ve never come to a clear conclusion.
And all this story was about a period before the two main 64 bit machines I use for testing antiX-25/26 (Vivobook and Vaio) came into my hands.
For the moment, I am satisfied with rc1. I’m facing the mentioned issues with 66, but 66 is not my preferred init program; this is runit (and as an alternative for further testing, I’m thinking of dinit).
However, you make me think of doing a memory consumption comparison between 23.2 and 26, when the latter is released.antiX utilities for processing PDF files (posted by ChPol)
Brian Masinick
However, you make me think of doing a memory consumption comparison between 23.2 and 26, when the latter is released.
Take a look at my “What are you “here” with today?” topic to get a pretty comprehensive look at memory usage on my hardware.
It clearly won’t show your hardware, but by looking at the last few pages you can get a pretty good look at the different init alternatives in antiX 26 RC1. There may be a few instances of 23.2 within the last 3-4 pages but if you look at https://www.antixforum.com/forums/topic/what-are-you-here-with-today/page/385/#post-196461 you can see that antiX 23.2 Full Runit consumes just over 220 MB at startup, anywhere from 30 to 60 MB LESS than the current version.However, if you DISABLE a bunch of running services, reduce the number of TTYs, remove services you don’t and will never use and you MAY be able to get pretty close to what I customized for antiX 23.2. When antiX is finalized I’ll probably do the same for it too.
--
Brian Masinick
Alternate "Search B":
This search pageBrian Masinick
By the way @christophe has put together some really nice scripts to create the perfect (for him) version of antiX Core.
IF you have a 100% older setup where you can START with a V5 kernel and Ethernet connections maybe building an antiX Core, adding just what you want (and maybe starting with one of his “recipes”) will be really cool. IF you ONLY have WiFi, that’s kind of tough, but I do recall that we’re offering WiFi on some of the recent antiX Core ISO images; something to think about if you want a super custom setup.One thing you may appreciate though, once you try that is that the FULL versions of antiX really do offer a lot of useful conveniences and APPEARANCE items; unless you have the perfect theme that you add in and a specific set of tools to make the system just right for you, maybe it’s NOT the thing to do; I just thought that if you’re ready to try something DIFFERENT, that might be an exercise to consider some day. I try it every 3-4 years just to keep my mind well engaged and functioning! 🙂
NOTE: For ‘Joe Average’ antiX Core is not appropriate, so if you consider yourself ‘Joe Average’ that idea isn’t worth much consideration. This old geek likes to build em up from the ground every so often and that’s the value of the Net and Core builds.
--
Brian Masinick
Alternate "Search B":
This search pageif you look at https://www.antixforum.com/forums/topic/what-are-you-here-with-today/page/385/#post-196461 you can see that antiX 23.2 Full Runit consumes just over 220 MB at startup, anywhere from 30 to 60 MB LESS than the current version.
Thank you.
The dinit service manager shows the available tty devices(…)
The tool seems to ignore any attempt to remove unused, unwanted tty devices
How do you determine which of these tty devices are unused/unwanted? I would also remove them, but I’m not sure how many of them are necessary.
- This reply was modified 6 months, 3 weeks ago by andfree.
antiX utilities for processing PDF files (posted by ChPol)
By the way @christophe has put together some really nice scripts to create the perfect (for him) version of antiX Core.
IF you have a 100% older setup where you can START with a V5 kernel and Ethernet connections maybe building an antiX Core, adding just what you want (and maybe starting with one of his “recipes”) will be really cool.Yes, your previous relevant suggestions are still in the back of my mind.
IF you ONLY have WiFi, that’s kind of tough, but I do recall that we’re offering WiFi on some of the recent antiX Core ISO images
I think that offering WiFi was what mainly differentiated Core from Net ISO. I also think that there’s no definitive answer yet to the question about whether wifi support should be included in antiX-26-core. In any case, I have access to Ethernet connections.
something to think about if you want a super custom setup.
I really want a super custom setup, but as I have written:
building from core is the hardest way (most complicated configurations).
For this reason,
that might be an exercise to consider some day.
Some day after antiX-26 will have been released.
- This reply was modified 6 months, 3 weeks ago by andfree.
antiX utilities for processing PDF files (posted by ChPol)
Brian Masinick
@andfree as I interpret what @anticapitalista wrote concerning antiX Core, WiFi support WILL be included for 64-bit systems and V6 kernels:
anticapitalista wrote:
I just built antiX-26-core 64-bit with wifi support and the final iso comes to 665MB.--
Brian Masinick
Alternate "Search B":
This search pageBrian Masinick
I *think* it’s the 32-bit systems where WiFi support is omitted. Not sure about 64-bit support for V5 kernel, but confirmed that 64-bit support for V6 kernel has already been tested and built (just not released yet).
--
Brian Masinick
Alternate "Search B":
This search pageolsztyn
Just a question:
This recent update (antiX’ modification) of Connman from 1.44-3.0antix3 to 1.44-3.0antix4 causes Network Manager to not work im my experience. Although Network Manager service is up, connection to Internet fails due to DNS, as it appears to me.
Shutting down Network Manager service and starting Connman service brings back Internet connection.
Not a big problem, as I do keep both installed for contingency, as the above malfunction demonstrates it is useful to have, but I am curious what was this antiX modification done to Connman that would introduce this malfunction…- This reply was modified 6 months, 3 weeks ago by olsztyn.
Live antiX Boot Options (Previously posted by Xecure):
http://antixlinuxfan.miraheze.org/wiki/Table_of_antiX_Boot_ParametersProwlerGr
Just a question:
This recent update (antiX’ modification) of Connman from 1.44-3.0antix3 to 1.44-3.0antix4 causes Network Manager to not work im my experience. Although Network Manager service is up, connection to Internet fails due to DNS, as it appears to me.
Shutting down Network Manager service and starting Connman service brings back Internet connection.
Not a big problem, as I do keep both installed for contingency, as the above malfunction demonstrates it is useful to have, but I am curious what was this antiX modification done to Connman that would introduce this malfunction…I don’t get this problem in a test VM (I have both connman & network-manager installed but only network-manager enabled on all inits).
DNS works as expected after the upgrade.I don’t get this problem in a test VM (I have both connman & network-manager installed but only network-manager enabled on all inits).
DNS works as expected after the upgrade.I do not know if this happens in a VM. I have tested this on two laptops and my above report reflects my experience. I tested this behavior being on runit but Network Manager was enabled on other inits as well.
An additional very strange thing I observed was the an ttempt to purge Connman resulted in the same – Internet connection on Network Manager became broken, if this helps to identify the issue…
As I mentioned, because of this issue – caused by either upgrade of Connman or an attempt to purge it, I disabled Network Manager and enabled Connman on my antiX 26RC1 image so it is my ‘official’ working image now.
Other than that, no particular complaints, except on Dinit logging to lightdm session still does not work for me in spite of the latest fix you suggested, change on line 29 of lightdm.conf, which worked for you. I think I identified recently that lightdm service could not be started on dinit and this may have caused this problem.I am working on real hardware in my tests, multiple laptops, so not sure if this makes a difference from a VM. A similar issue I have observed with one of my machines and MX 25 ahs, where it stops before lightdm session because lightdm service could not be started, reported on MX forum. The same image works fine on other laptops. But this is MX 25 with systemd, so just a similarity of behavior.
I am hoping that eventually these issues will find their resolution, particularly Connman behavior seems to instill some uncertainty of why and therefore more concerning.
Thanks and regards – thanks again for your tremendous help in my exploring hardware support expansion recently…Live antiX Boot Options (Previously posted by Xecure):
http://antixlinuxfan.miraheze.org/wiki/Table_of_antiX_Boot_Parameters
- You must be logged in to reply to this topic.