| DistroWatch Weekly
1 • swap (by brad on 2016-02-08 00:23:47 GMT from North America) |
I use a partition, but honestly, I can count on the fingers of one hand the times that swap was actually accessed, other than hibernation. I don't even use hibernation much any more.
2 • swap (by Gustavo on 2016-02-08 00:28:35 GMT from South America)
I need a swap partition having only 2GB of RAM, but I try not to use it very much. Linux has a very bad behavior in low free memory situations.
3 • swap (by Ferris on 2016-02-08 00:30:31 GMT from Oceania)
On my main system with 16GB RAM I don’t use swap at all. Small VMs running with SSD backing, I do tend to include a swap partition.
4 • Swap (by 67GTA on 2016-02-08 00:39:02 GMT from North America)
I always create swap partition out of habit, but never use it. Does the kernel still require it for sleep/hibernation?
5 • swap as a zfs zvol (by Jonathan on 2016-02-08 02:22:28 GMT from North America)
I have a swap but it is running as a zvol in my ZFS pool (ZFS on Linux - Native ZFS implementation).
6 • swap, using x86 (by bluphoenyx on 2016-02-08 03:05:34 GMT from North America)
With large memory pools there is little need for swapping memory unless the system is already pushing its limits and the user(s) opens a graphic intensive application. Smaller environments, say 2 to 4GB, may also eliminate the swap file, provided the number of programs that gets run are limited. In a normal environment, one with a swapping system, programs that make use of swap should easily survive low memory conditions. Even when it isn't used, swapped ram is helpful. It provides a buffer for low ram conditions. Isn't that worth a few gigs on that massive HD?
Of course, solid state systems complicate things. A swap file in this situation could require an additional drive to avoid intensive access of the SSD. One alternative is a small swap drive on a ram disk. This might be practical on systems with double digit gigabytes of memory. Keep in mind that most programs cannot use all of the ram in the system because the OS won't allow them to. The use of swap may allow programs to edit larger documents or buffer data.
Glad to hear that Ubuntu is keeping 32 bits for now. I like tinkering with really old 586 & 686 systems. Options are already limited because many distro's use desktops that require modern graphics hardware. Once the more universal operating systems drop support, x86 distro's will be only a niche system. X86 distro's are also a great tool for emulation and virtual machines. One reason, they consume fewer resources when running.
7 • Swap (by Patrick on 2016-02-08 03:23:01 GMT from Oceania)
I still create a swap partition, but its more force-of-habit than anything. It's accessed very rarely as I have 8GB of RAM in my primary system.
8 • Ubuntu (by Solar Bay on 2016-02-08 03:24:27 GMT from North America)
Meet The World’s First Ubuntu Tablet (That Is Also a Desktop PC)
Canonical has officially unveiled the world’s FIRST Ubuntu tablet, the Aquaris M10 Ubuntu Edition.
Made by Spanish OEM Bq, the 10-inch tablet runs the same core Ubuntu operating system as the Ubuntu Phone but the user experience dynamically adapts based on input and screen size.
This means it’s able to work both as a tablet and provide a full Ubuntu desktop experience — the first such device to do so
9 • To swap or not to swap...is that the question? (by Tom Joad on 2016-02-08 03:56:05 GMT from Europe)
I swap out of habit. When I install an OS the system just creates the swap and I let it be. I have 5 gig on my tower MB. I just looked to see what is used and it is zero. Most times I look the value is zero.
Maybe, since the question came up and since I always have one, I will wipe the swap partition. I guess the system will just run along away. I expect that will happen. But being a former and long ago Windows user the swap file thing is just a hang over from those days.
I swap...so therefore I am? I guess.
10 • Swap (by slick on 2016-02-08 04:53:00 GMT from North America)
Yes I do format my drive with a swap partition. There has been some instances where the swap was being used, however not frequently.
As long as it is still occasionally used then I will prefer having the partition with swap. I have 16Gb of ram, and 35Gb of swap partition on a 1Tb drive.
It is one of those things that hasn't been broken yet, and does work when the need arises, so I will continue with a swap partition.
11 • Swap (by ReallyBubba on 2016-02-08 05:13:13 GMT from North America)
I ONLY use swap for laptops, for hibernation. I do not use swap on desktops. IMHO if you have at least 4GB RAM on desktop, swap really isn't needed. Giddy up. :)
12 • SWAP & 32 bit (by M.Z. on 2016-02-08 05:20:40 GMT from North America)
As with several others I always have a SWAP partition, though on my two main machines with 6 + GB of RAM I doubt it is ever used or is ever likely to be used. On the other hand my old 32 bit machine with only a couple of GB or so of RAM does SWAP some, and while I don't use the machine for much I like to have it around and able to handle tasks that need might need more memory.
Speaking of 32 bit machines, I for don't care much about what happens to the main edition of Ubuntu in terms of 32 bit support, but I want to be able to have Mint in 32 bit. This is especially true of the 32 bit version of Mint MATE & XFCE, both of which I have used on my old 32 bit PC. I currently have LMDE 2 & Mageia w/ XFCE on it. The post does seem to indicate 32 bit support will continue for others versions, so as long as downstream projects aren't affected I do think it makes since to ship some of the heavier desktops as 64 bit only. I have used KDE 4 on 32 bit hardware, but honestly that machine was generally better served by XFCE or something else a bit lighter, though it wasn't too bad with KDE. I'm not sure but I get the impression that Unity may be even more resource intensive than KDE, so it really does make sense to prioritize 64 bit versions for such big DEs & ensure that lighter DEs have a 32 bit version.
13 • App Count Not Head Count Matters Most (by Arch Watcher 402563 on 2016-02-08 06:02:26 GMT from North America)
Sure, BSD mentioned several inits, but only nosh is viable. Interestingly too, nosh works on Linux. The launchd talk has gone on forever at BSD.
The one thing I do believe about init stuff is that it should be factored out. There's no reason to keep rc scripts or systemd service files inside packages. They merit their own. Factoring them means easier init changes, too.
The more interesting news was "Linux Kernel as a Library Added to the Ports Collection."
As usual, BSD quantifies a head count. There's never a statement on the only metric that matters: HOW MANY OF BSD's 25,000 PORTS ARE OUT OF DATE?
Just one dev at Void Linux keeps thousands of packages up to date. So what is wrong with BSD ports tree? Why is it so bit-rotted? I am tired of shiny-shiny in the kernel and init rig and code frameworks. Okay, LLVM is wonderful. BSD finally got there. Now, what about userland apps? Wasn't LLVM supposed to make them easier to build?
There's plenty of porting work in apps; stop fiddling with launchd! Go anywhere online for some app. You will find Windows, Mac, and if you're lucky, Linux packages. BSD is always missing. I wish BSD would focus on app porting. Meanwhile get the ports tree up to date and keep it so, like Void Linux and Gentoo.
P.S. Swap is archaic; turn it off except in special use cases.
14 • UEFI bug (by Didier Spaier on 2016-02-08 06:47:44 GMT from Europe)
Matthew Garrett has published what himself calls a short term fix, see:
As a fix didn't show up in the kernel yet (at least I didn't see one) I assume that he and or someone else is figuring out a more definitive solution.
15 • swap (by zykoda on 2016-02-08 07:35:53 GMT from Europe)
A single 2-4 GB swap partition per hard drive is shared over multi-boot distros mainly as a guard against memory depletion with standard 60 swapiness reduced to 10. No swap partition on SSDs. Never use hibernation or sleep on desktops and have no laptops I note generally very small swap space use except on some ancient 512MB RAM machines. Multiple VM machines may be a valid use of swap.
16 • Swap or not to swap (by Andy Mender on 2016-02-08 07:51:56 GMT from Europe)
I think the poll question needs an additional answer - I have a swap file/partition, but I don't use it often/ever. There is a difference between having swap (always nice, even if only as a backup) and using it explicitly.
Some good news in the FreeBSD quarterly report. We'll see how things develop ;).
17 • Tails_2.0 (by k on 2016-02-08 08:02:42 GMT from Europe)
Thank you Joshua for the comprehensive review of Tails 2.0.
Re: "it is designed more for storing documents and retaining system settings, than for trying to add a large amount of software to the system"
Actually, I understand from earlier Tails developer/support's prompt and well-explained response that addition of software by non-expert user is not advised.
Re: "Selecting "yes" to the "More options?" prompt and clicking the "login" button takes the user to a screen that asks if the user would like to set an administrator password (if the user does not, there is no way to do anything that would require root privileges...
... Because there is no administrator password set by default, some things will not work, which is understandable, but not all of them fail in a user friendly manner. For example, trying to launch the "Root Terminal" application, results in absolutely nothing happening."
Might confuse would-be (new) users of Tails. Like previous builds of Tails, 2.0 gives the user option of "administrative" tasks -- e.g. mounting other disks, directories -- with root privilege/password, at "greeting", after installation, or manual upgrade, and reboot:
click on "More options" --> "login" changes to "forward", click on forward --> options, including administrative password.
Best wishes to all.
18 • Impressions (swap) (by none on 2016-02-08 08:09:52 GMT from Europe)
Your question "Do you use swap space?" makes it almost sound like it would be optional... which in itself is really misguided. https://www.google.fi/search?q=linux+swap+mandatory
19 • swap (by dbrion on 2016-02-08 08:25:33 GMT from Europe)
It would be a bad idea of using swap when GNUlinux is **installed** on an USB stick (it has a limited number of writes : this is enough for editing what one can write, have a package manager working -FC and Mageia can install on flash drives, I learnt Tails could-). As USB sticks are smaller than ordinary disks, swap space is often a -dangerous- waste. If needed -that should be an exception, if one adds enough RAM - , one can ask the Holy Manual for swapon/swapoff or https://access.redhat.com/documentation/en-US/Red_Hat_Enterprise_Linux/5/html/Deployment_Guide/s2-swap-creating-file.html (I am afraid it has been known for years).
20 • /sys/firmware/efi (by Head_on_a_Stick on 2016-02-08 08:27:38 GMT from Europe)
Who wrote the news item on the /sys/firmware/efi problems?
They really should do their research before posting ignorant FUD...
I can assure you that /sys/firmware/efi is indeed mounted rw under SysVinit as otherwise it would not be possible to modify the motherboard NVRAM entries; this would render UEFI boot entry modification impossible.
The problem is *nothing* to do systemd and a correction should be added to the news article ASAP.
21 • tails (by dimitrij on 2016-02-08 08:27:48 GMT from Europe)
This weeks distro review seems pretty flat, not really up to DW standards. Tails being more of a special distribution, not really meant for FB surfing, there should be some introduction for new interested users first, with some background on tor and why and when to use it. I mean not a lot of privacy concerns will warrant the pain of install let alone the big penalty in browsing speed. When not dealing with really sensitive information, it might well be regarded as crap by uneducated users - which it clearly is not, as a lot of people trust tails with their life. For existing users, a link to the changelog would provide more information in a better format. I mean this as a positive, constructive comment to DW.
Surprised of still a rather big percentage using swap. In an era of rather cheap RAM and not so cheap (quality) SSDs, it pays to get rid of old habits. New installers should consider integrating some more logic at auto-partitioning.
22 • swap (by geev03 on 2016-02-08 09:46:26 GMT from Europe)
swap partition keeps load-averages low on this type of single board systems- http://goo.gl/7qAjtr
23 • Surprises with swap. (by dbrion on 2016-02-08 09:57:14 GMT from Europe)
"Surprised of still a rather big percentage using swap". This not that surprising, as it is the default install option in many distributions -at least Mageia (it even mildly protests if one does not need swap partition .... : for FC I do not remember). Swap partition is discouraged on RPis http://www.blaess.fr/christophe/2015/12/08/creation-dun-systeme-complet-avec-buildroot-2015-11/ -but installed on their images-, and seems absurd/dangerous on limited write supports.
24 • swap partition out of habit (by far2fish on 2016-02-08 09:58:46 GMT from Europe)
Like a few others have mentioned too, I also create a swap partition out of habit.
On my laptop I have 16GB memory, and with a 8GB swap partion on SSD, because that is the only HDD I have in my laptop.
On a few occations, I run an IDE with en embedded runtime environment, I sometimes come close to the 16GB RAM limit, but I do not believe I have crossed it yet.
And I do hibernate - all the time. The laptop is never powered off. Perhaps a monthly reboot to activate the latest updated kernel.
25 • @20 - EFI (by Stan on 2016-02-08 10:05:55 GMT from Europe)
Interesting point, I also read a bit on:
Now I feel so lucky that I'm still using the old BIOS. Now I would dare to say is the manufacturer fault as this does not affect all mobos, just some of them. Seriously why the hardware vendor will allow the user to be able to brick the system in that way.
26 • /sys/firmware/efi/efivars/ (by Linux Apocalypsis on 2016-02-08 10:43:39 GMT from Europe)
As announced, these premeditate conflicts between systemd developers and kernel developers are going to become more and more frequent. At some point, systemd developers will say "ok, things cannot go on like this, we have no choice but forking the kernel". RetHat will adopt the forked systemd-kernel, then they will make Debian swallow it as well, Ubuntu will have to follow and Linus will find himself off-side... One ring to rule them all.
27 • KDE Neon (by Frosch on 2016-02-08 13:15:48 GMT from Europe)
I don't understand what the interest of this project is...
If you want the latest KDE packages, then you can use for example Chakra, KaOS, openSUSE Tumbleweed, ...
28 • KDE Neon (by Lizard on 2016-02-08 13:52:53 GMT from South America)
@27 • KDE Neon.
KDE Neon looks interesting if you want to test bleeding edge KDE in a "serious work" machine while having a reliable base with an alternative DE installed to fall back to in case KDE breaks down.
29 • @26 so, Linux world is doomed ? Really ? (by Frederic Bezies on 2016-02-08 14:23:03 GMT from Europe)
Systemd is evil, so kill evil. It is what you're saying ?
I thought it was kernel job to manage hardware, not init system one. How can I possibly be wrong for so long ?
BSDs are the way to go, so... And going back to 2005-2006 for ease of use.
30 • firmware issue (by Jesse on 2016-02-08 14:51:26 GMT from North America)
>> "I can assure you that /sys/firmware/efi is indeed mounted rw under SysVinit as otherwise it would not be possible to modify the motherboard NVRAM entries; this would render UEFI boot entry modification impossible."
I tested it under my machines that use SysV init. The firmware data (located in "efivars") is not mounted. Not even read-only. By comparison, on a systemd machine, the systemd software is hard coded to mount the firmware data in read-write mode.
>> "The problem is *nothing* to do systemd and a correction should be added to the news article ASAP."
It does have something to do with systemd. The bug happens on platforms running systemd, because of systemd's hard coded approach to mount the data in read-write mode. If the data were mounted in read-only mode, the vulnerability would not be exposed. The systemd developers have refused to change the behaviour of their software and mount the data in read-only mode, which means the bug continues to be exposed to userspace applications.
systemd is not the root cause of the issue, but it does leave the door open to the underlying bug and that is a problem.
31 • Swap? (by Vukota on 2016-02-08 15:02:39 GMT from Europe)
Lately I try to avoid using swap partitions at all costs. Most of my systems have enough RAM (thus providing enough swap space would be costly), or have SSD in them (and I do not want to burn them with SWAP partition) or they are very slow machines and using swap on them would make them crawl. Thus logical questions: When do you really need swap partition? Is it going to save you or cause more performance (or reliability) problems?
32 • 32-bit Ubuntu, UEFI in Arch (by jg53 on 2016-02-08 17:19:01 GMT from Poland)
Hopefully, limiting Ubuntu exclusively to 64-bit is only meant for the Unity version. I kindly ask somebody from Canonical confirm, that 32-bit support shall be continued for other Ubuntu flavors - main reason being there are still hundreds of millions of older machines that will still be running for the next 10 years or so.
Ubuntu or Debian, in contrast to Arch, can be today (I mean February 8th ) installed on any machine (bios-mbr, uefi-gpt) without a problem, the user, in the most complicated scenario simply preparing the partition table with gparted. Just yesterday, I tried in vain to install ARCHITECT LINUX ( I installed it twice already, a year and some months before today) - same uefi error coming up and system not booting. Bugs can happen. But this one clearly derives from the wrong development path or strategy. Modern operating systems are already too complicated for an ad hoc citizen militia type approach. The Debian way is IMHO producing stable results, but there ar
e hundreds of developers contributing to a perfectly organized project and every issue can be addressed. The smaller distros, if there will be no progress on the interoperability part, will be doomed to fail. There is simply too much to take care about, and if somebody, like the systemd team, makes arbitrary and unexpected choices, then the rest of the linux community will suffer.
33 • Book review (by dbrion on 2016-02-08 18:11:05 GMT from Europe)
First, it is a nice book review, and I was glad to read it (and to read one could be humorous with command line).
However, something struck me
"Linux manual pages usually do not include helpful examples of how programs are supposed to be used."
Kerrisk (The Linux Programming Interface : http://man7.org/tlpi/ ) wrote a full book, with running demos showing each -maybe- aspect of functions described in the Manual , with C examples and discussion . This can show well written C (and it is a language most microcontroller-oriented people try to use -there are few alternatives...) , Linux specificities (one should have some experience of command line to appreciate it) . These examples can be adapted if one needs something specific.
34 • 32 • 32-bit Ubuntu, UEFI in Arch by jg53 (by AsWas on 2016-02-08 18:20:45 GMT from Europe)
If Ubuntu is going to stop supporting 32 bit, its OK, time to go forward. Debian way might be good, but the Ubuntu way might be different. If one wants to install Arch in a UEFI system, it can be done, but don't complain about Architect Linux. Anyway, Architect Linux can be installed on an uefi system.
35 • Swap creation during install (by Jordan on 2016-02-08 18:26:36 GMT from North America)
Some distros offer the swap partition during install, and some even default to it without a choice by the user (unless it's a user savvy in Gparted etc).
I've not done benchmarks, but it does seem to be a useless partition if there is adequate RAM.
36 • Swap is a Relic (by Justin on 2016-02-08 18:39:59 GMT from North America)
I used to teach people how to speed up their XP machines by removing the swapfile. Most of the time, 1GB+ was plenty of RAM. Nowadays, I run everything I can on RAM disks because it's much faster (and very easy to do in Linux). Maybe it saves only a handful of GBs, but those are _my_ GBs...
FYI, for people who don't know: you can control the "swappiness" of the system: https://en.wikipedia.org/wiki/Swappiness
Also, for the disk-space conscious, you can recover GBs of space from the ext4 filesystem using tune2fs. You want to turn off the default reserve of 5% (~250GBs on 5TB!) or set it to something more sane: tune2fs -m 0 /dev/sda (for 0%). I boot off a small SSD and mount /home elsewhere, so for me, this is getting back a lot of wasted space.
37 • /sys/firmware/efi (by Head_on_a_Stick on 2016-02-08 21:17:29 GMT from Europe)
Users can mount /sys/firmware/efi read only by adding a line to /etc/fstab, as per this ArchWiki article:
38 • @26 efi (by vw72 on 2016-02-08 21:32:43 GMT from North America)
Actually, there isn't a conflict at all between systemd and kernel developers as neither has created this problem. It is bad hardware design that has created it and both groups are trying to figure out the best way to mitigate the problem. Both systemd can handle it and so can the kernel. However, it is waste for both to expend effort doing so.
As such, the conflict is not a conflict at all, but a healthy discussion on how best to help people avoid a manufacturing defect that can render their hardware useless if they make a mistake.
39 • 32-bit Ubuntu, UEFI in Arch (by jg53 on 2016-02-08 22:45:01 GMT from Europe)
Hopefully, limiting Ubuntu exclusively to 64-bit is only meant for the Unity version. I kindly ask somebody from Canonical confirm, that 32-bit support shall be continued for other Ubuntu flavors - main reason being there are still hundreds of millions of older machines that will still be running for the next 10 years or so.
Ubuntu or Debian, in contrast to Arch, can be today (I mean February 8th ) installed on any machine (bios-mbr, uefi-gpt) without a problem, the user, in the most complicated scenario simply preparing the partition table with gparted. Just yesterday, I tried in vain to install ARCHITECT LINUX ( I installed it twice already, a year and some months before today) - same uefi error coming up and system not booting. Bugs can happen. But this one clearly derives from the wrong development path or strategy. Modern operating systems are already too complicated for an ad hoc citizen militia type approach. The Debian way is IMHO producing stable results, but there are hundreds of developers contributing to a perfectly organized project and every issue can be addressed. The smaller distros, if there will be no progress on the interoperability part, will be doomed to fail. There is simply too much to take care about, and if somebody, like the systemd team, makes arbitrary and unexpected choices, then the rest of the linux community will suffer.
40 • Swap (by Nathan Vance on 2016-02-08 23:07:47 GMT from North America)
The only reason I ever used to use swap is when some errant app started eating up memory. One all the RAM was taken, the computer would bog down as the app would start eating swap. If I didn't press the power button, Linux would crash a few minutes later. Now I don't use swap, and my computer crashes earlier without going through elongated death throes.
41 • systemd & 32 bit support (by M.Z. on 2016-02-08 23:21:45 GMT from North America)
@26 - sytemd
Sure people always say the sky is falling, but how often does it actually happen? That kind of talk is nonsense & if you don't like systemd there are plenty of good distros to choose from, so stop playing chicken little.
@39 - 32 bit support
If you read the mailing list link it's really only Unity DE versions being considered for 64-bit only. As I mentioned above, it sort of makes sense if it's limited to the biggest DEs. A quote from the link:
"In 2016, people with i386-only hardware are unlikely to be capable to
run Unity 7 Desktop, and probably run other Ubuntu variants. I guess
there are some accidental i386 users, e.g. those that have installed
i386 variant on amd64 hardware.
Does it still make sense to build ubuntu-desktop-i386.iso? Validate
it? Test it on amd64 hardware? Ship it?"
Of course I always recommend against Ubuntu anyway because they got sleazy & started stuffing spyware into the edition they are discussing, but if anyone needs another excuse to try a more ethical distro I guess this 32 bit thing is an OK one.
42 • healthy discussion (by anticapitalista on 2016-02-08 23:54:20 GMT from Europe)
@38 Bad hardware is certainly the cause of the problem, but systemd devs REFUSED to change (at least initially - not sure what their plan is now) from mounting /sys/firmware/efi default as rw.
43 • Hacking vs Security (by null security on 2016-02-09 02:29:52 GMT from North America)
If you compare pentesting (hacking) distros to security/privacy distros there is a huge difference
* Pentesting distros promote lots of (mostly hacking) tools - BlackArch over 1300.
* Security/privacy distros promote only about 10 or 20 tools.
It shows that Linux is primarily a hacker's distro, and security/privacy focused distros are sorely lacking in features.
44 • Swap - last century hardware. (by Greg Zeng on 2016-02-09 03:33:49 GMT from Oceania)
My Dell notebook (XPS-15, 2013) has 16 GB DDR3 memory. No swap memory is needed for both Linux and Windows-10 operating systems.
The exceptions in Linux is booting from a USB stick or drive, which is much slower than the motherboard's electronic, spinning-drive or SSD drive. My boot menu offers any of two Win-10, or ten Linux partitions on the motherboard, or any drives connected to the three (or more) USB ports. For the non-motherboard operating systems, I do have a "spare" 2GB partition, accessible by any Linux operating system running on my lightweight notebook.
Users who seem to be limited with USB ports (not enough or "too close together") are unaware of modern technologies. USB hubs, four ports or more, start at just a few dollars (US). I have very few problems using the later (2014+) versions of these multi-port hubs, even with many high speed USB3 devices.
SSD FUD is repeated in comments here. Linux's recent kernels auto-adjust for today's SSDs. Old Linux, trapped into poor Linux kernels (PCLOS, etc) cannot use SSD technology properly. The older wear-tear with limited read-writes does not apply to today's SSDs. IMHO they might soon become the most reliable version of long term data storage ever invented; far better than ink and/ or magnetic memory.
45 • 32 bit support (by OkVern on 2016-02-09 04:16:32 GMT from Europe)
I am sure there are a lot of people that have 32 bit only rigs...so if Ubuntu does drop 32bit, there are plenty other Linux alternatives that will support it.
Futurecast: Bring on the 128 bit'ers. *Really?* Yes really. :D
46 • systemd (by Bill on 2016-02-09 05:10:18 GMT from Oceania)
Happily running Slackware, NetBSD and Manjaro OpenRC here, no systemd in sight!
47 • @42 - healthy discussion (by Stan on 2016-02-09 08:26:50 GMT from Europe)
Systemd refused WITH an explanation why it is like that:
Someone else already linked the explanation:
In any case setting up to RO means that none of this will work:
Is it really systemd the problem? They already gave the workaround and reason, now we all have to figure out how to fix the mess that some hardware manufacturer caused.
48 • @46 is an init system responsible for an hardware bug ? (by Frederic Bezies on 2016-02-09 08:29:36 GMT from Europe)
So, an init system is responsible for an hardware bug ? I can remember back in the middle 2000 how many people had problems with VIA chipsets. Or with wifi chipsets.
Do we have to blame kernel for these low quality chipset ?
Looks like kindergarten war here. I do not have an uefi crippled computer in my home.
Not for now, but I will have one. And by the way, there is a workaround which is horribly long to set up - thanks "head on a stick" for it.
Just waiting for a big bad bug with OpenRC in order to laugh at you. But I won't. I'm a grown up, which doesn't care if his computer is starting with OpenRC, systemd or runit or upstart or sysVinit. I just want to him to work.
49 • systemd (by Bill on 2016-02-09 08:35:29 GMT from Oceania)
"I'm a grown up, which doesn't care if his computer is starting with OpenRC, systemd or runit or upstart or sysVinit. I just want to him to work."
If you don't care, cool, just ignore it. But I do think that caring about some things is worthwhile, including UNIX principles, so I avoid systemd.
50 • @44 SSD!= SD (RPi's disk/booting partition ); (by dbrion on 2016-02-09 08:42:08 GMT from Europe)
What does FUD mean.
Oh, BTW : RapsberryPis are getting very popular, and boot on a SD . Many people try to avoid writes, not to wear it out too soon.
FramboisePi (in French) recommended keeping the boot partition -FAT, RO-and putting everything one needs (i.e /usr -to upgrade- /home -to keep one's work - /var ) on en external USB rotating drive. Of course, swap partition is removed (this leads to very small and cheap SD "disks"). Less extreme tweakings involve removing swap partition http://www.ideaheap.com/2013/07/stopping-sd-card-corruption-on-a-raspberry-pi/
It should not be that difficult to know how many writes are involved when swpa occurs, and, knowing the number of writes a USB stick can endure... choose, when one has a system installed (not lazy installed) on an USB -reads are as comfortable as a rotating disk, writes may be slower (another reason to avoid swap) -, whether swap is dangerous or not...
51 • Majaro-ARM (by Peter on 2016-02-09 12:59:26 GMT from Europe)
Manjaro-ARM seems to be a complete effort to port Manjaro over to ARM. Why don't they just help the Archlinux ARM project and build their own software on top of that? Seems like less work and the greater community would profit from it.
52 • @50 FUD (by Jordan on 2016-02-09 20:41:17 GMT from North America)
Fear. Uncertainty. Doubt.
No relation to Elmer.
53 • ZRAM, ZSWAP and compressed filesystems (by technosaurus on 2016-02-09 21:14:17 GMT from North America)
Currently the kernel has some good features to help with swap.
ZRAM can be helpful to store more (compressed) data in RAM
ZSWAP can be helpful to store more (compressed) data on a fast drive (SSDs, etc...)
JFFS2, BTRFS and other filesystems support the same types of compression.
Unfortunately even if the filesystem uses page sized compressed blocks with the same compression type, all data is still decompressed and then recompressed. This is also true between zram and zswap even when they use the same compression type.
Its not just the kernel, most file downloaders and browsers support gzip compression and will still decompress the data only for it to be recompressed on disk... of course this would require kernel support as well.
Its great that the linux kernel can work with so many different devices, but its too bad there aren't enough resources left to make the components work with each other.
54 • Time for FOSS Firmware (by Arch Watcher 402563 on 2016-02-10 00:52:53 GMT from North America)
Firmware is too complicated. Why do FOSS folks accept any crud hardware vendors foist on them? If we're going to have complex firmware, it should be FOSS, like coreboot.org or openwrt.org or libreboot.org, allowing proper initrd / initramfs images.
And I'm even wondering how much sense has the whole dance: UEFI -> InitRamFS -> Kernel. Why not just burn a kernel into firmware? FOSS firmware code can be added to or called from the kernel.
And I would like motherboard hacks to replace vendor chips. A commercialized kit would sell like crazy. Hardware OEMs don't realize what an untapped market and software labor savings they'd have in open-source firmware or just mobos with bare sockets.
55 • Time for FOSS Firmware (by Arch Watcher 402563 on 2016-02-10 01:17:36 GMT from North America)
56 • 32-bit Ubuntu support, Aquaris M10 Ubuntu Edition (by RO on 2016-02-10 04:12:34 GMT from North America)
@39, 41, 45:
57 • 32-bit Ubuntu support, Aquaris M10 Ubuntu Edition (v.2) (by RO on 2016-02-10 04:31:04 GMT from North America)
[Seems I hit enter in the wrong place/time for "v.1" ...]
@39, 41, 45:
It would indeed seem to be necessary for Canonical to continue 32-bit support for both their older LTS versions that are running the 32-bit mode, and for their shiny new Aquaris M10 since it runs on an ARM Cortex A7 CPU, which I understand to be 32-bit architecture, right?
Further regarding that tablet, I would be MUCH more impressed if Canonical, and other distros, would work on expanding Linux availability to all the existing 32-bit Windows 8/10 tablets that are out there now. It seems that although most of them have 64-bit capable Atom CPU's, they are set up with 32-bit UEFI as well as their 32-bit Windows installations (to save money/licensing costs per an Intel insider article I found once on the matter - GoogDuckBing for yourself to see if you can find it...). The only exception I am aware of so far is the Atom-based Microsoft Surface 3 (not "Pro" model - those are i3/5/7), which does run 64-bit Windows 8/10 - anyone have any luck installing any Linux distro on that?
The only distro I have found so far that can handle such tablets as my Acer W3, HP Stream 7, and Lenovo Miix 10 2, is Knoppix 7.4 , which can boot UEFI either in 32 or 64-bit mode (7.6 seems broken in that regard, but maybe I have not generated the flash drive correctly unetbootin?)., however, its kernel does not have drivers for the wonky wifi/touchscreen/etc hardware. Fix those issues to repurpose such tablets, and that would get a lot more uptake than this one-off ARM tablet that is getting so much attention.
58 • Rebecca_black_back... (by k on 2016-02-10 07:10:01 GMT from Europe)
Wayland intrigues. Please Distrowatch, post review after thorough trial. Thank you.
59 • 57 • Aquaris M10 Ubuntu Ed. (by Kragle von Schnitzelbank on 2016-02-10 16:43:39 GMT from North America)
MediaTek MT8163A based on ARM Cortex-A53 with A64 instruction set ARMv8-A, AArch64 architecture, and 32-bit Thumb instruction set. RISC. LPAE. AES, SHA-256.
It's complicated, and _not_ like x86. Power-efficient, ...
60 • 32-bit Ubuntu (by Drone on 2016-02-11 05:09:54 GMT from Asia)
Dropping 32-bit support should ONLY be for the bloated Unity abomination. The rest of us NEED 32-bit support. Even today, the likes of Lenovo, Acer, Asus, etc. are selling millions of intentionally hobbled new laptops with only 2GB of SDRAM that CAN NOT be user upgraded. These machines need 32-bit support to have a hope in heck of being useful for any real work. So it is not just the millions of older machines out there that need 32-bit support, it's plenty of new machines too.
61 • Tor_Browser_5.5_on_Tails_2.0... (by k on 2016-02-11 13:08:07 GMT from Europe)
... does not block fingerprint/tracking as previous Tor Browser on Tails 1.8.2, nor is it as privacy-protecting as latest Tor Browser 5.5.1 on other Debian-based operating system.
62 • ubuntu (by torz on 2016-02-11 21:33:05 GMT from North America)
dang, i'm still using intel atom which only supports 32-bit!
63 • 62 • Intel Atom (by Kragle on 2016-02-12 06:25:52 GMT from North America)
Some Atom CPUs are able to do 64-bit; few are given enough RAM to do it well.
With an OS that doesn't use resources skillfully, that is.
64 • Re: Intel Atom (by Andy Bear on 2016-02-12 07:19:28 GMT from Europe)
I believe the BSDs would deserve some praise in that respect as they are very good at RAM handling. I noticed that the same GNU apps (namely, Chromium and Firefox), running on FreeBSD, use substantially less RAM than on Linux.
If the wireless card is supported, I think it's actually a pretty good combination (FreeBSD/OpenBSD + Intel Atom ultrabook/netbook) :).
65 • @57 32bit (by greg on 2016-02-12 08:15:07 GMT from Europe)
the issue discussed here is not ARM image, but "intel" (i386) 32 bit which is a totally different CPU architecture than ARM.
66 • Manjaro ease of use etc.. and now ARM (by Jordan on 2016-02-12 13:47:14 GMT from North America)
I've never been a one distro fanboi, but Manjaro may be changing that. Rolling releases have always attracted me, this one's no exception and the devs wait a bit for stability issues to be washed out before releasing updates, making Manjaro even more attractive as a rolling distro.
Hardware detection, and all that. Even a vast array of drivers in AUR, obscure printer drivers and on and on.
And now embedded device universe: Manjaro-ARM.
67 • Swap (by ErikWestre on 2016-02-13 00:54:19 GMT from Europe)
I haven't been using swap for many years. Don't need hibernation, as SSD boots superfast any way. Swap is waisted space.
68 • Basic Linux Operating System (by Jeff Manis on 2016-02-13 09:08:41 GMT from North America)
I would like to see someone develop a Basic Operating System (BOS) that only addresses the issues of system components on installation. For example, allowing the selection of all video drivers, all network drivers, all sound drivers, all package managers, all desktops, all partition options, etcetera. Then after the BOS is installed allow a selection of software components to be installed from the selected repositories. The problem with most distributions are; that hardware compatibility seems to be the least important and bundled applications the most important, and in my opinion it should be the other way around.
69 • Basic Linux Operating System (by Jeff Manis on 2016-02-13 09:13:21 GMT from North America)
I would like to see someone develop a Basic Operating System (BOS) that only addresses the issues of system components on installation. For example, allowing the selection of all video drivers, all network drivers, all sound drivers, all package managers, all desktops, all partition options, etcetera. Then after the BOS is installed allow a selection of software components to be installed from the selected repositories. The problem with most distributions are; that hardware compatibility seems to be the least important and bundled applications the most important, and in my opinion it should be the other way around. SWAP should also be option-able as a choice also during installation, such as RAM swap.
70 • Basic Linux Operating System (by GuntherT on 2016-02-13 16:21:22 GMT from North America)
Have you tried Arch?
71 • Gentoo & others? (by M.Z. on 2016-02-13 19:45:07 GMT from North America)
@69 & 70
Aren't there multiple distros that do that in addition to Arch, such as Gentoo? Perhaps Slackware as well? I think there are a number DIY distro options that let you make it do exactly as you want, but they are more manually intensive & less known to average users. I don't think that there are many distro niches like that that aren't already handled to some extent, & if you try to start digging into the thousands of tiny lesser known distro projects that are done by individuals around the world I seriously doubt there are many such niches that aren't filled to some extant; however, many small niches may be filled sporadically by a few tiny fly by night efforts that disappear & reappear seemingly at random based of available time & interest from the developers.
Number of Comments: 71
Display mode: DWW Only • Comments Only • Both DWW and Comments
|• Issue 658 (2016-04-25): Kali Linux 2016.1, elementary OS 0.3.2, Debian elects Project Leader, Fedora 24 feature preview, Nard reaches 1.0|
|• Issue 657 (2016-04-18): Redox, Linux Mint improves update manager, planned Fedora 24 features, Ubuntu 16.04 getting Snappy packages|
|• Issue 656 (2016-04-11): Qubes OS 3.1, Whonix offers bug bounties, Puppy's family tree, setting up disk partitions and running bash on Windows|
|• Issue 655 (2016-04-04): Parsix 8.5, Sabayon's Community repository, Red Hat offers free subscriptions, Ubuntu tablets, command line tips|
|• Issue 654 (2016-03-28): PCLinuxOS 2016.03, Using signatures to create a web of trust, Arch Linux rolls out Pacman update, GuixSD packages GNOME|
|• Issue 653 (2016-03-21): Antergos 2016.02.21, Debian prepares for election, a Unix-like OS written in Rust, watching Netflix on FreeBSD|
|• Issue 652 (2016-03-14): ReactOS 0.4.0, Debian swaps Iceweasel for Firefox, Fedora moving forward with Wayland, Verifying ISO files|
|• Issue 651 (2016-03-07): Korora 23, Linux Mint improves security, Ubuntu MATE on Raspberry Pi 3 computers, trying different file systems|
|• Issue 650 (2016-02-29): Haiku in 2016, running Android apps on GNU/Linux, 30 years of MINIX, Fedora plans Atomic Workstation|
|• Issue 649 (2016-02-22): Zorin OS 11, openSUSE launches new editions, Linux Mint website compromised, sandboxing applications using Firejail|
|• Issue 648 (2016-02-15): XStream Desktop 153, Raspbian unveils OpenGL feature, free hardware, Ikey Doherty talks desktop design|
|• Issue 647 (2016-02-08): Tails 2.0, KDE project launches Neon, Manjaro unveils ARM support, FreeBSD's quarterly report|
|• Issue 646 (2016-02-01): deepin 15, Mint plans X-Apps, FreeBSD to support boot environments, logging into the desktop as root|
|• Issue 645 (2016-01-25): Linux Mint 17.3 "Xfce", Chromixium changes its name, Ubuntu tablets coming soon, Linux vs BSD comparision|
|• Issue 644 (2016-01-18): Kwort 4.3, Sabayon tests ARM images, Slackware adopts PulseAudio, running Linux without GNU software|
|• Issue 643 (2016-01-11): Solus 1.0, Mint provide upgrade path to 17.3, Fedora developers work on stability, running the LXQt desktop|
|• Issue 642 (2016-01-04): paldo GNU/Linux, vetting distro repositories, Fedora plans to adopt GCC 6, Ian Murdock passes|
|• Issue 641 (2015-12-21): Arch Linux, Qubes OS to ship on Librem laptops, ALT offers start kit images, the spread of systemd and launchd|
|• Issue 640 (2015-12-14): Chakra GNU/Linux 2015.11, removing meta-data from files, Ubuntu to remove on-line dash searches|
|• Issue 639 (2015-12-07): OpenBSD 5.8, openSUSE gathers Summer of Code proposals, running WINE on a live disc, Enlightenment adds Wayland support|
|• Issue 638 (2015-11-30): Qubes OS 3.0, KaOS with Plasma, NetBSD 7.0, Fedora seeks Wayland testers, scheduling tasks|
|• Issue 637 (2015-11-23): NixOS 15.09, Antergos introduces ZFS support, MINIX shares new features, copying an OS to a new computer|
|• Issue 636 (2015-11-16): openSUSE 42.1, Fedora uses Wayland by default, Debian replaces live CD project, Steam consoles launch|
|• Issue 635 (2015-11-09): Fedora 23, Cinnamon 2.8 released, a Fedora KDE packager quits, Red Hat signs deal with Microsoft|
|• Issue 634 (2015-11-02): Ubuntu 15.10, Chakra upgrades to Plasma 5, OpenMandriva plans new editions, MINIX plans conference|
|• Issue 633 (2015-10-26): GhostBSD 10.1, Bodhi Linux to get new settings panel, Fedora 23 delayed, creating live image of existing OS|
|• Issue 632 (2015-10-19): Linux Lite 2.6, 32-bit build of CentOS, OpenBSD turns 20, Bodhi Linux releases AppPack|
|• Issue 631 (2015-10-12): Parsix 8.0, Manjaro seeks new artwork, sending commands to multiple servers, Debian drops LSB support|
|• Issue 630 (2015-10-05): Android-x86 4.4-r3, Ubuntu's new installer, Raspbian defaults to GUI interface, cleaning out dot files|
|• Issue 629 (2015-09-28): Open source desktops and touch interfaces, locking down user accounts, OpenMandriva opens gaming documentation|
|• Issue 628 (2015-09-21): Neptune 4.4, changes to pfSense, Pinguy OS releases updated ISO images, accessing hard disk images|
|• Issue 627 (2015-09-14): Mageia 5, Snappy co-exists with Debian packages, creating PDF/A documents, Antergos previews Poodle|
|• Issue 626 (2015-09-07): Status of Wayland and Mir, Cinnamon improvements, an OpenBSD hypervisor, HAMMER2 gets deduplication|
|• Issue 625 (2015-08-31): OpenELEC 5.0.8, Fedora's new Wayland features, Tails releases update, the LILO boot loader|
|• Issue 624 (2015-08-24): Zorin OS 10, Sabayon's new features, Solus seeks funding, Debian turns 22, new PC-BSD repository|
|• Issue 623 (2015-08-17): VectorLinux 7.1, Ubuntu One source released, Moksha Desktop ships in Bodhi, Fedora developers debate Chromium|
|• Issue 622 (2015-08-10): antiX 15, Fedora tests kdbus, Debian tracks UEFI issues, word processors for the CLI|
|• Issue 621 (2015-08-03): Point Linux 3.0, Debian drops Sparc, Fedora package stats, VirtualBox 5.0|
|• Issue 620 (2015-07-27): Debian GNU/Hurd 2015, Linux Bible, Ubuntu MATE gets new Welcome app, Telegram on Fedora, Plasma Mobile|
|• Issue 619 (2015-07-20): SolydXK 201506, Tanglu's new bug tracker, FSF and Canonical negotiate licensing, Haiku unveils new init system|
|• Issue 618 (2015-07-13): Semplice Linux 7, openSUSE derivatives, Debian adopts GCC 5, Docker ported to FreeBSD|
|• Issue 617 (2015-07-06): Alpine linux 3.2.0, Fedora on MIPS CPUs, Solus offers daily builds, Ubuntu migrating to Snappy|
|• Issue 616 (2015-06-29): MidnightBSD 0.6, openSUSE's "42", encryption added to the ext4 file system, FreeBSD on a Raspberry Pi|
|• Issue 615 (2015-06-22): Raspbian 2015, Fedora works around Intel driver issue, openSUSE adopts GCC 5, frozen desktop while copying files|
|• Issue 614 (2015-06-15): Chromixium OS 1.0, Debian 8.1 released, OpenBSD running in the cloud, sudo myths|
|• Issue 613 (2015-06-08): Fedora 22, Cinnamon 2.6 released, FreeBSD's history, working around Secure Boot|
|• Issue 612 (2015-06-01): Manjaro OpenRC, Debian, Devuan and systemd, Fedora 22 released, Mandriva closes its doors|
|• Issue 611 (2015-05-25): Kubuntu 15.04, openSUSE adopts Plasma 5, Ubuntu's Snappy, words from Debian's Neil McGovern|
|• Issue 610 (2015-05-18): NethServer 6.6, interview with Neil McGovern, CentOS supports AArach64, Foresight discontinued|
|• Issue 609 (2015-05-11): OpenIndiana 2015.03, LXLE 14.04, PC-BSD Current, creating ISO images, Ask A Leader with Peter Ganten|
|• Issue 608 (2015-05-04): Debian 8.0, Bodhi forks Enlightenment, new Debian GNU/Hurd release, distribution release frequency|
|• Issue 607 (2015-04-27): Ubuntu 15.04, Chapeau 21, Debian 8.0 features, Fedora 22 Beta details|
|• Full list of all issues|