Re: [DNG] vdev
> We all miss Jude and a decisive contribution from his part to have a > running Vdev on Devuan. It's more than 2 years since the inception of > Vdev and more than one year since Jude has "almost finished" it. Yet we > haven't a replacement for Udev, while it is more and more integrated to > systemd. I see this as a major threat. I'm deceived that Skarnet hasn't > written their own, because it's more critical than a supervisor. > Surprisingly the last update on Jude's github project for vdev is 23rd April 2016, only a couple of month's ago. See: https://github.com/jcnelson/vdev/commits/master So it's not that long ago... I guess what we need is to get a working package in Devuan to attract his attention back. Once I've got completed a couple of key items we need to get on the road to a final release for Devuan, I'll look at the state of various packaging attempts for vdev and work towards an official devuan package. D -- Daniel Reurich Centurion Computer Technology (2005) Ltd. 021 797 722 signature.asc Description: OpenPGP digital signature ___ Dng mailing list Dng@lists.dyne.org https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng
Re: [DNG] vdev
As i announced yersterday in the IRC Channel, i just uploaded the repository of gnuinos containing (in addition to linux-libre-4.6.2 and simple-netaid) the packages of vdev: deb http://packages.gnuinos.org/ jessie main deb-src http://packages.gnuinos.org/ jessie main So does it work with devuan then? Can you go past the login screen? The login screen? What are you referring to? Aitor. OK I was not using your vdev package but a hacked in git version of vdev. If I do "vdev --once ..." I get device files in the designated directory however at reboot with vdev when I get to loginmanager - Is it slim in devuan? Anyway at this login view I have no keyboard or mouse. Now I have had some problems with devtmpfs and vdev on another OS? Not sure this is the problem though? This guy seems to be getting the same behavior as me https://github.com/jcnelson/vdev/issues/99 I was interested if you were getting the same behavior or if you had found a solution? ___ Dng mailing list Dng@lists.dyne.org https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng
Re: [DNG] vdev
On Sun, 7 Aug 2016 19:50:14 +1200, Daniel wrote in message <57a6e836.4040...@centurion.net.nz>: > > > We all miss Jude and a decisive contribution from his part to > > have a running Vdev on Devuan. It's more than 2 years since the > > inception of Vdev and more than one year since Jude has "almost > > finished" it. Yet we haven't a replacement for Udev, while it is > > more and more integrated to systemd. I see this as a major threat. > > I'm deceived that Skarnet hasn't written their own, because it's > > more critical than a supervisor. > > > > Surprisingly the last update on Jude's github project for vdev is 23rd > April 2016, only a couple of month's ago. See: > https://github.com/jcnelson/vdev/commits/master > > So it's not that long ago... > > I guess what we need is to get a working package in Devuan to attract > his attention back. ..by "working", we should mean flawless packaging and easy bug reporting so the vdev developers get feedback from people trying the packages so they get help finding the bugs. ..with no feedback, he can justify the "no interest" conclusion. ..which proves the current "git clone -> .configure -> make -> make install" route unworkable, at the very least we need a way to make our omn vdev .dev package from our own git trees. > Once I've got completed a couple of key items we need to get on the > road to a final release for Devuan, I'll look at the state of various > packaging attempts for vdev and work towards an official devuan > package. > > D > -- ..med vennlig hilsen = with Kind Regards from Arnt Karlsen ...with a number of polar bear hunters in his ancestry... Scenarios always come in sets of three: best case, worst case, and just in case. ___ Dng mailing list Dng@lists.dyne.org https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng
Re: [DNG] vdev
On Sat, 6 Aug 2016 10:15:59 +0200 aitor_czr wrote: > As i announced yersterday in the IRC Channel, i just uploaded the > repository of gnuinos containing (in addition to linux-libre-4.6.2 > and simple-netaid) the packages of vdev: > > deb http://packages.gnuinos.org/ jessie main > deb-src http://packages.gnuinos.org/ jessie main > > Linux-libre has been built using libudev-dev. Now, i'll try it again > using Jude Nelson's *libudev-compat* (which should be recalled to > libudev-compat-dev ?? ), instead of libudev-dev. > > I also should add the following line to debian/control: > > Provides: libudev-dev Great. I'll make an extra install of Jessie this week to test the packages (I run Ascii for the moment). Packaging is essential for testing vdev by many people. A mature Devuan absolutely needs a replacement for udev. It's essential IMHO. If Devuan stays with udev, it will be swallowed by systemd one day or another. R. -- richard lucassen http://contact.xaq.nl/ ___ Dng mailing list Dng@lists.dyne.org https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng
[DNG] libvirt - can't create snapshot
Hi, I test devuan with libvirt (Virtual Machine Manager 0.9.1) on Debian 7.11. I would like to create snapshot but when I type as root: 1) "virsh snapshot-create-as --domain devuan_jessie --name snapshot1 --description "clear system" error: failed to get domain 'devuan_jessie' error: An error occurred, but the cause is unknown". 2) "virsh list --all" command return empty list. Has anyone had this problem before? Regards, Paweł ___ Dng mailing list Dng@lists.dyne.org https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng
Re: [DNG] libvirt - can't create snapshot
On 08/08/16 00:37, Paweł Cholewiński wrote: Hi, I test devuan with libvirt (Virtual Machine Manager 0.9.1) on Debian 7.11. I would like to create snapshot but when I type as root: 1) "virsh snapshot-create-as --domain devuan_jessie --name snapshot1 --description "clear system" error: failed to get domain 'devuan_jessie' error: An error occurred, but the cause is unknown". 2) "virsh list --all" command return empty list. Has anyone had this problem before? Only when the machine(s) were created as non-root. Ralph. Regards, Paweł ___ Dng mailing list Dng@lists.dyne.org https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng ___ Dng mailing list Dng@lists.dyne.org https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng
Re: [DNG] vdev
On Sun, 7 Aug 2016 08:14:18 +0200 Didier Kryn wrote: > We all miss Jude and a decisive contribution from his part to > have a running Vdev on Devuan. It's more than 2 years since the > inception of Vdev and more than one year since Jude has "almost > finished" it. Yet we haven't a replacement for Udev, while it is more > and more integrated to systemd. I see this as a major threat. I'm > deceived that Skarnet hasn't written their own, because it's more > critical than a supervisor. It's also much harder to write than a supervisor. If I had a couple months with nothing else to do, I could write a supervisor. I'd need more than a year to write software that loads and maintains all devices, and might not be able to do it at all. Also, Skarnet is in the business of supervisors, that's what they do. That's their core competency. SteveT Steve Litt August 2016 featured book: Manager's Guide to Technical Troubleshooting Brand new, second edition http://www.troubleshooters.com/mgr ___ Dng mailing list Dng@lists.dyne.org https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng
Re: [DNG] vdev
On 07/08/16 22:09, richard lucassen wrote: On Sat, 6 Aug 2016 10:15:59 +0200 aitor_czr wrote: As i announced yersterday in the IRC Channel, i just uploaded the repository of gnuinos containing (in addition to linux-libre-4.6.2 and simple-netaid) the packages of vdev: deb http://packages.gnuinos.org/ jessie main deb-src http://packages.gnuinos.org/ jessie main Linux-libre has been built using libudev-dev. Now, i'll try it again using Jude Nelson's *libudev-compat* (which should be recalled to libudev-compat-dev ?? ), instead of libudev-dev. I also should add the following line to debian/control: Provides: libudev-dev Great. I'll make an extra install of Jessie this week to test the packages (I run Ascii for the moment). Packaging is essential for testing vdev by many people. A mature Devuan absolutely needs a replacement for udev. It's essential IMHO. If Devuan stays with udev, it will be swallowed by systemd one day or another. R. FWIW, I know very little about Devuan package building, so I jumped straight to compiling the github source for vdev, and installing from it. After just a couple of tweaks of the runtime and initramfs configurations, I've made a successful replacement on a pristine Devuan 1.0.0 install, with the 3.16.0-4-amd64 kernel. Of course I don't know anything about how complete it is in respect of handling all sorts of devices. Ralph. ___ Dng mailing list Dng@lists.dyne.org https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng
[DNG] (no subject)
Running testing/unstable applications in Debian stable ___ Dng mailing list Dng@lists.dyne.org https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng
[DNG] SystemD's brownie points over non-systemd OSs.
Hi All, http://forums.debian.net/viewtopic.php?f=16&p=621842#p621842 Excuse me for the topic title. But the above link at first looked like some inherent advantage in using SystemD. However, after a little reflection, a couple of minutes, it seems there are actually no extra brownie-points in using SystemD. This *appears* like an advantage SystemD users have over non-systemd users. But if Devuan allows the use of virtualisation software, the same can be achieved without requiring any sort of benediction SystemD. virtualbox is the Devuan 64 bit repository. So, the above is not an advantage but another way to make a cup of tea rather than the known ways. -- If you can't explain it simply, you don't understand it well enough. Albert Einstein ___ Dng mailing list Dng@lists.dyne.org https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng
[DNG] Fearsome rumblings from GNU
Greetings everyone. This is just a heads up, GNU has had a few changes floating around for a while, and it looks like they are finally making it into distributions. Currently, only Fedora is affected (which I doubt anyone here uses), but it is possible that these changes will make it to Debian as well. Attached is a copy of one user's thoughts on the `ls` command. I'm not sure on the specifics of the change, but for most scripted uses you would probably pass it the `-1` flag anyway. If it doesn't list one file per line anymore, that could be quite serious: we would need to use `dir` instead of `ls`! That's a whole extra keystroke! If the changes turn out to be serious, you might want to try suckless.org's version. (I believe it is called "sutils" and "putils", I can confirm that they work.) - Forwarded message from Mark Clarkson - Date: Thu, 04 Aug 2016 17:16:29 +0100 From: Mark Clarkson To: coreut...@gnu.org Subject: ls output changed !! Hi, This change nearly gave me a heart attack - do you want that on your conscience? hmm? Now we'll have to wait months after you change it back for the change to make it back into Fedora. I really can't believe someone would make a change like this to a core utility - I was stunned. For a brief moment I thought I'd been hacked. Please don't do stuff like that without saying first. I mean it should really be coded into the boot loader, yep, put it in grub so you can't even start the OS without out typing, 'I understand that ls output has changed and it will be shocking'. Please put it back ASAP. Thank you. Cheers! Mark Clarkson - End forwarded message - -- Success usually occurs in private. Failure, in public. signature.asc Description: Digital signature ___ Dng mailing list Dng@lists.dyne.org https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng
Re: [DNG] Fearsome rumblings from GNU
> On August 7, 2016 at 1:22 PM Brian Nash wrote: > > Greetings everyone. > > This is just a heads up, GNU has had a few changes floating around for a > while, and it looks like they are finally making it into distributions. > > Currently, only Fedora is affected (which I doubt anyone here uses), > but it is possible that these changes will make it to Debian as well. > > Attached is a copy of one user's thoughts on the `ls` command. As far as I can see, the original poster doesn't say what the change is. (Maybe some attachments got lost in the forwarding?) I haven't looked into the changelog of coreutils to try to find out, but it would be helpful to know what is at issue here. Peter Olson ___ Dng mailing list Dng@lists.dyne.org https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng
Re: [DNG] Fearsome rumblings from GNU
On Sun, Aug 07, 2016 at 01:33:32PM -0400, Peter Olson wrote: > > On August 7, 2016 at 1:22 PM Brian Nash wrote: > > > > Greetings everyone. > > > > This is just a heads up, GNU has had a few changes floating around for a > > while, and it looks like they are finally making it into distributions. > > > > Currently, only Fedora is affected (which I doubt anyone here uses), > > but it is possible that these changes will make it to Debian as well. > > > > As far as I can see, the original poster doesn't say what the change is. > (Maybe some attachments got lost in the forwarding?) > > I haven't looked into the changelog of coreutils to try to find out, but it > would be helpful to know what is at issue here. I guess this is about weird, superflous quoting in “ls”: https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=813164 Debian was dilligent enough to revert the patch (half a year ago). Fedora – not so much. -- Tomasz Torcz ,,(...) today's high-end is tomorrow's embedded processor.'' xmpp: zdzich...@chrome.pl -- Mitchell Blank on LKML ___ Dng mailing list Dng@lists.dyne.org https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng
Re: [DNG] vdev
FWIW, I know very little about Devuan package building, so I jumped straight to compiling the github source for vdev, and installing from it. After just a couple of tweaks of the runtime and initramfs configurations, I've made a successful replacement on a pristine Devuan 1.0.0 install, with the 3.16.0-4-amd64 kernel. Of course I don't know anything about how complete it is in respect of handling all sorts of devices. By any chance you could share procedure since I tried the same and failed? Not git and compiling of course but how you got it to cooperate with Devuan /scooby ___ Dng mailing list Dng@lists.dyne.org https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng
Re: [DNG] libvirt - can't create snapshot
W dniu 07.08.2016 o 16:46, Ralph Ronnquist pisze: On 08/08/16 00:37, Paweł Cholewiński wrote: Hi, I test devuan with libvirt (Virtual Machine Manager 0.9.1) on Debian 7.11. I would like to create snapshot but when I type as root: 1) "virsh snapshot-create-as --domain devuan_jessie --name snapshot1 --description "clear system" error: failed to get domain 'devuan_jessie' error: An error occurred, but the cause is unknown". 2) "virsh list --all" command return empty list. Has anyone had this problem before? Only when the machine(s) were created as non-root. Ralph. Regards, Paweł ___ Dng mailing list Dng@lists.dyne.org https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng ___ Dng mailing list Dng@lists.dyne.org https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng Thanks Ralph. I read that I must add "--connect qemu:///system" parameter to be able to create snapshot. Regards Paweł ___ Dng mailing list Dng@lists.dyne.org https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng
Re: [DNG] Fearsome rumblings from GNU
On Sun, 7 Aug 2016 20:00:17 +0200 Tomasz Torcz wrote: > On Sun, Aug 07, 2016 at 01:33:32PM -0400, Peter Olson wrote: > > > On August 7, 2016 at 1:22 PM Brian Nash wrote: > > > > > > Greetings everyone. > > > > > > This is just a heads up, GNU has had a few changes floating > > > around for a while, and it looks like they are finally making it > > > into distributions. > > > > > > Currently, only Fedora is affected (which I doubt anyone here > > > uses), but it is possible that these changes will make it to > > > Debian as well. > > > > As far as I can see, the original poster doesn't say what the > > change is. (Maybe some attachments got lost in the forwarding?) > > > > I haven't looked into the changelog of coreutils to try to find > > out, but it would be helpful to know what is at issue here. > > I guess this is about weird, superflous quoting in “ls”: > https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=813164 > > Debian was dilligent enough to revert the patch (half a year ago). > Fedora – not so much. So... Systemd is gobbling up our beloved OS, Secure Boot is forcing us to use corporate distros, Gnome is trying to rule our world, and we're worried about a small change to the ls command that affects only those so silly as to allow spaces and non-dash, non-dot, non-underscore punctuation into a filename? When the dust finally settles, if necessary one of us can write a ls substitute that does the right thing, whatever that turns out to be. We have much bigger fish to fry right now. SteveT Steve Litt August 2016 featured book: Manager's Guide to Technical Troubleshooting Brand new, second edition http://www.troubleshooters.com/mgr ___ Dng mailing list Dng@lists.dyne.org https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng
Re: [DNG] vdev
Right. Perhaps the first caveat is that I trimmed down the installation (without DE etc and only leaving standard utilites) by purging as much as I thought sensible. Other than that, I made vdev installation as per README, then: a) changed /etc/vdev/vdevd.conf to in particular have logfile=/run/vdev/vdev.log (the /var directory is ro when it starts) b) add missing /etc/vdev/acl directory, and copied in the file 00-whitelist-root.acl from the example directory (the acl directory was missing and it just looked like a good idea to include that file; not sure why) c) [also for testing] fixed the paths in example/vdevd.conf for acls, actions, ifnames and hwdb, which pointed wrong. The hwdb in particular needs to be right for making the initramfs. I believe that's all. I'm planning to retry this on a DE installation shortly. regards, Ralph. On 08/08/16 04:35, shraptor wrote: FWIW, I know very little about Devuan package building, so I jumped straight to compiling the github source for vdev, and installing from it. After just a couple of tweaks of the runtime and initramfs configurations, I've made a successful replacement on a pristine Devuan 1.0.0 install, with the 3.16.0-4-amd64 kernel. Of course I don't know anything about how complete it is in respect of handling all sorts of devices. By any chance you could share procedure since I tried the same and failed? Not git and compiling of course but how you got it to cooperate with Devuan /scooby ___ Dng mailing list Dng@lists.dyne.org https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng ___ Dng mailing list Dng@lists.dyne.org https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng
Re: [DNG] libvirt - can't create snapshot
In my experience, there's no need to be root for creating and running libvirt VM's. I needed to be root only to set up the VDE host-only network for them to attach to, and then all else can be done non-root. Ralph. On 08/08/16 04:52, Paweł Cholewiński wrote: W dniu 07.08.2016 o 16:46, Ralph Ronnquist pisze: On 08/08/16 00:37, Paweł Cholewiński wrote: Hi, I test devuan with libvirt (Virtual Machine Manager 0.9.1) on Debian 7.11. I would like to create snapshot but when I type as root: 1) "virsh snapshot-create-as --domain devuan_jessie --name snapshot1 --description "clear system" error: failed to get domain 'devuan_jessie' error: An error occurred, but the cause is unknown". 2) "virsh list --all" command return empty list. Has anyone had this problem before? Only when the machine(s) were created as non-root. Ralph. Regards, Paweł ___ Dng mailing list Dng@lists.dyne.org https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng ___ Dng mailing list Dng@lists.dyne.org https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng Thanks Ralph. I read that I must add "--connect qemu:///system" parameter to be able to create snapshot. Regards Paweł ___ Dng mailing list Dng@lists.dyne.org https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng ___ Dng mailing list Dng@lists.dyne.org https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng
Re: [DNG] Fearsome rumblings from GNU
On Sun, Aug 07, 2016 at 04:26:59PM -0400, Steve Litt wrote: > On Sun, 7 Aug 2016 20:00:17 +0200 > > I guess this is about weird, superflous quoting in “ls”: > > https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=813164 > > > > Debian was dilligent enough to revert the patch (half a year ago). > > Fedora – not so much. > > So... > > Systemd is gobbling up our beloved OS, Secure Boot is forcing us to use > corporate distros, Gnome is trying to rule our world, and we're worried > about a small change to the ls command that affects only those so silly > as to allow spaces and non-dash, non-dot, non-underscore punctuation > into a filename? If your needs are so small they're served adequately by restricting filenames to not even the set of basic ASCII, that's you. As for me, I prefer media files to have Unicode, spaces and apostrophes in their names, and if it breaks a tool, it's a bug in the tool. And regressing such a basic tool as "ls" is not acceptable. > When the dust finally settles, if necessary one of us can write a ls > substitute that does the right thing, whatever that turns out to be. I don't accept init getting broken, neither do I accept that with ls. -- An imaginary friend squared is a real enemy. ___ Dng mailing list Dng@lists.dyne.org https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng
Re: [DNG] vdev
.. correction: I wrote /etc/vdev/acl but meant /etc/vdev/acls On 08/08/16 07:25, Ralph Ronnquist wrote: Right. Perhaps the first caveat is that I trimmed down the installation (without DE etc and only leaving standard utilites) by purging as much as I thought sensible. Other than that, I made vdev installation as per README, then: a) changed /etc/vdev/vdevd.conf to in particular have logfile=/run/vdev/vdev.log (the /var directory is ro when it starts) b) add missing /etc/vdev/acl directory, and copied in the file 00-whitelist-root.acl from the example directory (the acl directory was missing and it just looked like a good idea to include that file; not sure why) c) [also for testing] fixed the paths in example/vdevd.conf for acls, actions, ifnames and hwdb, which pointed wrong. The hwdb in particular needs to be right for making the initramfs. I believe that's all. I'm planning to retry this on a DE installation shortly. regards, Ralph. On 08/08/16 04:35, shraptor wrote: FWIW, I know very little about Devuan package building, so I jumped straight to compiling the github source for vdev, and installing from it. After just a couple of tweaks of the runtime and initramfs configurations, I've made a successful replacement on a pristine Devuan 1.0.0 install, with the 3.16.0-4-amd64 kernel. Of course I don't know anything about how complete it is in respect of handling all sorts of devices. By any chance you could share procedure since I tried the same and failed? Not git and compiling of course but how you got it to cooperate with Devuan /scooby ___ Dng mailing list Dng@lists.dyne.org https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng -- Chief Scientist Realthing Entertainment Pty Ltd ___ Dng mailing list Dng@lists.dyne.org https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng
Re: [DNG] SystemD's brownie points over non-systemd OSs.
On 08/08/16 00:31, Edward Bartolo wrote: Hi All, http://forums.debian.net/viewtopic.php?f=16&p=621842#p621842 Excuse me for the topic title. But the above link at first looked like some inherent advantage in using SystemD. However, after a little reflection, a couple of minutes, it seems there are actually no extra brownie-points in using SystemD. This *appears* like an advantage SystemD users have over non-systemd users. But if Devuan allows the use of virtualisation software, the same can be achieved without requiring any sort of benediction SystemD. Unless I'm mistaken, it looks like someone re-invented a more complicated chroot. ___ Dng mailing list Dng@lists.dyne.org https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng
Re: [DNG] SystemD's brownie points over non-systemd OSs.
On Sun, Aug 07, 2016 at 06:31:10PM +0200, Edward Bartolo wrote: > http://forums.debian.net/viewtopic.php?f=16&p=621842#p621842 > > Excuse me for the topic title. But the above link at first looked like > some inherent advantage in using SystemD. However, after a little > reflection, a couple of minutes, it seems there are actually no extra > brownie-points in using SystemD. > > This *appears* like an advantage SystemD users have over non-systemd > users. But if Devuan allows the use of virtualisation software, the > same can be achieved without requiring any sort of benediction > SystemD. > > virtualbox is the Devuan 64 bit repository. So, the above is not an > advantage but another way to make a cup of tea rather than the known > ways. virtualbox is _not_ an equivalent. What virtualbox and qemu-kvm, or proprietary vmware and MS Hyper-V, do, is full machine virtualization. You put an entire operating system inside, with its own kernel, and it can be anything, even Windows or SCO ClosedServer if you fancy so. systemd-nspawn is a worse clone of lxc which in turn is a worse remake of vserver and openvz. These run using host's kernel. The difference between these two is mostly in efficiency and memory use. While for most devices (network, disk) the cost of full virtualization isn't significant (but always noticeable), the memory needs are MASSIVE. With OS-level virtualization, it costs you only for processes that are running. That's around 100KB for init (obvious snide skipped), 1.1MB for rsyslogd, perhaps 380KB for sshd if you dislike "vserver exec" or "lxc-attach" and that's it. Anything more are daemons that do productive work. It's not uncommon to put 400ish vservers on a single physical machine in production[1], or tens of thousands to prove a point. And on a load spike, any vserver can take most of the machine's memory and resources (of course, if others stay mostly quiet at that time), which works wonders if you do maintenance serially. On the other hand, full-machine virtualization costs you the max of assigned memory to that system, at all time. systemd-nspawn, lxc and docker[2] are built upon chroot+unshare+cgroups+ seccomp. The purposes of those are: * chroot: separating the filesystem * unshare: separating namespaces. Of note are the mount namespace (a container can have mounts of its own), hostname, network (containers have their own IPs, routing tables, etc), user (you get to be root inside yet can't break the rest) * cgroups: resource limiting: caps and fair share of CPU/memory/IO/etc * seccomp: syscalls which could harm the rest of the system are vetoed You can try playing with those on your own. Especially "unshare -n" (read the manpage) is fun! But for a whole package, use lxc. It will configure all of the above for you. systemd-nspawn is merely a NIH copy of it. [1]. Depends on what your customers do, obviously. [2]. Docker is mostly about what you put _inside_ the container, but can manage them on its own. -- An imaginary friend squared is a real enemy. ___ Dng mailing list Dng@lists.dyne.org https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng
Re: [DNG] SystemD's brownie points over non-systemd OSs.
> Sun, Aug 07, 2016 at 06:31:10PM +0200, Edward Bartolo wrote: > > On the other hand, full-machine virtualization costs you the max of > assigned > memory to that system, at all time. > Going off-topic but I just wanted to correct this statement, when you're using full machine virtualization in the worst case it will use all the memory assigned to the server (unless you're using xen). Products like vmware or hyper-v uses in memory deduplication and swapping the virtual machine files to the disk to reduce the memory usage. -- aldemir ___ Dng mailing list Dng@lists.dyne.org https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng
Re: [DNG] SystemD's brownie points over non-systemd OSs.
On 08/08/16 13:44, Aldemir Akpinar wrote: Sun, Aug 07, 2016 at 06:31:10PM +0200, Edward Bartolo wrote: On the other hand, full-machine virtualization costs you the max of assigned memory to that system, at all time. Going off-topic but I just wanted to correct this statement, when you're using full machine virtualization in the worst case it will use all the memory assigned to the server (unless you're using xen). Products like vmware or hyper-v uses in memory deduplication and swapping the virtual machine files to the disk to reduce the memory usage. Just for the record, so does KVM if you turn on ksm. It actually works very well for a free solution. This is a tiny array of linux & windows VMs (5 linux / 3 windows). ksm effectively saves half the memory footprint. root@srv:~# ./ksmstat2 Shared memory is 488 MB Saved memory is 6557 MB Shared pages usage ratio is 13.43 Unshared pages usage ratio is .95 Sure, not as efficient as any of the containers, but then the containers can't run a blackberry BES or any of the other horrid windows software you sometimes have to run as part of day to day business life catering for others. ___ Dng mailing list Dng@lists.dyne.org https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng
Re: [DNG] SystemD's brownie points over non-systemd OSs.
On Mon, Aug 08, 2016 at 08:44:39AM +0300, Aldemir Akpinar wrote: > > Sun, Aug 07, 2016 at 06:31:10PM +0200, Edward Bartolo wrote: > > > > On the other hand, full-machine virtualization costs you the max of > > assigned > > memory to that system, at all time. > > Going off-topic but I just wanted to correct this statement, when you're > using full machine virtualization in the worst case it will use all the > memory assigned to the server (unless you're using xen). Products like > vmware or hyper-v uses in memory deduplication and swapping the virtual > machine files to the disk to reduce the memory usage. qemu-kvm (but not virtualbox) has support for KSM which, at a noticeable CPU cost, deduplicates anonymous pages (on kvm all are anonymous) after a delay. Both qemu-kvm and virtualbox can swap, too, but in that case you have double swapping which means performance will be beyond terrible. On the other hand, processes which mmap the same file will share the pages at no cost or delay whatsoever. And under memory pressure, the system can drop those frames without pointlessly swapping them to the disk. -- An imaginary friend squared is a real enemy. ___ Dng mailing list Dng@lists.dyne.org https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng
[DNG] KVM/ksm: was SystemD's brownie points over non-systemd OSs.
On Mon, 8 Aug 2016 14:00:48 +0800 Brad Campbell wrote: > Just for the record, so does KVM if you turn on ksm. It actually > works very well for a free solution. I use Qemu all the time, with hardware assist. How would I turn on ksm? SteveT Steve Litt August 2016 featured book: Manager's Guide to Technical Troubleshooting Brand new, second edition http://www.troubleshooters.com/mgr ___ Dng mailing list Dng@lists.dyne.org https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng
Re: [DNG] KVM/ksm: was SystemD's brownie points over non-systemd OSs.
On 08/08/16 14:17, Steve Litt wrote: On Mon, 8 Aug 2016 14:00:48 +0800 Brad Campbell wrote: Just for the record, so does KVM if you turn on ksm. It actually works very well for a free solution. I use Qemu all the time, with hardware assist. How would I turn on ksm? KSM=/sys/kernel/mm/ksm echo 500 > $KSM/sleep_millisecs echo 1000 > $KSM/pages_to_scan echo 1 > $KSM/run I do it by putting this in a script that gets run at bootup. My numbers are relatively arbitrary and are designed to work slowly with a low load on the system. Redhat published a daemon to do it called ksmtuned, but for my use case it just made things more complex for no real gain. ___ Dng mailing list Dng@lists.dyne.org https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng
