Re: Upstream Default (FOSS) DDX Driver for NVidia GPUs is not Nouveau
On 2021-01-20 at 21:17, Felix Miata wrote: > The Wanderer composed on 2021-01-20 19:44 (UTC-0500): > >> Felix Miata wrote: > >>> Sven Joachim composed on 2021-01-20 23:18 (UTC+0100): > the upstream Xorg server still uses a graphics card specific driver if available. > >>> That was exactly one of my points. "If available" equates to they >>> are optional, not needed by primarily AMD, Intel and NVidia >>> graphics devices made since somewhere around 2010. > >> This looks like another case (which I hadn't noticed before) of a >> difference in possible meanings of "default". > >> Some people may read "the default DDX driver" to mean "the DDX >> driver which will be used unless something tells it to do >> otherwise". I'm guessing this is how Sven read it. > > Above was mine. In that case I'm not sure how this fits with your statement about "if available", above. >> Other people may read it to mean "the DDX driver which will be used >> in the default case where no better one is available". I'm guessing >> this is how you intended it. > > Others might, but not me as worded. The crutch word there is > "better". Acknowledged; discard that. Try "where no alternative is available" instead. (I couldn't think of preferable wording at the time of posting, but the latter has come to me since.) > To me in this context default means the one that will be used, if > supported, when no /unneeded/ and thus optional DDX drivers drivers > are installed, aka none of xf86-video-* are installed. In other words, "in the default case where no alternative is available". I used "better" originally because presumably the system would not choose the other over this, when the other is available, if it did not expect the other to provide better results. > Because most mainstream hardware of the day, and since, is supported, > none of the optionals are /needed/ (as long as a fully-functional > DRM/KMS kernel module exists). The installation of the optional > drivers equates to something telling it to do otherwise. I don't read the action of installing an optional driver as telling the system to use that driver. Installing a driver merely makes it available, it does not explicitly tell the system to use it; telling the system to use that driver would require something like specifying which driver to use in a configuration file. >>> Since a year or so of server 1.17.0 appearance, the GPU-specific >>> DDX drivers spend a bare minimum of time installed on the vast >>> majority of my PCs. Other than AMDGPU, the others occasionally >>> get installed only temporarily for quick testing and then are >>> removed. > >> One thing which I haven't been able to make out yet in reading >> about these modesetting drivers: how well do they work for 3D >> performance? Or for that matter, high-end 2D, such as you get with >> Dwarf Fortress? > > I have no experience with, or interest in, any sort of gaming, > computer or otherwise. 3D performance does not always mean gaming (though I'll admit that's the overwhelmingly most common case for it I can think of), but fair enough. That does mean that your recommendations don't necessarily generalize to what many other people are going to want, however, so it would be worth making clear when you advocate for the modesetting drivers only that you can't testify to whether they work for 3D performance. >> The only positive statements I've found so far, that I recall, are >> that they work well for 2D acceleration on at least some (if not >> most) GPU models. > >> If you're using them near-exclusively on most of your PCs, that >> could imply that they work fine for 3D etc., or simply that you >> don't tend to need those things on those PCs. > > The latter. To me, 2D and 3D have no relevant meaning in a computing > context. I can't tell where 2D stops and 3D starts. 3D only simulates > 3 dimensions. A screen only displays 2 dimensions. What's the fuss? A 2D representation of a 3D space is much harder to calculate and render on the fly than is a 2D representation of a 2D space, and so requires much more computing power. If you happen to want the former, and your system is only set up to be able to provide the latter, you may wind up with an unpleasant low-performance surprise one day. > When I want TV I use a TV and remote controller and an easy chair or > sofa, not a workchair and desk and rodent and keyboard and all the > distractions of multitasking. My computers are general purpose tools > for getting work done. Computers are tools for automation. What makes things interesting is the question of what one chooses to automate. > Most videos are primarily time devouring entertainers, amusement. > > a (as first letter as affects meaning of word): anti or not > > muse: v: to think or ruminate upon. > > thus: > > amuse: literally not think or unthink https://www.etymonline.com/word/amuse states that this is the "to, at" prefix from Latin in a sense of "causing to", not the "not, wi
Re: installing debian 10 without a cd and without usb but could use ethernet
On Wed 20 Jan 2021 at 20:31:53 -0800, Dan Hitt wrote:
[...]
> menuentry "debian-10-iso" {
> set isofile="/USER/iso/debian-10.7.0-amd64-xfce-CD-1.iso"
> loopback loop (hd0,gptNN)$isofile
> linux (loop)/install.amd/vmlinuz boot=install.amd
> iso-scan/filename=$isofile noprompt noeject
> initrd (loop)/install.amd/initrd.gz
> }
This technique is doomed to failure. debian-10.7.0-amd64-xfce-CD-1.iso
needs to be mounted when it is found. However, the installer's initrd
does not contain a loop module, so this is not possible.
David Wright's advice to use the hd-media kernel and initrd is your way
forward. The simplest GRUB stanza possible is
menuentry 'Debian 10' {
linux /boot/vmlinuz
initrd /boot/initrd.gz
}
--
Brian.
Re: Upstream Default (FOSS) DDX Driver for NVidia GPUs is not Nouveau
> https://www.etymonline.com/word/amuse states that this is the "to, at" > prefix from Latin in a sense of "causing to", not the "not, without" one > from Greek. Funnily [oops, I guess I meant "amusingly"] enough, the two end up meaning kind of the same: in that what happens is that attention is diverted, so when you "amuse", you cause thinking of something which prevents thinking of what you'd have been thinking of if it weren't for that amusement. Stefan
Re: new harddrive degraded speed
On 1/20/21 11:41 PM, Stefan Monnier wrote: i bought a new harddrive western digital red plus 6tb === START OF INFORMATION SECTION === Model Family: Western Digital Red Device Model: WDC WD60EFRX-68L0BN1 Serial Number:WD-WX62D60C9L6X LU WWN Device Id: 5 0014ee 212e6b9ad Firmware Version: 82.00A82 User Capacity:6,001,175,126,016 bytes [6.00 TB] Sector Sizes: 512 bytes logical, 4096 bytes physical Rotation Rate:5700 rpm Device is:In smartctl database [for details use: -P show] ATA Version is: ACS-2, ACS-3 T13/2161-D revision 3b SATA Version is: SATA 3.1, 6.0 Gb/s (current: 3.0 Gb/s) Local Time is:Wed Jan 20 17:12:52 2021 CST SMART support is: Available - device has SMART capability. SMART support is: Enabled why is the speed 3.0 gbs i have 4 other drive i have tested with this device and they all use 6.0 gbs I can't tell you why, but I'd recommend you look at dmesg | grep -i ata It might give you some clues, Stefan Is it an SMR drive? Interface speed may not be too critical if it is. Mark
Re: new harddrive degraded speed
Mark Allums wrote: > Is it an SMR drive? Interface speed may not be too critical if it is. WD says it is a CMR drive, and that they are separating the Red designations to make it clear to consumers which are SMR and which are CMR because people don't like buying SMR by accident, which surprises WD for some reason. In any case, the drive can't saturate 3 or 6 Gb/s, and would probably not be noticeably different with a SATA-1 1.5Gb/s connection. 135 MB/s seems to be a normal transfer rate for it. -dsr-
Re: installing debian 10 without a cd and without usb but could use ethernet
On Thu, Jan 21, 2021 at 3:33 AM Brian wrote:
> On Wed 20 Jan 2021 at 20:31:53 -0800, Dan Hitt wrote:
>
> [...]
>
> > menuentry "debian-10-iso" {
> > set isofile="/USER/iso/debian-10.7.0-amd64-xfce-CD-1.iso"
> > loopback loop (hd0,gptNN)$isofile
> > linux (loop)/install.amd/vmlinuz boot=install.amd
> > iso-scan/filename=$isofile noprompt noeject
> > initrd (loop)/install.amd/initrd.gz
> > }
>
> This technique is doomed to failure. debian-10.7.0-amd64-xfce-CD-1.iso
> needs to be mounted when it is found. However, the installer's initrd
> does not contain a loop module, so this is not possible.
>
> David Wright's advice to use the hd-media kernel and initrd is your way
> forward. The simplest GRUB stanza possible is
>
>menuentry 'Debian 10' {
>linux /boot/vmlinuz
>initrd /boot/initrd.gz
>}
>
> --
> Brian.
>
>
Brian, thanks so much for your advice. Thank you also Felix, David, and
Bastien --- i need to study what you have all written.
However, Brian's final stanza is so simple that i can ask a question about
it immediately.
And that is: how can grub2 or any other software know what partition
'/boot' refers to?
So i presume that in this very very short stanza you provide, there will
also have to be a search line like David has (search --no-floppy ..) to
identify just where '/boot' is (???).
Thanks everybody for your help!!!
dan
Re: new harddrive degraded speed
On 21.01.2021 05:36, brainf...@posteo.net wrote: i guess i'm asking what decides how a device is going to connect Both host controller and device controller are negotiating for a highest possible bandwidth. If a host controller supports speeds up to 3.0Gb only, then device controller will operate at 3.0Gb, and it doesn't matter if the device supports faster speeds. Usually, SATA connectors on motherboards are separated by capability and could be marked by different colors. Some of them support up to 3.0Gb only and others support up to 6.0Gb. Read a manual for your motherboard to find out, if there are any 6.0Gb connectors and where are they located. -- With kindest regards, Alexander. ⢀⣴⠾⠻⢶⣦⠀ ⣾⠁⢠⠒⠀⣿⡁ Debian - The universal operating system ⢿⡄⠘⠷⠚⠋⠀ https://www.debian.org ⠈⠳⣄
Re: installing debian 10 without a cd and without usb but could use ethernet
Dan Hitt wrote:
> On Thu, Jan 21, 2021 at 3:33 AM Brian wrote:
>
> >
> >menuentry 'Debian 10' {
> >linux /boot/vmlinuz
> >initrd /boot/initrd.gz
> >}
> >
> > --
> > Brian.
> >
> >
> Brian, thanks so much for your advice. Thank you also Felix, David, and
> Bastien --- i need to study what you have all written.
>
> However, Brian's final stanza is so simple that i can ask a question about
> it immediately.
>
> And that is: how can grub2 or any other software know what partition
> '/boot' refers to?
>
> So i presume that in this very very short stanza you provide, there will
> also have to be a search line like David has (search --no-floppy ..) to
> identify just where '/boot' is (???).
https://en.wikipedia.org/wiki/GNU_GRUB#Startup_on_systems_using_BIOS_firmware
stage 2: core.img loads /boot/grub/i386-pc/normal.mod from the
partition configured by grub-install. If the partition index has
changed, GRUB will be unable to find the normal.mod, and
presents the user with the GRUB Rescue prompt.
So the answer to your question is, it's been configured at
install time, not discovered at runtime.
-dsr-
Re: Upstream Default (FOSS) DDX Driver for NVidia GPUs is not Nouveau
On Jo, 21 ian 21, 06:30:16, The Wanderer wrote: > On 2021-01-20 at 21:17, Felix Miata wrote: > > > The Wanderer composed on 2021-01-20 19:44 (UTC-0500): > > > >> Felix Miata wrote: > > > >>> Sven Joachim composed on 2021-01-20 23:18 (UTC+0100): > > > the upstream Xorg server still uses a graphics card specific > driver if available. > > > >>> That was exactly one of my points. "If available" equates to they > >>> are optional, not needed by primarily AMD, Intel and NVidia > >>> graphics devices made since somewhere around 2010. > > > >> This looks like another case (which I hadn't noticed before) of a > >> difference in possible meanings of "default". > > > >> Some people may read "the default DDX driver" to mean "the DDX > >> driver which will be used unless something tells it to do > >> otherwise". I'm guessing this is how Sven read it. > > > > Above was mine. > > In that case I'm not sure how this fits with your statement about "if > available", above. > > >> Other people may read it to mean "the DDX driver which will be used > >> in the default case where no better one is available". I'm guessing > >> this is how you intended it. > > > > Others might, but not me as worded. The crutch word there is > > "better". > > Acknowledged; discard that. Try "where no alternative is available" > instead. (I couldn't think of preferable wording at the time of posting, > but the latter has come to me since.) > > > To me in this context default means the one that will be used, if > > supported, when no /unneeded/ and thus optional DDX drivers drivers > > are installed, aka none of xf86-video-* are installed. > > In other words, "in the default case where no alternative is available". > > I used "better" originally because presumably the system would not > choose the other over this, when the other is available, if it did not > expect the other to provide better results. > > > Because most mainstream hardware of the day, and since, is supported, > > none of the optionals are /needed/ (as long as a fully-functional > > DRM/KMS kernel module exists). The installation of the optional > > drivers equates to something telling it to do otherwise. > > I don't read the action of installing an optional driver as telling the > system to use that driver. Installing a driver merely makes it > available, it does not explicitly tell the system to use it; telling the > system to use that driver would require something like specifying which > driver to use in a configuration file. In my opinion installing an *optional* driver to enable it makes for a nice user experience (as opposed to having to additionally edit configuration files). It's unfortunate that all drivers are installed by default (as in a standard Debian installation with a Desktop Environment) in order to ensure maximum compatibility. > > The latter. To me, 2D and 3D have no relevant meaning in a computing > > context. I can't tell where 2D stops and 3D starts. 3D only simulates > > 3 dimensions. A screen only displays 2 dimensions. What's the fuss? > > A 2D representation of a 3D space is much harder to calculate and render > on the fly than is a 2D representation of a 2D space, and so requires > much more computing power. If you happen to want the former, and your > system is only set up to be able to provide the latter, you may wind up > with an unpleasant low-performance surprise one day. As far as I understand the "3D" engines of GPUs are now being used also to render traditional 2D applications (i.e. without even trying to simulate three dimensions), so yes, at least some basic "3D" support and performance is important. Kind regards, Andrei -- http://wiki.debian.org/FAQsFromDebianUser signature.asc Description: PGP signature
Re: new harddrive degraded speed
On Jo, 21 ian 21, 12:09:12, Dan Ritter wrote: > Mark Allums wrote: > > Is it an SMR drive? Interface speed may not be too critical if it is. > > WD says it is a CMR drive, and that they are separating the Red > designations to make it clear to consumers which are SMR and > which are CMR because people don't like buying SMR by accident, > which surprises WD for some reason. According to [1] the newer WD Red drives are SMR, the CMR drives are now the Red Pro series. [1] https://www.truenas.com/community/resources/list-of-known-smr-drives.141/ Kind regards, Andrei -- http://wiki.debian.org/FAQsFromDebianUser signature.asc Description: PGP signature
Re: Upstream Default (FOSS) DDX Driver for NVidia GPUs is not Nouveau
On 2021-01-21 at 13:35, Andrei POPESCU wrote: > On Jo, 21 ian 21, 06:30:16, The Wanderer wrote: > >> On 2021-01-20 at 21:17, Felix Miata wrote: >>> Because most mainstream hardware of the day, and since, is >>> supported, none of the optionals are /needed/ (as long as a >>> fully-functional DRM/KMS kernel module exists). The installation >>> of the optional drivers equates to something telling it to do >>> otherwise. >> >> I don't read the action of installing an optional driver as telling >> the system to use that driver. Installing a driver merely makes it >> available, it does not explicitly tell the system to use it; >> telling the system to use that driver would require something like >> specifying which driver to use in a configuration file. > > In my opinion installing an *optional* driver to enable it makes for > a nice user experience (as opposed to having to additionally edit > configuration files). I don't necessarily disagree, although there can be difficulties in practice with some use cases. My point was more musing about (differences in people's) definitions of "default". In one approach, the system defaults to using the optional driver if it is available, and falls back to a simpler one if it is not. In another approach, the system defaults to using the simpler driver even when the optional one is available, unless something has told it to use the optional driver. In the latter case, the simpler driver is the default by every definition that's occurring to me offhand. In the former case, by some definitions the simpler driver is the default, and by the other definition it is not. I think that's the case we're looking at in this instance. As a result of these multiple possible definitions being able to apply, the use of the bare term "default" can be confusing and result in people potentially talking past one another, which I've seen happen too many times in the past. (That being part of why I have such an interest in the multiple meanings of "default" that a manifestation of a new one grabs my attention.) >>> The latter. To me, 2D and 3D have no relevant meaning in a >>> computing context. I can't tell where 2D stops and 3D starts. 3D >>> only simulates 3 dimensions. A screen only displays 2 dimensions. >>> What's the fuss? >> >> A 2D representation of a 3D space is much harder to calculate and >> render on the fly than is a 2D representation of a 2D space, and so >> requires much more computing power. If you happen to want the >> former, and your system is only set up to be able to provide the >> latter, you may wind up with an unpleasant low-performance surprise >> one day. > > As far as I understand the "3D" engines of GPUs are now being used > also to render traditional 2D applications (i.e. without even trying > to simulate three dimensions), so yes, at least some basic "3D" > support and performance is important. I believe I recall this myself, but I wasn't sufficiently certain of it to declare it without doing digging which I did not at that point have time for. IIRC, Dwarf Fortress is one example of this being done, although it's irrelevant to Felix because it's still gaming (in which he's said he's not interested). -- The Wanderer The reasonable man adapts himself to the world; the unreasonable one persists in trying to adapt the world to himself. Therefore all progress depends on the unreasonable man. -- George Bernard Shaw signature.asc Description: OpenPGP digital signature
Re: installing debian 10 without a cd and without usb but could use ethernet
On Mi, 20 ian 21, 20:31:53, Dan Hitt wrote: > > Or perhaps there's some other way to approach the problem? For example, > i've already created a partition to hold the debian system i want to put on > the machine. Is there some way of hand-populating it? Yes, either debootstrap (used also by the Debian Installer) or mmdebstrap can create a "blank" Debian installation. Please mind that it will need additional configuration in order to make it work, including but not limited to creating users, setting up your fstab (in case you more partitions), etc. that are normally done by the installer. > I do have a running > debian 10 system on another machine, and i suppose i could tar it up and > unpack it into the new partition on the mint machine. Cloning another installation presents its own challenges, as you might end with stuff duplicated that should be unique for each system (unless you really know what you are doing). It's probably easier to just let debootstrap/mmdebstrap take care of the bulk installation and use the other Debian installation only as a source of configuration files. > But i'm not sure if > there's something outside the filesystem but inside the partition which is > necessary for it to be bootable. It depends. With GRUB on traditional BIOS (non-UEFI) systems there are several approaches possible, depending on which (if any) is your "primary" system (i.e. the one that provides/controls the boot loader). More details are needed here to provide useful recommendations. Kind regards, Andrei -- http://wiki.debian.org/FAQsFromDebianUser signature.asc Description: PGP signature
Re: installing debian 10 without a cd and without usb but could use ethernet
On Thu, Jan 21, 2021 at 10:02 AM Dan Ritter wrote:
> Dan Hitt wrote:
> > On Thu, Jan 21, 2021 at 3:33 AM Brian wrote:
> >
> > >
> > >menuentry 'Debian 10' {
> > >linux /boot/vmlinuz
> > >initrd /boot/initrd.gz
> > >}
> > >
> > > --
> > > Brian.
> > >
> > >
> > Brian, thanks so much for your advice. Thank you also Felix, David, and
> > Bastien --- i need to study what you have all written.
> >
> > However, Brian's final stanza is so simple that i can ask a question
> about
> > it immediately.
> >
> > And that is: how can grub2 or any other software know what partition
> > '/boot' refers to?
> >
> > So i presume that in this very very short stanza you provide, there will
> > also have to be a search line like David has (search --no-floppy ..)
> to
> > identify just where '/boot' is (???).
>
>
> https://en.wikipedia.org/wiki/GNU_GRUB#Startup_on_systems_using_BIOS_firmware
>
> stage 2: core.img loads /boot/grub/i386-pc/normal.mod from the
> partition configured by grub-install. If the partition index has
> changed, GRUB will be unable to find the normal.mod, and
> presents the user with the GRUB Rescue prompt.
>
> So the answer to your question is, it's been configured at
> install time, not discovered at runtime.
>
> -dsr-
>
Thanks Dan for your mail, and for the reference to the wikipedia article.
When you say 'configured at install time', does that refer to the time at
which i run 'sudo update-grub' (on my mint host)?
(I presume that it is impossible that this refers to the time when grub
itself was last installed on the box, several years ago.)
Anyhow, i added an entry to /etc/grub.d just to see what would happen if i
took the simple menu entry quite literally:
menuentry "simple-test" {
linux /boot/vmlinuz
initrd /boot/initrd.gz
}
I ran 'sudo update-grub', and the entry was copied into /boot/grub/grub.cfg
without modification. And then i tried booting into it, just to see what
grub would do. And, it did what i think was the only thing it possibly
could: it reported:
error: file `/boot/vmlinuz' not found.
error: you need to load the kernel first.
Press any key to continue...
Now, Brian said that "the installer's initrd does not contain a loop
module", so that would indicate that if i want to use
debian-10.7.0-amd64-xfce-CD-1.iso, i'll need to get it on the disk
(presumably by just unpacking it somewhere --- prior to booting, i can loop
mount it and copy it to a 'real' directory), and then modifying
/boot/vmlinuz and /boot/initrd.gz to be paths that grub understands. Or,
maybe the debootstrap method Bastien suggests would be good.
Anyhow, thanks for your message, and thanks everybody else for these
important pieces of knowledge that i need to learn.
dan
Re: new harddrive degraded speed
On 2021-01-21 12:44 PM, Andrei POPESCU wrote: On Jo, 21 ian 21, 12:09:12, Dan Ritter wrote: Mark Allums wrote: > Is it an SMR drive? Interface speed may not be too critical if it is. WD says it is a CMR drive, and that they are separating the Red designations to make it clear to consumers which are SMR and which are CMR because people don't like buying SMR by accident, which surprises WD for some reason. According to [1] the newer WD Red drives are SMR, the CMR drives are now the Red Pro series. [1] https://www.truenas.com/community/resources/list-of-known-smr-drives.141/ what i have is wd red plus
Re: installing debian 10 without a cd and without usb but could use ethernet
Dan Hitt wrote:
> On Thu, Jan 21, 2021 at 10:02 AM Dan Ritter wrote:
> > > also have to be a search line like David has (search --no-floppy ..)
> > to
> > > identify just where '/boot' is (???).
> >
> >
> > https://en.wikipedia.org/wiki/GNU_GRUB#Startup_on_systems_using_BIOS_firmware
> >
> > stage 2: core.img loads /boot/grub/i386-pc/normal.mod from the
> > partition configured by grub-install. If the partition index has
> > changed, GRUB will be unable to find the normal.mod, and
> > presents the user with the GRUB Rescue prompt.
> >
> > So the answer to your question is, it's been configured at
> > install time, not discovered at runtime.
> >
> > -dsr-
> >
>
> Thanks Dan for your mail, and for the reference to the wikipedia article.
>
> When you say 'configured at install time', does that refer to the time at
> which i run 'sudo update-grub' (on my mint host)?
That's a possible install time. Another possible install time is
when the creator of an ISO image - say, the netinst.iso - runs
update-grub on the bootloader for that image. (Somebody with
more expertise in CD booting might be able to say more.)
Consider that your computer decides which disk to read the boot
sector from, and after that it's all up to what that boot sector
does.
> (I presume that it is impossible that this refers to the time when grub
> itself was last installed on the box, several years ago.)
>
> Anyhow, i added an entry to /etc/grub.d just to see what would happen if i
> took the simple menu entry quite literally:
> menuentry "simple-test" {
> linux /boot/vmlinuz
> initrd /boot/initrd.gz
> }
>
> I ran 'sudo update-grub', and the entry was copied into /boot/grub/grub.cfg
> without modification. And then i tried booting into it, just to see what
> grub would do. And, it did what i think was the only thing it possibly
> could: it reported:
> error: file `/boot/vmlinuz' not found.
> error: you need to load the kernel first.
>
> Press any key to continue...
>
> Now, Brian said that "the installer's initrd does not contain a loop
> module", so that would indicate that if i want to use
> debian-10.7.0-amd64-xfce-CD-1.iso, i'll need to get it on the disk
> (presumably by just unpacking it somewhere --- prior to booting, i can loop
> mount it and copy it to a 'real' directory), and then modifying
> /boot/vmlinuz and /boot/initrd.gz to be paths that grub understands. Or,
> maybe the debootstrap method Bastien suggests would be good.
Both should be workable.
-dsr-
Re: installing debian 10 without a cd and without usb but could use ethernet
On Mi, 20 ian 21, 20:31:53, Dan Hitt wrote: > I have a machine that currently has linux mint 16.04 on it. > > I would like to install debian 10 on it, but the installer really wants > access to a cd drive, and one just isn't available. Any kind of medium that can be used by your motherboard to boot should work (e.g. a spare SATA drive), same as you would be using a USB stick (including loosing all data when copying the .iso to it). If using the 'mini.iso' in the 'netboot' category you can even install to the same drive (the other images don't allow repartitioning of the drive (partition?) they are on). Kind regards, Andrei -- http://wiki.debian.org/FAQsFromDebianUser signature.asc Description: PGP signature
Re: new harddrive degraded speed
On Jo, 21 ian 21, 13:12:38, brainf...@posteo.net wrote: > > > On 2021-01-21 12:44 PM, Andrei POPESCU wrote: > > On Jo, 21 ian 21, 12:09:12, Dan Ritter wrote: > > > Mark Allums wrote: > > > > Is it an SMR drive? Interface speed may not be too critical if it is. > > > > > > WD says it is a CMR drive, and that they are separating the Red > > > designations to make it clear to consumers which are SMR and > > > which are CMR because people don't like buying SMR by accident, > > > which surprises WD for some reason. > > > > According to [1] the newer WD Red drives are SMR, the CMR drives are now > > the Red Pro series. > > > > [1] > > https://www.truenas.com/community/resources/list-of-known-smr-drives.141/ > > what i have is wd red plus The 5TB EFRX is WD Red though. Kind regards, Andrei -- http://wiki.debian.org/FAQsFromDebianUser signature.asc Description: PGP signature
Re: installing debian 10 without a cd and without usb but could use ethernet
On Thu 21 Jan 2021 at 09:34:56 -0800, Dan Hitt wrote:
> On Thu, Jan 21, 2021 at 3:33 AM Brian wrote:
>
> > On Wed 20 Jan 2021 at 20:31:53 -0800, Dan Hitt wrote:
> >
> > [...]
> >
> > > menuentry "debian-10-iso" {
> > > set isofile="/USER/iso/debian-10.7.0-amd64-xfce-CD-1.iso"
> > > loopback loop (hd0,gptNN)$isofile
> > > linux (loop)/install.amd/vmlinuz boot=install.amd
> > > iso-scan/filename=$isofile noprompt noeject
> > > initrd (loop)/install.amd/initrd.gz
> > > }
> >
> > This technique is doomed to failure. debian-10.7.0-amd64-xfce-CD-1.iso
> > needs to be mounted when it is found. However, the installer's initrd
> > does not contain a loop module, so this is not possible.
> >
> > David Wright's advice to use the hd-media kernel and initrd is your way
> > forward. The simplest GRUB stanza possible is
> >
> >menuentry 'Debian 10' {
> >linux /boot/vmlinuz
> >initrd /boot/initrd.gz
> >}
> >
> > --
> > Brian.
> >
> >
> Brian, thanks so much for your advice. Thank you also Felix, David, and
> Bastien --- i need to study what you have all written.
>
> However, Brian's final stanza is so simple that i can ask a question about
> it immediately.
>
> And that is: how can grub2 or any other software know what partition
> '/boot' refers to?
You said originally:
> I have a machine that currently has linux mint 16.04 on it
I assummed that that this was the only partition and that vmlinuz and
initrd.gz would go in /boot. How many partitions do you have?
> So i presume that in this very very short stanza you provide, there will
> also have to be a search line like David has (search --no-floppy ..) to
> identify just where '/boot' is (???).
With more than one partition: David's suggestion is more or less
obligatory.
--
Brian.
Re: installing debian 10 without a cd and without usb but could use ethernet
On Thu 21 Jan 2021 at 11:08:56 -0800, Dan Hitt wrote: [...] > Now, Brian said that "the installer's initrd does not contain a loop > module", so that would indicate that if i want to use > debian-10.7.0-amd64-xfce-CD-1.iso, i'll need to get it on the disk > (presumably by just unpacking it somewhere --- prior to booting, i can loop > mount it and copy it to a 'real' directory), and then modifying > /boot/vmlinuz and /boot/initrd.gz to be paths that grub understands. Or, > maybe the debootstrap method Bastien suggests would be good. You can manipulate debian-10.7.0-amd64-xfce-CD-1.iso in whatever way you want. It will still not be loop mounted by hd-media. You are heading off into the wild, blue yonder. -- Brian.
Re: installing debian 10 without a cd and without usb but could use ethernet
Le 21/01/2021 à 21:46, Brian a écrit : > On Thu 21 Jan 2021 at 11:08:56 -0800, Dan Hitt wrote: > > [...] > >> Now, Brian said that "the installer's initrd does not contain a loop >> module", so that would indicate that if i want to use >> debian-10.7.0-amd64-xfce-CD-1.iso, i'll need to get it on the disk >> (presumably by just unpacking it somewhere --- prior to booting, i can loop >> mount it and copy it to a 'real' directory), and then modifying >> /boot/vmlinuz and /boot/initrd.gz to be paths that grub understands. Or, >> maybe the debootstrap method Bastien suggests would be good. > You can manipulate debian-10.7.0-amd64-xfce-CD-1.iso in whatever way > you want. It will still not be loop mounted by hd-media. > > You are heading off into the wild, blue yonder. > A possibility would be to set up a PXE server, tehre are ways to do a netinstall with a PXE boot (boot on the network). There surely are tutorials on the net for this.
Re: installing debian 10 without a cd and without usb but could use ethernet
Le 21/01/2021 à 21:52, Erwan David a écrit : > Le 21/01/2021 à 21:46, Brian a écrit : >> On Thu 21 Jan 2021 at 11:08:56 -0800, Dan Hitt wrote: >> >> [...] >> >>> Now, Brian said that "the installer's initrd does not contain a loop >>> module", so that would indicate that if i want to use >>> debian-10.7.0-amd64-xfce-CD-1.iso, i'll need to get it on the disk >>> (presumably by just unpacking it somewhere --- prior to booting, i can loop >>> mount it and copy it to a 'real' directory), and then modifying >>> /boot/vmlinuz and /boot/initrd.gz to be paths that grub understands. Or, >>> maybe the debootstrap method Bastien suggests would be good. >> You can manipulate debian-10.7.0-amd64-xfce-CD-1.iso in whatever way >> you want. It will still not be loop mounted by hd-media. >> >> You are heading off into the wild, blue yonder. >> > A possibility would be to set up a PXE server, tehre are ways to do a > netinstall with a PXE boot (boot on the network). > > There surely are tutorials on the net for this. > > Most obvious : https://wiki.debian.org/PXEBootInstall
Re: new harddrive degraded speed
On 2021-01-20 15:21, brainf...@posteo.net wrote: i bought a new harddrive western digital red plus 6tb === START OF INFORMATION SECTION === Model Family: Western Digital Red Device Model: WDC WD60EFRX-68L0BN1 Serial Number: WD-WX62D60C9L6X LU WWN Device Id: 5 0014ee 212e6b9ad Firmware Version: 82.00A82 User Capacity: 6,001,175,126,016 bytes [6.00 TB] Sector Sizes: 512 bytes logical, 4096 bytes physical Rotation Rate: 5700 rpm Device is: In smartctl database [for details use: -P show] ATA Version is: ACS-2, ACS-3 T13/2161-D revision 3b SATA Version is: SATA 3.1, 6.0 Gb/s (current: 3.0 Gb/s) Local Time is: Wed Jan 20 17:12:52 2021 CST SMART support is: Available - device has SMART capability. SMART support is: Enabled why is the speed 3.0 gbs i have 4 other drive i have tested with this device and they all use 6.0 gbs On 2021-01-20 16:36, brainf...@posteo.net wrote: > this device is a odroid hc4 > it supports all my other drives a 6.0 gbs > i used a sata to usb adapter to connect this drive to my pc and it > connects at 6.0 gbs > i guess i'm asking what decides how a device is going to connect > is there a database of devices that the kernel uses > i should say on my pc - 4.19.0-12-amd64 #1 SMP Debian 4.19.152-1 > (2020-10-18) x86_64 > and the hc4 - 4.9.236-arm64 #1 SMP PREEMPT Sun Dec 20 21:57:12 CET 2020 > aarch64 STFW 'odroid hc4': https://ameridroid.com/products/odroid-hc4 The drive firmware and computer kernel negotiate the speed via device drivers and interface chips. Everything has tolerances. Sometimes, two devices can be within tolerances individually when connected to a test stand, but go out of tolerance(s) when connected to each other. I've seen this with memory modules and motherboards. As another reader suggested, look for error messages in dmesg(1) and/or /var/log. Post anything interesting. Does the drive connect at 3 Gbps in both bays? Do you have another identical drive and/or NAS that you can use for A/B testing? Can you return the drive or the NAS? David
Re: installing debian 10 without a cd and without usb but could use ethernet
On Thu, Jan 21, 2021 at 12:29 PM Brian wrote:
> On Thu 21 Jan 2021 at 09:34:56 -0800, Dan Hitt wrote:
>
> > On Thu, Jan 21, 2021 at 3:33 AM Brian wrote:
> >
> > > On Wed 20 Jan 2021 at 20:31:53 -0800, Dan Hitt wrote:
> > >
> > > [...]
> > >
> > > > menuentry "debian-10-iso" {
> > > > set isofile="/USER/iso/debian-10.7.0-amd64-xfce-CD-1.iso"
> > > > loopback loop (hd0,gptNN)$isofile
> > > > linux (loop)/install.amd/vmlinuz boot=install.amd
> > > > iso-scan/filename=$isofile noprompt noeject
> > > > initrd (loop)/install.amd/initrd.gz
> > > > }
> > >
> > > This technique is doomed to failure. debian-10.7.0-amd64-xfce-CD-1.iso
> > > needs to be mounted when it is found. However, the installer's initrd
> > > does not contain a loop module, so this is not possible.
> > >
> > > David Wright's advice to use the hd-media kernel and initrd is your way
> > > forward. The simplest GRUB stanza possible is
> > >
> > >menuentry 'Debian 10' {
> > >linux /boot/vmlinuz
> > >initrd /boot/initrd.gz
> > >}
> > >
> > > --
> > > Brian.
> > >
> > >
> > Brian, thanks so much for your advice. Thank you also Felix, David, and
> > Bastien --- i need to study what you have all written.
> >
> > However, Brian's final stanza is so simple that i can ask a question
> about
> > it immediately.
> >
> > And that is: how can grub2 or any other software know what partition
> > '/boot' refers to?
>
> You said originally:
>
> > I have a machine that currently has linux mint 16.04 on it
>
> I assummed that that this was the only partition and that vmlinuz and
> initrd.gz would go in /boot. How many partitions do you have?
>
Thanks Brian for your mail.
I didn't mean to imply that linux mint 16.04 was the only OS on the
machine, or that there was only one partition.
There are 30 partitions, although not all of them have an OS on them; i
have a partition for users so that when i update the box with a new OS, i
can have access to all the previous user accounts. That's what i intend to
do this time as well.
I haven't put a new OS on the box in the last 5 years, but i would like to
update it to debian 10 now; my general procedure in such cases is to add a
new partition for the OS, but not discard what i have already installed.
The only disadvantage to this is that operations that scan the entire disk
looking for OSes take longer.
The challenges i face this time are that the CD is not available, so that
everything has to be done through the hd. Further, i only have access to
the bios with great difficulty. If i could get into the bios, i could
change the boot order, and possibly boot from usb. However, it is very
difficult to interrupt the boot process successfully (by pressing F2),
because it is so fast. (I could conceivably even get a usb
But the hd is big, and i can make more partitions as needed and populate
them as desired.
Thanks again everybody for your help!
dan
>
> > So i presume that in this very very short stanza you provide, there will
> > also have to be a search line like David has (search --no-floppy ..)
> to
> > identify just where '/boot' is (???).
>
> With more than one partition: David's suggestion is more or less
> obligatory.
>
> --
> Brian.
>
>
Re: installing debian 10 without a cd and without usb but could use ethernet
Dan Hitt wrote: > On Thu, Jan 21, 2021 at 12:29 PM Brian wrote: > > > I have a machine that currently has linux mint 16.04 on it > > > > I assummed that that this was the only partition and that vmlinuz and > > initrd.gz would go in /boot. How many partitions do you have? > > > > Thanks Brian for your mail. > > I didn't mean to imply that linux mint 16.04 was the only OS on the > machine, or that there was only one partition. > > There are 30 partitions, although not all of them have an OS on them; i > have a partition for users so that when i update the box with a new OS, i > can have access to all the previous user accounts. That's what i intend to > do this time as well. Oh, debootstrap is definitely the way to go, then. Pick your partition, mkfs.ext4 a filesystem, mount it, and: debootstrap stable /PATH/TO/MOUNTPOINT http://deb.debian.org/debian/ then you can arrange booting via your existing grub menu or whatever. -dsr-
Re: new harddrive degraded speed
>> Is it an SMR drive? Interface speed may not be too critical if it is. > > WD says it is a CMR drive, and that they are separating the Red > designations to make it clear to consumers which are SMR and > which are CMR because people don't like buying SMR by accident, > which surprises WD for some reason. In any case, we're talking about the speed of the SATA connection, which seems unrelated to CMR-vs-SMR, especially since AFAIK the slowness of SMR only affects writes. Stefan
Re: installing debian 10 without a cd and without usb but could use ethernet
Brian composed on 2021-01-21 11:33 (UTC):
> David Wright's advice to use the hd-media kernel and initrd is your way
> forward. The simplest GRUB stanza possible is
>menuentry 'Debian 10' {
>linux /boot/vmlinuz
>initrd /boot/initrd.gz
>}
I'm having trouble thinking of how this could work. Grub needs a way to know
where
to find the correct partition/filesystem. My 2021-01-21 00:06 (UTC-0500)
response
included one example:
search --no-floppy --label --set=root p03res
It remains for the reader to determine the simplest possible example.
https://www.gnu.org/software/grub/manual/grub/grub.html
--
Evolution as taught in public schools, like religion,
is based on faith, not on science.
Team OS/2 ** Reg. Linux User #211409 ** a11y rocks!
Felix Miata *** http://fm.no-ip.com/
Re: installing debian 10 without a cd and without usb but could use ethernet
On Thu, Jan 21, 2021 at 12:58 PM Erwan David wrote: > Le 21/01/2021 à 21:52, Erwan David a écrit : > > Le 21/01/2021 à 21:46, Brian a écrit : > >> On Thu 21 Jan 2021 at 11:08:56 -0800, Dan Hitt wrote: > >> > >> [...] > >> > >>> Now, Brian said that "the installer's initrd does not contain a loop > >>> module", so that would indicate that if i want to use > >>> debian-10.7.0-amd64-xfce-CD-1.iso, i'll need to get it on the disk > >>> (presumably by just unpacking it somewhere --- prior to booting, i can > loop > >>> mount it and copy it to a 'real' directory), and then modifying > >>> /boot/vmlinuz and /boot/initrd.gz to be paths that grub understands. > Or, > >>> maybe the debootstrap method Bastien suggests would be good. > >> You can manipulate debian-10.7.0-amd64-xfce-CD-1.iso in whatever way > >> you want. It will still not be loop mounted by hd-media. > >> > >> You are heading off into the wild, blue yonder. > >> > > A possibility would be to set up a PXE server, tehre are ways to do a > > netinstall with a PXE boot (boot on the network). > > > > There surely are tutorials on the net for this. > > > > > Most obvious : https://wiki.debian.org/PXEBootInstall Thanks Erwin for your message and the link. The problem with pxeboot, is that i would have to get into the bios to change the boot order. The bios does have an option for pxe boot, but i cannot reliably get into the bios: after many tries i was only able to get into it a couple of times. So if i had a way to reliably force my way into the bios, it would be very valuable, whether for pxe boot or just anything else. Thanks again everybody for your help. dan
Re: installing debian 10 without a cd and without usb but could use ethernet
Dan Hitt composed on 2021-01-21 13:06 (UTC-0800): > ...i only have access to > the bios with great difficulty. If i could get into the bios, i could > change the boot order, and possibly boot from usb. However, it is very > difficult to interrupt the boot process successfully (by pressing F2), > because it is so fast I cannot recall having been in a BIOS where there was not a fast boot option enabled by default. To set yours off, boot with no bootable devices available and you should get a chance to get into the BIOS if it doesn't automatically present. If that doesn't work either, do a mechanical BIOS reset. All BIOS I've encountered will present an opportunity to enter when the clock is not set. -- Evolution as taught in public schools, like religion, is based on faith, not on science. Team OS/2 ** Reg. Linux User #211409 ** a11y rocks! Felix Miata *** http://fm.no-ip.com/
Re: new harddrive degraded speed
Am 21.01.21 um 22:01 schrieb David Christensen: > On 2021-01-20 15:21, brainf...@posteo.net wrote: >> i bought a new harddrive >> western digital red plus 6tb >> >> === START OF INFORMATION SECTION === >> Model Family: Western Digital Red >> Device Model: WDC WD60EFRX-68L0BN1 >> Serial Number: WD-WX62D60C9L6X >> LU WWN Device Id: 5 0014ee 212e6b9ad >> Firmware Version: 82.00A82 >> User Capacity: 6,001,175,126,016 bytes [6.00 TB] >> Sector Sizes: 512 bytes logical, 4096 bytes physical >> Rotation Rate: 5700 rpm >> Device is: In smartctl database [for details use: -P show] >> ATA Version is: ACS-2, ACS-3 T13/2161-D revision 3b >> SATA Version is: SATA 3.1, 6.0 Gb/s (current: 3.0 Gb/s) >> Local Time is: Wed Jan 20 17:12:52 2021 CST >> SMART support is: Available - device has SMART capability. >> SMART support is: Enabled >> >> why is the speed 3.0 gbs >> i have 4 other drive i have tested with this device and they all use >> 6.0 gbs >> 3.0 gbs is round about 300 MB/s, WD60EFRX is much slower ... don't pull out your hair
Re: installing debian 10 without a cd and without usb but could use ethernet
On Thu 21 Jan 2021 at 16:20:10 -0500, Felix Miata wrote:
> Brian composed on 2021-01-21 11:33 (UTC):
>
> > David Wright's advice to use the hd-media kernel and initrd is your way
> > forward. The simplest GRUB stanza possible is
>
> >menuentry 'Debian 10' {
> >linux /boot/vmlinuz
> >initrd /boot/initrd.gz
> >}
>
> I'm having trouble thinking of how this could work. Grub needs a way to know
> where
> to find the correct partition/filesystem. My 2021-01-21 00:06 (UTC-0500)
> response
> included one example:
>
> search --no-floppy --label --set=root p03res
>
> It remains for the reader to determine the simplest possible example.
> https://www.gnu.org/software/grub/manual/grub/grub.html
In the absence of being told the disk and partition, GRUB will use
(hd0,msdos1). The OP did not specify more than one disk and partition.
--
Brian.
Re: installing debian 10 without a cd and without usb but could use ethernet
On Thu 21 Jan 2021 at 13:06:25 -0800, Dan Hitt wrote:
> On Thu, Jan 21, 2021 at 12:29 PM Brian wrote:
>
> > On Thu 21 Jan 2021 at 09:34:56 -0800, Dan Hitt wrote:
> >
> > > On Thu, Jan 21, 2021 at 3:33 AM Brian wrote:
> > >
> > > > On Wed 20 Jan 2021 at 20:31:53 -0800, Dan Hitt wrote:
> > > >
> > > > [...]
> > > >
> > > > > menuentry "debian-10-iso" {
> > > > > set isofile="/USER/iso/debian-10.7.0-amd64-xfce-CD-1.iso"
> > > > > loopback loop (hd0,gptNN)$isofile
> > > > > linux (loop)/install.amd/vmlinuz boot=install.amd
> > > > > iso-scan/filename=$isofile noprompt noeject
> > > > > initrd (loop)/install.amd/initrd.gz
> > > > > }
> > > >
> > > > This technique is doomed to failure. debian-10.7.0-amd64-xfce-CD-1.iso
> > > > needs to be mounted when it is found. However, the installer's initrd
> > > > does not contain a loop module, so this is not possible.
> > > >
> > > > David Wright's advice to use the hd-media kernel and initrd is your way
> > > > forward. The simplest GRUB stanza possible is
> > > >
> > > >menuentry 'Debian 10' {
> > > >linux /boot/vmlinuz
> > > >initrd /boot/initrd.gz
> > > >}
> > > >
> > > > --
> > > > Brian.
> > > >
> > > >
> > > Brian, thanks so much for your advice. Thank you also Felix, David, and
> > > Bastien --- i need to study what you have all written.
> > >
> > > However, Brian's final stanza is so simple that i can ask a question
> > about
> > > it immediately.
> > >
> > > And that is: how can grub2 or any other software know what partition
> > > '/boot' refers to?
> >
> > You said originally:
> >
> > > I have a machine that currently has linux mint 16.04 on it
> >
> > I assummed that that this was the only partition and that vmlinuz and
> > initrd.gz would go in /boot. How many partitions do you have?
> >
>
> Thanks Brian for your mail.
>
> I didn't mean to imply that linux mint 16.04 was the only OS on the
> machine, or that there was only one partition.
But you did say
> I have a machine that currently has linux mint 16.04 on it.
That installation will be on (hdX,msdosY). So
menuentry 'Debian 10' {
linux (hdX,msdosY)/boot/vmlinuz
initrd (hdX,msdosY)/boot/initrd.gz
}
or use a "search" line.
--
Brian.
Re: installing debian 10 without a cd and without usb but could use ethernet
On 2021-01-20 20:31, Dan Hitt wrote:
I have a machine that currently has linux mint 16.04 on it.
I would like to install debian 10 on it, but the installer really
wants access to a cd drive, and one just isn't available.
However, the linux mint 16.04 system does have grub2 on it.
So it is possible for me to boot from an iso image stored in the
filesystem just like a regular file. It's just a matter of writing
a menu entry in /etc/grub/40_custom.
I know this works because i've booted into a live cd image of linux
mint 20.1 (using the filesystem, not a cdrom), and started an
installation process. I backed out of it, because i would like to
install debian 10, not just a later version of mint.
When i do the same thing with debian, it starts off ok, doing some
simple things like setting the language and the keyboard layout, but
then it complains that it cannot find a cd rom. This is true with
both the netinst image, as well as with a jigdo xfce image which i
think should have everything necessary and not need a cd rom. For
reference, my debian menu entry is
menuentry "debian-10-iso" { set
isofile="/USER/iso/debian-10.7.0-amd64-xfce-CD-1.iso" loopback loop
(hd0,gptNN)$isofile linux (loop)/install.amd/vmlinuz boot=install.amd
iso-scan/filename=$isofile noprompt noeject initrd
(loop)/install.amd/initrd.gz }
Here, USER is the user name in whose account the iso image is, and
gptNN stands for the particular partition where home is mounted for
the user USER.
So my first question is whether there's a better iso image i can
use, or if i can fix this up by giving more arguments to the linux
invocation line or something else in the menu entry.
The second question is whether there's a way, from grub (grub2,
actually), of dropping down to the bios. I imagine this is quite
impossible, but if i'm wrong, please let me know. The reason i
would like to do this is that it is very hard for me to interrupt the
boot process fast enough to get to the bios, and i've only managed to
do it once or twice after many tries. If i were in the bios, i might
be able to figure out if it could boot from usb, and i could set the
boot order to do this, and make a bootable usb version of netinst
(perhaps).
Or perhaps there's some other way to approach the problem? For
example, i've already created a partition to hold the debian system
i want to put on the machine. Is there some way of hand-populating
it? I do have a running debian 10 system on another machine, and i
suppose i could tar it up and unpack it into the new partition on
the mint machine. But i'm not sure if there's something outside the
filesystem but inside the partition which is necessary for it to be
bootable.
Thanks in advance for any advice!
On 2021-01-21 13:06, Dan Hitt wrote:
I didn't mean to imply that linux mint 16.04 was the only OS on the
machine, or that there was only one partition.
There are 30 partitions, although not all of them have an OS on
them; i have a partition for users so that when i update the box with
a new OS, i can have access to all the previous user accounts.
That's what i> intend to do this time as well.
I haven't put a new OS on the box in the last 5 years, but i would
like to update it to debian 10 now; my general procedure in such
cases is to add a new partition for the OS, but not discard what i
have already installed. The only disadvantage to this is that
operations that scan the entire disk looking for OSes take longer.
The challenges i face this time are that the CD is not available, so
that everything has to be done through the hd. Further, i only have
access to the bios with great difficulty. If i could get into the
bios, i could change the boot order, and possibly boot from usb.
However, it is very difficult to interrupt the boot process
successfully (by pressing F2), because it is so fast. (I could
conceivably even get a usb
But the hd is big, and i can make more partitions as needed and
populate them as desired.
On 2021-01-21 13:20, Dan Hitt wrote:
he problem with pxeboot, is that i would have to get into the bios to
change the boot order. The bios does have an option for pxe boot,
but i cannot reliably get into the bios: after many tries i was only
able to get into it a couple of times.
So if i had a way to reliably force my way into the bios, it would
be very valuable, whether for pxe boot or just anything else.
Providing more information will allow readers to provide better answers:
- The make and model of the computer. If homebrew, the make and model
of the chassis and motherboard. Please provide URL's.
- How many drive bays the chassis has.
- Makes and models of drives in the computer, and which bay each is in.
Please provide URL's.
- How many and type of drive interfaces on the motherboard and/or
HBA(s), and which drive is connected to which interface.
- The purpose of each drive, what it contains (e.g. partitioning scheme,
slices/
Re: installing debian 10 without a cd and without usb but could use ethernet
On Thu 21 Jan 2021 at 23:22:32 +, Brian wrote:
> menuentry 'Debian 10' {
> linux (hdX,msdosY)/boot/vmlinuz
> initrd (hdX,msdosY)/boot/initrd.gz
> }
>
> or use a "search" line.
The latter might br easiest for you.
Put linux, initrd.gz and the ISO file in /boot on the Linux Mint
partition. Get the UUID of this partition (YOUR_UUID) from
lsblk -o +UUID
Then
search --no-floppy --UUID --set=root YOUR_UUID
menuentry 'Debian 10' {
linux /boot/vmlinuz
initrd /boot/initrd.gz
}
--
Brian.
Re: installing debian 10 without a cd and without usb but could use ethernet
On Thu 21 Jan 2021 at 11:08:56 (-0800), Dan Hitt wrote:
> On Thu, Jan 21, 2021 at 10:02 AM Dan Ritter wrote:
> > Dan Hitt wrote:
> > > On Thu, Jan 21, 2021 at 3:33 AM Brian wrote:
> > > >
> > > >menuentry 'Debian 10' {
> > > >linux /boot/vmlinuz
> > > >initrd /boot/initrd.gz
> > > >}
> > >
> > > And that is: how can grub2 or any other software know what partition
> > > '/boot' refers to?
> > >
> > > So i presume that in this very very short stanza you provide, there will
> > > also have to be a search line like David has (search --no-floppy ..)
> > > to identify just where '/boot' is (???).
Yes, my PCs all have two root filesystems (normally stable and
oldstable), hence the search line. And with a BIOS Boot partition and
an EFI one too, neither of the roots is going to be as early as the
first partition.
The stanza is embedded into grub.cfg at 40_custom, so there are loads
of modules already set up.
> > https://en.wikipedia.org/wiki/GNU_GRUB#Startup_on_systems_using_BIOS_firmware
> >
> > stage 2: core.img loads /boot/grub/i386-pc/normal.mod from the
> > partition configured by grub-install. If the partition index has
> > changed, GRUB will be unable to find the normal.mod, and
> > presents the user with the GRUB Rescue prompt.
> >
> > So the answer to your question is, it's been configured at
> > install time, not discovered at runtime.
>
> Thanks Dan for your mail, and for the reference to the wikipedia article.
>
> When you say 'configured at install time', does that refer to the time at
> which i run 'sudo update-grub' (on my mint host)?
>
> (I presume that it is impossible that this refers to the time when grub
> itself was last installed on the box, several years ago.)
As dsr wrote, it's when grub-install is run that you decide where most
of the Grub stuff is loaded from, including the grub.cfg file. So this
could indeed be years ago if grub itself hasn't needed upgrading.
Commands like update-grub (and grub-mkconfig) merely play around with
the /boot/grub/grub.cfg that belongs to the system you're running.
So, for example, I boot using the (newest) Grub on buster. When I
(occasionally) upgrade stretch on that PC, I update-grub so that the
stretch grub.cfg information is correct, but I don't grub-install
on stretch (ie I don't touch the MBR or UEFI).
Then I boot buster and run its update-grub so that the new information
in stretch's "BEGIN /etc/grub.d/10_linux" section gets incorporated
into buster's "BEGIN /etc/grub.d/30_os-prober" section. (That's the
only way in which stretch's grub.cfg is ever used.)
> Anyhow, i added an entry to /etc/grub.d just to see what would happen if i
> took the simple menu entry quite literally:
> menuentry "simple-test" {
> linux /boot/vmlinuz
> initrd /boot/initrd.gz
> }
>
> I ran 'sudo update-grub', and the entry was copied into /boot/grub/grub.cfg
> without modification. And then i tried booting into it, just to see what
> grub would do. And, it did what i think was the only thing it possibly
> could: it reported:
> error: file `/boot/vmlinuz' not found.
> error: you need to load the kernel first.
>
> Press any key to continue...
>
> Now, Brian said that "the installer's initrd does not contain a loop
> module", so that would indicate that if i want to use
> debian-10.7.0-amd64-xfce-CD-1.iso, i'll need to get it on the disk
> (presumably by just unpacking it somewhere --- prior to booting, i can loop
> mount it and copy it to a 'real' directory), and then modifying
> /boot/vmlinuz and /boot/initrd.gz to be paths that grub understands. Or,
> maybe the debootstrap method Bastien suggests would be good.
My own priority is just to circumvent the BIOS limitation of my oldest
PC, but otherwise get the most similar installation to my usual
net-install. Because of that PC's architecture, bullseye may well turn
out to be its final installation after 20 years.
Cheers,
David.
Re: Debian jessie > buster IPv6 link scope change of behaviour
On Thu 21 Jan 2021 at 08:16:23 (+0100), Marc SCHAEFER wrote: > On Thu, Jan 21, 2021 at 08:04:05AM +0100, Marc SCHAEFER wrote: > > fe80::1 is specifically a link-local scope, a bit like if you try to > > access a class variable without telling in what class it is. > > Reading RFC-4291 [1], 2.5.6 (link-local addresses) and RFC-4007 [2] 6, > Zones Indices: > >Because the same non-global address may be in use in more than one >zone of the same scope (e.g., the use of link-local address fe80::1 >in two separate physical links) and a node may have interfaces >attached to different zones of the same scope (e.g., a router >normally has multiple interfaces attached to different links), a node >*requires* an internal means to identify to which zone a non-global >address belongs. This is accomplished by assigning, within the node, >a distinct "zone index" to each zone of the same scope to which that >node is attached, and by allowing all internal uses of an address to >be qualified by a zone index. Yes, that documents what we normally observe as a %eth0 or %1 suffix for IPv6 addresses which selects the interface to use. "Requires" (unemphasised in the original) mean that it is necessary to identify a particular zone, but IMHO doesn't mean that a choice of zone is necessary for the use of an address. I think the last clause means that every address specification should allow for a zone ID to be present as well. In the past, some software could handle a zone, but some couldn't because it hadn't been taken into account. > Also: > >An implementation should also support the concept of a "default" zone >for each scope. And, when supported, the index value zero at each >scope SHOULD be reserved to mean "use the default zone". Unlike >other zone indices, the default index does not contain any scope, and >the scope is determined by the address that the default index >accompanies. An implementation may additionally define a separate >default zone for each scope. Those default indices can also be used >as the zone qualifier for an address for which the node is attached >to only one zone; e.g., when using global addresses. Yes, and I understand that to mean that the absence of the index in your fe80::1 (as opposed to fe80::1%eth0) should allow fe80::1 to be a valid address when there is only one zone or a default zone has been chosen from amongst the many that are available. > If I read well, recent Linux kernels might have decided that the first > Ethernet interface is the default zone. Or at least this is how I > understand the *requires* above. I don't think it's quite that simple. > [1] https://tools.ietf.org/html/rfc4291#section-2.5.6 > [2] https://tools.ietf.org/html/rfc4007 As [2] says: "For example, two different physical links may each contain a node with the link-local address fe80::1. "Zones of the different scopes are instantiated as follows: […] " o Each link and the interfaces attached to that link comprise a single zone of link-local scope (for both unicast and multicast)." which covers, for example, two PCs connected by both a Cat5 cable and an ad hoc wireless link. Happening to have unpacked kernel sources for jessie and buster lying around, I took a look at diffs between code in net/ipv6/, in particular (icmp,ping,route,addrconf,anycast}.c. Without really understanding most of the code, obviously, I can see things that look like handling multiple devices, with reference counts, metrics, expiration times and so on, resulting in a lot more code. My hunch is that jessie threw an error merely because it couldn't handle making a decision on which interface to use, whereas buster is able to do so. What's going on might be explained by this page I came across, from which I quote: Neighbor Discovery If you have not specified the sin6_scope_id member in the sockaddr_in6 structure, have not bound a source address, and have not specified a route for link-local addresses, then the IPv6 protocol will try Neighbor Discovery to resolve the destination link-local address. For a given packet being sent, one interface is tried. This first interface that is tried is considered the most preferred interface. If Neighbor Discovery fails to resolve the link-local address on an interface, the packet to be sent is dropped and the system remembers that the destination link-local address is not reachable over that interface. On the next packet to be sent under all of the same conditions, a different interface is tried for Neighbor Discovery. This process continues through each of the interfaces on a local computer for each new packet until Neighbor Discovery responds for the destination link-local address or all of the possible interfaces have been tried and failed. Each time an attempt to resolve the neighbor fails, one interface is eliminated for that neighbor. If the destination link-local
Re: [OT] Upstream Default (FOSS) DDX Driver for NVidia GPUs is not Nouveau
On Thu 21 Jan 2021 at 01:14:31 (-0500), Felix Miata wrote: > David Wright composed on 2021-01-20 23:58 (UTC-0600): > > On Wed 20 Jan 2021 at 21:17:30 (-0500), Felix Miata wrote: > > >> When I want TV I use a TV and remote controller and an easy chair or sofa, > >> not a > >> workchair and desk and rodent and keyboard and all the distractions of > >> multitasking. My computers are general purpose tools for getting work > >> done. Most > >> videos are primarily time devouring entertainers, amusement. > > >> a (as first letter as affects meaning of word): anti or not > > >> muse: v: to think or ruminate upon. > > >> thus: > > >> amuse: literally not think or unthink > > > Ouch. It's derived from French: the à means to/at, not anti/without. > > In many cases yes. It has several others as well, such as in avocation. In > others, > such as amuse, clearly not to or at: > > "a- (3) > > prefix meaning "not, without," from Greek a-, an- "not" (the "alpha > privative"), > from PIE root *ne- "not" (source also of English un-). Quite simply, you arbitrarily picked the wrong choice (3) and skipped over: a- (1) […] In words from Romanic languages, often it represents reduced forms of Latin ad "to, toward; for" (see ad-), or ab "from, away, off" (see ab-); both of which by about 7c. had been reduced to a in the ancestor of Old French. In a few cases it represents Latin ex. > In words from Greek, such as abysmal, adamant, amethyst; also partly > nativized as > a prefix of negation (asexual, amoral, agnostic). The ancient alpha privatum, > denoting want or absence." > https://www.etymonline.com/word/a- On Thu 21 Jan 2021 at 10:14:03 (-0500), Stefan Monnier wrote: > > https://www.etymonline.com/word/amuse states that this is the "to, at" > > prefix from Latin in a sense of "causing to", not the "not, without" one > > from Greek. > > Funnily [oops, I guess I meant "amusingly"] enough, the two end up > meaning kind of the same: in that what happens is that attention is > diverted, so when you "amuse", you cause thinking of something which > prevents thinking of what you'd have been thinking of if it weren't for > that amusement. But that *is* just a fanciful rationalisation of a false dichotomy. I can find no hint of negation in the etymology of amuse. Cheers, David.
Re: Too many levels of symbolic links
On Thu 21 Jan 2021 at 08:34:34 (+0100), steve wrote: > I have rebooted with udev_log=debug in /etc/udev/udev.conf. I see > > Jan 21 08:15:28 box systemd-udevd[607]: sdc6: Failed to update device > symlinks: Too many levels of symbolic links > Jan 21 08:15:28 box systemd-udevd[607]: sdc6: Preserve already existing > symlink '/dev/disk/by-path/pci-:00:17.0-ata-2.0-part6' to '../../sdc6' > Jan 21 08:15:28 box systemd-udevd[607]: sdc6: Preserve already existing > symlink '/dev/disk/by-partuuid/7acd2c90-b372-4bfb-a517-e14d2a17e342' to > '../../sdc6' > Jan 21 08:15:28 box systemd-udevd[607]: sdc6: Preserve already existing > symlink '/dev/disk/by-path/pci-:00:17.0-ata-2-part6' to '../../sdc6' > > And the same for other partitions which are all part of a Raid1 array. > Other partitions are not impacted. So I guess this is related. > > But I'm blocked now. You said it's working, so that's a good thing. But I can't help wondering whether there's some mistake in the way your raid is configured—a part that perhaps aids monitoring it rather than being essential to its assembly. Not that it would help me at all (I haven't used raid), but perhaps you could show how your raid is configured, for others to cast an eye over. Experienced raiders might spot something. I always archive the output of udevadm for my disks (actually I just copy the /run/udev/data/b8\:* files). Does mdadm have a equivalent manner of displaying its configuration? Cheers, David.
Help me setup home-office hardware with free software
I am setting up one (or may be two) home-office servers. I have 6 ATI Radian RX580 graphics cards that are lying in cold storage. I used them a couple of years back to experiment with various crypto-mining technologies. I want these cards to be put to use for experimenting with multi-threading work loads like may be Machine Learning or may be Image Processing of deep space data or may be run GNU Radio for experimenting with signals. I don’t want any proprietary software to run on these systems. And hence exploring possibilities. My questions: 1. I had a look at i7 and Ryzen 7. For multi-threading applications, it appears that Ryzen is better choice because it allows more number of simultaneous threads. 16 as compared to 8 on i7. Does anybody have experience with these? Please share your recommendation. 2. Are there good general purpose motherboards available for 3 graphics cards? I have experience with motherboards with 12 graphics cards. But those were not general purpose. They were specifically designed for cryto mining purpose only. 3. I’ll also put these machine to serve local copy of all my data. I’ll synchronize it with some cloud space. I need recommendations for reliable hard disks. A combination of SSD/HDD to balance between cost, performance and reliability will probably be best. But I really don’t know much about the hardware. I already have good modular corsair 1200w power-supply. So I’ll put that to use. I need your help to setup a system completely free of proprietary software. Please do reply. -- Regards, Pankaj Jangid GnuPG Fingerprint: 0B62 7424 3B26 A911 052A DDE6 7C95 6E6F F858 7689
Re: Too many levels of symbolic links
Hi, Le 21-01-2021, à 20:41:05 -0600, David Wright a écrit : On Thu 21 Jan 2021 at 08:34:34 (+0100), steve wrote: I have rebooted with udev_log=debug in /etc/udev/udev.conf. I see Jan 21 08:15:28 box systemd-udevd[607]: sdc6: Failed to update device symlinks: Too many levels of symbolic links Jan 21 08:15:28 box systemd-udevd[607]: sdc6: Preserve already existing symlink '/dev/disk/by-path/pci-:00:17.0-ata-2.0-part6' to '../../sdc6' Jan 21 08:15:28 box systemd-udevd[607]: sdc6: Preserve already existing symlink '/dev/disk/by-partuuid/7acd2c90-b372-4bfb-a517-e14d2a17e342' to '../../sdc6' Jan 21 08:15:28 box systemd-udevd[607]: sdc6: Preserve already existing symlink '/dev/disk/by-path/pci-:00:17.0-ata-2-part6' to '../../sdc6' And the same for other partitions which are all part of a Raid1 array. Other partitions are not impacted. So I guess this is related. But I'm blocked now. You said it's working, so that's a good thing. But I can't help wondering whether there's some mistake in the way your raid is configured—a part that perhaps aids monitoring it rather than being essential to its assembly. Not that it would help me at all (I haven't used raid), but perhaps you could show how your raid is configured, for others to cast an eye over. Experienced raiders might spot something. cat /proc/mdstat Personalities : [raid1] [linear] [multipath] md1 : active raid1 sdg5[3] sdc5[5] sdd5[4] 117120896 blocks super 1.2 [3/3] [UUU] md0 : active raid1 sdg1[3] sdc1[5] sdd1[4] 19514240 blocks super 1.2 [3/3] [UUU] md2 : active raid1 sdg6[3] sdc6[5] sdd6[4] 97589120 blocks super 1.2 [3/3] [UUU] unused devices: and cat /etc/mdadm/mdadm.conf # mdadm.conf # # Please refer to mdadm.conf(5) for information about this file. # # by default (built-in), scan all partitions (/proc/partitions) and all # containers for MD superblocks. alternatively, specify devices to scan, using # wildcards if desired. #DEVICE partitions containers # auto-create devices with Debian standard permissions CREATE owner=root group=disk mode=0660 auto=yes # automatically tag new arrays as belonging to the local system HOMEHOST # instruct the monitoring daemon where to send mail alerts MAILADDR root # definitions of existing MD arrays ARRAY /dev/md/0 metadata=1.2 name=paros:0 UUID=9bb7d2a7:e96b5e9e:54c12ef1:4e8992fc ARRAY /dev/md/1 metadata=1.2 name=paros:1 UUID=be24cc5a:5a29d2b8:10385878:4ae00056 ARRAY /dev/md/2 metadata=1.2 name=paros:2 UUID=78920a51:c8858cf6:a6027ee6:a016643e I always archive the output of udevadm for my disks (actually I just copy the /run/udev/data/b8\:* files). I don't understand. Thanks.
Re: Too many levels of symbolic links
Le 21-01-2021, à 09:45:18 +0200, Andrei POPESCU a écrit : On Jo, 21 ian 21, 08:34:34, steve wrote: I have rebooted with udev_log=debug in /etc/udev/udev.conf. I see Jan 21 08:15:28 box systemd-udevd[607]: sdc6: Failed to update device symlinks: Too many levels of symbolic links Jan 21 08:15:28 box systemd-udevd[607]: sdc6: Preserve already existing symlink '/dev/disk/by-path/pci-:00:17.0-ata-2.0-part6' to '../../sdc6' Jan 21 08:15:28 box systemd-udevd[607]: sdc6: Preserve already existing symlink '/dev/disk/by-partuuid/7acd2c90-b372-4bfb-a517-e14d2a17e342' to '../../sdc6' Jan 21 08:15:28 box systemd-udevd[607]: sdc6: Preserve already existing symlink '/dev/disk/by-path/pci-:00:17.0-ata-2-part6' to '../../sdc6' And the same for other partitions which are all part of a Raid1 array. Other partitions are not impacted. So I guess this is related. Could you inspect the /dev directory from a live image or similar (i.e. with the affected system not running)? Yes I can do that. What should I look for? Do you have any custom udev rules or any other custom scripts, services, etc. plugged into the early startup? Don't know: tree /etc/udev /etc/udev ├── hwdb.d ├── rules.d │ ├── 51-android.rules │ ├── 55-Argyll.rules │ ├── 69-libmtp.rules │ ├── 70-persistent-cd.rules │ └── 70-persistent-net.rules.old └── udev.conf How is the RAID assembled? See my other's morning mail. Thanks for your help. Steve
Re: Help me setup home-office hardware with free software
On Friday, January 22, 2021 12:06:11 AM EST Pankaj Jangid wrote: > I am setting up one (or may be two) home-office servers. I have 6 ATI > Radian RX580 graphics cards that are lying in cold storage. I used them > a couple of years back to experiment with various crypto-mining > technologies. > > I want these cards to be put to use for experimenting with > multi-threading work loads like may be Machine Learning or may be Image > Processing of deep space data or may be run GNU Radio for experimenting > with signals. > > I don’t want any proprietary software to run on these systems. And hence > exploring possibilities. My questions: > > 1. I had a look at i7 and Ryzen 7. For multi-threading applications, it >appears that Ryzen is better choice because it allows more number of >simultaneous threads. 16 as compared to 8 on i7. Does anybody have >experience with these? Please share your recommendation. > > 2. Are there good general purpose motherboards available for 3 graphics >cards? I have experience with motherboards with 12 graphics >cards. But those were not general purpose. They were specifically >designed for cryto mining purpose only. > > 3. I’ll also put these machine to serve local copy of all my data. I’ll >synchronize it with some cloud space. I need recommendations for >reliable hard disks. A combination of SSD/HDD to balance between >cost, performance and reliability will probably be best. But I really >don’t know much about the hardware. > > I already have good modular corsair 1200w power-supply. So I’ll put that > to use. > > I need your help to setup a system completely free of proprietary > software. Please do reply. Well I can say that I have a ryzen 7 2700X in my desktop and I am very happy with it. I don't know what question specifically I can answer for you but I have no complaints about it. I would look for a motherboard with components(sound, sensors, network controllers) that are well supported in Linux. -- Steven Mainor 0x8C7885A0309216F0C403DF2B4AFE11D89477C19B.asc Description: application/pgp-keys signature.asc Description: This is a digitally signed message part.
