Troubles installing openSUSE 15.3 on a HP Spectre laptop - Intel Optane memory
Hello - I started a thread on the OpenSuSE users mail list, and it was suggested I move the thread over here to the factory mail list in the hope that some of the developers will take note of it and perhaps offer a solution or some advice. In a nutshell I am unable to install either OpenSuSE 15.3 or 15.4 Alpha on an HP laptop that has Windows 10 Pro, and a 1TB SSD card with Optane memory. The thread can be followed here - https://lists.opensuse.org/archives/list/users@lists.opensuse.org/thread/D3F... It was also suggested that I submit a bug report but I will hold off until someone on this list tells me to do so.... I will much appreciate it if some of you can take the time to read this thread and get back to me with your thoughts, ideas, and suggestions. Thanks in advance! Marc -- --... ...-- .----. ... -.. . .-- .- --... .--. -..- .-- -- .- .-. -.-. <b>Computers: the final frontier. These are the voyages of the user Marc.<br> His mission: to explore strange new hardware. To seek out new software and new applications.<br> To boldly go where no Marc has gone before!<br></b>
* Larry Len Rainey
Most windows 10 installs are with the BIOS set to RAID. Linux will not install on m.2 drives set to RAID, It has to be set to SATA for Linux to see the drive.
The sad part is Windows will have to be reinstalled if you change the BIOS to SATA and Windows was installed as RAID. The performance difference is not normally noticeable.
I recently did this successfully. as I recall, the only adjustment I had to do to windows was the system recovery, not a "reinstall". I googled for the advised procedure and followed it.
If you really want both Windows 10 Pro and OpenSUSE - and keep the BIOS RAID - I would suggest installing VrtualBox in Windows 10 and loading OpenSUSE as a virtual machine in Windows 10. That way the Optane Windows performance will still be there for OpenSUSE - the speed difference on new hardware virtually vs running native is not but 4% (Unless you want games like Steam - they run better in Windows). For virtual OpenSUSE - I suggest 40tb virtual drive and 4 gb ram and at least 2 CPU's - 4 if you have 8 threads. I have one 15.3. one 15.4 beta and a Tumbleweed VM all on my 1TB Intel 670 nvme on a Dell i5-8350 with 32gb ram and a Windows 10 VM and a Windows 11 VM (patched to allow no TPM).
My 2 cents.
On 12/30/21 22:08, Marc Chamberlin wrote:
Hello - I started a thread on the OpenSuSE users mail list, and it was suggested I move the thread over here to the factory mail list in the hope that some of the developers will take note of it and perhaps offer a solution or some advice. In a nutshell I am unable to install either OpenSuSE 15.3 or 15.4 Alpha on an HP laptop that has Windows 10 Pro, and a 1TB SSD card with Optane memory. The thread can be followed here -
https://lists.opensuse.org/archives/list/users@lists.opensuse.org/thread/D3F...
It was also suggested that I submit a bug report but I will hold off until someone on this list tells me to do so.... I will much appreciate it if some of you can take the time to read this thread and get back to me with your thoughts, ideas, and suggestions.
Thanks in advance! Marc
-- (paka)Patrick Shanahan Plainfield, Indiana, USA @ptilopteri http://en.opensuse.org openSUSE Community Member facebook/ptilopteri Photos: http://wahoo.no-ip.org/piwigo paka @ IRCnet oftc What sort of day was it? A day like all days, filled with those events that alter and illuminate our times...
On 12/30/21 20:32, Larry Len Rainey wrote:
Most windows 10 installs are with the BIOS set to RAID. Linux will not install on m.2 drives set to RAID, It has to be set to SATA for Linux to see the drive.
The sad part is Windows will have to be reinstalled if you change the BIOS to SATA and Windows was installed as RAID. The performance difference is not normally noticeable.
If you really want both Windows 10 Pro and OpenSUSE - and keep the BIOS RAID - I would suggest installing VrtualBox in Windows 10 and loading OpenSUSE as a virtual machine in Windows 10. That way the Optane Windows performance will still be there for OpenSUSE - the speed difference on new hardware virtually vs running native is not but 4% (Unless you want games like Steam - they run better in Windows). For virtual OpenSUSE - I suggest 40tb virtual drive and 4 gb ram and at least 2 CPU's - 4 if you have 8 threads. I have one 15.3. one 15.4 beta and a Tumbleweed VM all on my 1TB Intel 670 nvme on a Dell i5-8350 with 32gb ram and a Windows 10 VM and a Windows 11 VM (patched to allow no TPM).
My 2 cents.
Hi Larry, many thanks for your "2 cents" worth! ;-) You set me off on an adventure poking around in all the dark corners of the BIOS (places I don't go) and I found (I think) the RAID setup/settings. Within the BIOS Configuration menu there is a sub menu titled UEFI HII Configuration (I have no idea what this is all about or why RAID setup would be there!). Under that I found a single sub-menu called Intel(R) Rapid Storage Technology (why not just call it RAID?) Opening that I get a menu titled (Intel(R) RST 17.8.0.4507 RAID Driver. Again there is yet another single item - Optane Volume: Intel Optane, 953.9GB. Opening that menu gives me yet another menu titled OPTANE VOLUME INFO. There it gives me a couple of non-settable bits of info - Disable mode: Safe, and Size: 953.8GB. Next there is a toggle item that allows me to either Enable or Disable the Optane memory. It is set to show Disable meaning that the Optane memory is enabled, and clicking on it will disable the Optane memory and set the toggle button to show Enable. Below that there is list titled Volume member disks: and there are two sub menu items, one for the Optane memory and the other for the SSD drive. Opening up the SSD drive menu shows a number of non-settable info items - Port: 4.0 Model Number: INTEL HBRPEKNX0203AH Serial Number: TE034405TW1POC-1 Size: 953.8GB Status: Non-RAID Controller Type: NVNe Controller Interface: TG Opening up the Optane drive menu also shows a number of non-settable info items - Port: 5.0 Model Number: INTEL HBRPEKNX0203AHO Serial Number: TE034405TW1POC-2 Size: 27.2GB Status: Cache Controller Type: NVNe Controller Interface: TG My take is that the two "drives" are not set up as RAID drives. Sorry for the wordiness but experience tells me it is best to be thorough and accurate when trying to explain things about computers. It is so easy for miscommunication! I have not found any place where a drive can be set to SATA explicitly. But as you can see, AMI BIOS likes to hide things underneath lots of menu levels so I will keep looking. The idea of using a VirtualBox under Windows sound intriguing, I know nothing about VirtualBox but will do some Googling to research it. HTHs, Marc
-- *"The Truth is out there" - Spooky* *_ _ . . . . . . _ _ . _ _ _ _ . . . . _ . . . . _ _ . _ _ _ . . . . _ _ . _ . . _ . _ _ _ _ . _ . _ . _ . _ . * Computers: the final frontier. These are the voyages of the user Marc. His mission: to explore strange new hardware. To seek out new software and new applications. To boldly go where no Marc has gone before! (/This email is digitally signed and the electronic signature is attached. If you know how, you can use my public key to prove this email indeed came from me and has not been modified in transit. My public key, which can be used for sending encrypted email to me also, can be found at - https://keys.openpgp.org/search?q=marc@marcchamberlin.com or just ask me for it and I will send it to you as an attachment. If you don't understand all this geek speak, no worries, just ignore this explanation and ignore the signature key attached to this email (it will look like gibberish if you open it) and/or ask me to explain it further if you like./)
Hi Marc On 31.12.21 18:24, Marc Chamberlin wrote:
On 12/30/21 20:32, Larry Len Rainey wrote:
My 2 cents.
Hi Larry, many thanks for your "2 cents" worth! ;-) You set me off on an adventure poking around in all the dark corners of the BIOS (places I don't go) and I found (I think) the RAID setup/settings. Within the BIOS Configuration menu there is a sub menu titled UEFI HII Configuration (I have no idea what this is all about or why RAID setup would be there!). Under that I found a single sub-menu called Intel(R) Rapid Storage Technology (why not just call it RAID?) Opening that I get a menu titled (Intel(R) RST 17.8.0.4507 RAID Driver. Again there is yet another single item - Optane Volume: Intel Optane, 953.9GB. Opening that menu gives me yet another menu titled OPTANE VOLUME INFO. There it gives me a couple of non-settable bits of info - Disable mode: Safe, and Size: 953.8GB. Next there is a toggle item that allows me to either Enable or Disable the Optane memory. It is set to show Disable meaning that the Optane memory is enabled, and clicking on it will disable the Optane memory and set the toggle button to show Enable. Below that there is list titled Volume member disks: and there are two sub menu items, one for the Optane memory and the other for the SSD drive.
Opening up the SSD drive menu shows a number of non-settable info items - Port: 4.0 Model Number: INTEL HBRPEKNX0203AH Serial Number: TE034405TW1POC-1 Size: 953.8GB Status: Non-RAID Controller Type: NVNe Controller Interface: TG
Opening up the Optane drive menu also shows a number of non-settable info items - Port: 5.0 Model Number: INTEL HBRPEKNX0203AHO Serial Number: TE034405TW1POC-2 Size: 27.2GB Status: Cache Controller Type: NVNe Controller Interface: TG
My take is that the two "drives" are not set up as RAID drives. Sorry for the wordiness but experience tells me it is best to be thorough and accurate when trying to explain things about computers. It is so easy for miscommunication!
Maybe it is not "RAID" but it might be they are sort of "combined" by the driver, the (smaller but supposedly faster) Optane drive as additional Cache for the (bigger) "Standard NVME" SSD. I seem to remember that in earlier days, the optane stuff was marketed as sort of "persistent RAM", like "(almost as) fast as RAM, but keeping the data over reset". If this is what it is, it might even make sense to use the optane storage as an additional "speed booster", even for fast NVME SSDs. Fortunately I only own old hardware so I didn't experience such setups yet :-)
I have not found any place where a drive can be set to SATA explicitly. But as you can see, AMI BIOS likes to hide things underneath lots of menu levels so I will keep looking.
there is probably no SATA in your machine (maybe for a DVD drive if it still has one), only NVMe storage.
The idea of using a VirtualBox under Windows sound intriguing, I know nothing about VirtualBox but will do some Googling to research it.
It's a Hypervisor solution, similar to VMWare Workstation (which might also work for you), available free of charge and developed as open source project. You install VirtualBox, then just create a new virtual machine from the GUI and install your openSUSE OS into that virtual machine. Good luck :-) -- Stefan Seyfried "For a successful technology, reality must take precedence over public relations, for nature cannot be fooled." -- Richard Feynman
Hi Marc, First question: do you have some Linux experience or are you a Linux newbie? On 31.12.21 05:08, Marc Chamberlin wrote:
Hello - I started a thread on the OpenSuSE users mail list, and it was suggested I move the thread over here to the factory mail list in the hope that some of the developers will take note of it and perhaps offer a solution or some advice. In a nutshell I am unable to install either OpenSuSE 15.3 or 15.4 Alpha on an HP laptop that has Windows 10 Pro, and a 1TB SSD card with Optane memory. The thread can be followed here -
https://lists.opensuse.org/archives/list/users@lists.opensuse.org/thread/D3F...
I glimpsed over it. I stopped after some people suggested it might be a rotating-rust drive inside (with an NVME interface??? No way) and to remove the optane module m( This advice is most likely totally wrong. You *might* remove the optane module *temporarily* later if we find out that the installer has problems with it, but I doubt that Linux will be interested in it. It will most likely just ignore it (or offer to install into it, but that should not be done ;-)) Most likely Larry Len Rainey's diagnosis is right: Firmware setting the controller into "raid" mode, causing it to be not easily manageable by linux. This would be probably in line with the error you are seeing about a "MdContainer" not being deleted. This is likely not about a docker container -- I don't think YaST knows anything about docker yet -- but about a md as in "multiple devices" -- Linux softraid -- "container", an agglomeration of disks, as which the windows raid setup is detected, but can't be managed. *IF* you have some experience in Linux (and do not mind reinstalling your windows installation...), then I'd suggest to boot with a Live System USB stick of Leap 15.3, there you can examine the system in more detail and maybe even repartition it wihtout reinstalling windows (Linux *might* actually be able to handle the Windows RAID thing, but the YaST installer might not). You could even try this with a Tumbleweed Live medium or 15.4 alpha, which might have newer tools. If your linux experience is low, then using VirtualBox as a Hypervisor and running openSUSE virtualized, as also suggested by Larry, might be a solution which gets you the best of both worlds.
It was also suggested that I submit a bug report but I will hold off until someone on this list tells me to do so.... I will much appreciate it if some of you can take the time to read this thread and get back to me with your thoughts, ideas, and suggestions.
A bug report is still a good idea, because the gained insights and data are later available for the development team in a condensed form. Nobody is later going to read the mailing list threads (I did not even dare to read the opensuse-users thread until its bitter end after lots of nonsense was suggested in there ;-)) Good luck! And have a lot of fun... Stefan -- Stefan Seyfried "For a successful technology, reality must take precedence over public relations, for nature cannot be fooled." -- Richard Feynman
On Friday 2021-12-31 09:48, Stefan Seyfried wrote:
it might be a rotating-rust drive inside (with an NVME interface??? No way)
Ya way. Though, it's too early to see them. https://www.pcgamer.com/nvme-2-specification/
On 31/12/2021 09.48, Stefan Seyfried wrote:
Hi Marc,
First question: do you have some Linux experience or are you a Linux newbie?
On 31.12.21 05:08, Marc Chamberlin wrote:
Hello - I started a thread on the OpenSuSE users mail list, and it was suggested I move the thread over here to the factory mail list in the hope that some of the developers will take note of it and perhaps offer a solution or some advice. In a nutshell I am unable to install either OpenSuSE 15.3 or 15.4 Alpha on an HP laptop that has Windows 10 Pro, and a 1TB SSD card with Optane memory. The thread can be followed here -
https://lists.opensuse.org/archives/list/users@lists.opensuse.org/thread/D3F...
I glimpsed over it. I stopped after some people suggested it might be a rotating-rust drive inside (with an NVME interface??? No way) and to remove the optane module m(
Not really. What we said is that Optane combo is, or was, typically done with a traditional hard disk plus an SSD, that works as fast cache. Not that this machine in particular has a rotating rust with nvme interface plus Optane. This machine in particular appears (to fdisk) to have two nvme ssd disks, one large and one Optane, a weird combo because the main disk is already very fast. The Optane combo is probably used by Windows as a "peculiar" raid array, one element big and slow, and another small and fast, working as cache of the other. Actual data: https://ark.intel.com/content/www/us/en/ark/products/189611/intel-optane-mem... Intel® Optane™ Memory H10 with Solid State Storage Intel® Optane™ Memory 32GB + Intel® QLC 3D NAND SSD 1TB, M.2 80mm PCIe 3.0 Capacity 1 TB Form Factor M.2 22 x 80mm Interface PCIe 3.0 x4, NVMe Description Introducing the new Intel® Optane™ Memory H10 with Solid State Storage which combines two breakthrough technologies, Low-latency Intel® Optane™ technology and high-density Intel® QLC 3D NAND in a single M.2 2280 form factor. https://www.hp.com/us-en/shop/tech-takes/what-is-intel-optane-memory Intel Optane Memory: What is It and Why Do You Need It? -- Cheers / Saludos, Carlos E. R. (from 15.2 x86_64 at Telcontar)
Am Freitag, 31. Dezember 2021, 11:47:15 CET schrieb Carlos E. R.:
https://www.hp.com/us-en/shop/tech-takes/what-is-intel-optane-memory
Intel Optane Memory: What is It and Why Do You Need It?
Putting the marketing gossip aside, that sounds like an additional (3rd level) cache to me. Corrections from experts are welcoem :-) Cheers Axel
On 12/31/21 00:48, Stefan Seyfried wrote:
Hi Marc,
First question: do you have some Linux experience or are you a Linux newbie? Hi Stefan, thanks for your reply, I will intersperse my comments after yours... As for my Linux experience, I have worked with Linux for many years but would not claim to be a Linux Guru. I mostly use Linux for running servers these days but in the past I have done lots of software development using Eclipse.
On 31.12.21 05:08, Marc Chamberlin wrote:
Hello - I started a thread on the OpenSuSE users mail list, and it was suggested I move the thread over here to the factory mail list in the hope that some of the developers will take note of it and perhaps offer a solution or some advice. In a nutshell I am unable to install either OpenSuSE 15.3 or 15.4 Alpha on an HP laptop that has Windows 10 Pro, and a 1TB SSD card with Optane memory. The thread can be followed here -
https://lists.opensuse.org/archives/list/users@lists.opensuse.org/thread/D3F...
I glimpsed over it. I stopped after some people suggested it might be a rotating-rust drive inside (with an NVME interface??? No way) and to remove the optane module m(
This advice is most likely totally wrong. You *might* remove the optane module *temporarily* later if we find out that the installer has problems with it, but I doubt that Linux will be interested in it. It will most likely just ignore it (or offer to install into it, but that should not be done ;-))
OK I kinda thought the same when I started out on this adventure. At least now I think I know how to disable the optane module in the BIOS.
Most likely Larry Len Rainey's diagnosis is right: Firmware setting the controller into "raid" mode, causing it to be not easily manageable by linux. This would be probably in line with the error you are seeing about a "MdContainer" not being deleted. This is likely not about a docker container -- I don't think YaST knows anything about docker yet -- but about a md as in "multiple devices" -- Linux softraid -- "container", an agglomeration of disks, as which the windows raid setup is detected, but can't be managed.
I replied to Larry's email and show what I found out about RAID on my laptop. My take is that it is not using the drives in a RAID configuration.
*IF* you have some experience in Linux (and do not mind reinstalling your windows installation...), then I'd suggest to boot with a Live System USB stick of Leap 15.3, there you can examine the system in more detail and maybe even repartition it wihtout reinstalling windows (Linux *might* actually be able to handle the Windows RAID thing, but the YaST installer might not). You could even try this with a Tumbleweed Live medium or 15.4 alpha, which might have newer tools.
I did bring up a live version of OpenSuSE Leap 15.2 and tried to use the YaST partitioner to set up partitions for Linux. But was unable to do so, the YaST partitioner got told that the SSD drive was "busy" and wouldn't touch it. I will try it again with 15.4 alpha or TW as you suggest and report back soon.
If your linux experience is low, then using VirtualBox as a Hypervisor and running openSUSE virtualized, as also suggested by Larry, might be a solution which gets you the best of both worlds.
Like I said to Larry, this sounds like an interesting approach. Will that affect my ability to run email, Apache and other servers? Will I still be able to support autofs, samba, and nfs services. I use all of these all the time.
It was also suggested that I submit a bug report but I will hold off until someone on this list tells me to do so.... I will much appreciate it if some of you can take the time to read this thread and get back to me with your thoughts, ideas, and suggestions.
A bug report is still a good idea, because the gained insights and data are later available for the development team in a condensed form. Nobody is later going to read the mailing list threads (I did not even dare to read the opensuse-users thread until its bitter end after lots of nonsense was suggested in there ;-))
OK I will try and submit one.....
Good luck! And have a lot of fun...
Oh yeah, solving computer issues is definitely my idea of having lots of fun! LOL Thanks so much for your thoughts, Marc
Stefan
-- *"The Truth is out there" - Spooky* *_ _ . . . . . . _ _ . _ _ _ _ . . . . _ . . . . _ _ . _ _ _ . . . . _ _ . _ . . _ . _ _ _ _ . _ . _ . _ . _ . * Computers: the final frontier. These are the voyages of the user Marc. His mission: to explore strange new hardware. To seek out new software and new applications. To boldly go where no Marc has gone before! (/This email is digitally signed and the electronic signature is attached. If you know how, you can use my public key to prove this email indeed came from me and has not been modified in transit. My public key, which can be used for sending encrypted email to me also, can be found at - https://keys.openpgp.org/search?q=marc@marcchamberlin.com or just ask me for it and I will send it to you as an attachment. If you don't understand all this geek speak, no worries, just ignore this explanation and ignore the signature key attached to this email (it will look like gibberish if you open it) and/or ask me to explain it further if you like./)
Am Freitag, 31. Dezember 2021, 18:55:53 CET schrieb Marc Chamberlin:
I did bring up a live version of OpenSuSE Leap 15.2 and tried to use the YaST partitioner to set up partitions for Linux. But was unable to do so, the YaST partitioner got told that the SSD drive was "busy" and wouldn't touch it. I will try it again with 15.4 alpha or TW as you suggest and report back soon.
With latest hardware you should definitely use Tumbleweed..... Cheers Axel
On Friday, December 31, 2021 11:55:53 AM CST Marc Chamberlin wrote:
On 12/31/21 00:48, Stefan Seyfried wrote:
Hi Marc,
First question: do you have some Linux experience or are you a Linux newbie?
Hi Stefan, thanks for your reply, I will intersperse my comments after yours... As for my Linux experience, I have worked with Linux for many years but would not claim to be a Linux Guru. I mostly use Linux for running servers these days but in the past I have done lots of software development using Eclipse.
On 31.12.21 05:08, Marc Chamberlin wrote:
Hello - I started a thread on the OpenSuSE users mail list, and it was suggested I move the thread over here to the factory mail list in the hope that some of the developers will take note of it and perhaps offer a solution or some advice. In a nutshell I am unable to install either OpenSuSE 15.3 or 15.4 Alpha on an HP laptop that has Windows 10 Pro, and a 1TB SSD card with Optane memory. The thread can be followed here -
https://lists.opensuse.org/archives/list/users@lists.opensuse.org/thread/ D3F4YSV6ARHNEVSMU6AE77JTYPB4XTR4/
I glimpsed over it. I stopped after some people suggested it might be a rotating-rust drive inside (with an NVME interface??? No way) and to remove the optane module m(
This advice is most likely totally wrong. You *might* remove the optane module *temporarily* later if we find out that the installer has problems with it, but I doubt that Linux will be interested in it. It will most likely just ignore it (or offer to install into it, but that should not be done ;-))
OK I kinda thought the same when I started out on this adventure. At least now I think I know how to disable the optane module in the BIOS.
Most likely Larry Len Rainey's diagnosis is right: Firmware setting the controller into "raid" mode, causing it to be not easily manageable by linux. This would be probably in line with the error you are seeing about a "MdContainer" not being deleted. This is likely not about a docker container -- I don't think YaST knows anything about docker yet -- but about a md as in "multiple devices" -- Linux softraid -- "container", an agglomeration of disks, as which the windows raid setup is detected, but can't be managed.
I replied to Larry's email and show what I found out about RAID on my laptop. My take is that it is not using the drives in a RAID configuration.
*IF* you have some experience in Linux (and do not mind reinstalling your windows installation...), then I'd suggest to boot with a Live System USB stick of Leap 15.3, there you can examine the system in more detail and maybe even repartition it wihtout reinstalling windows (Linux *might* actually be able to handle the Windows RAID thing, but the YaST installer might not). You could even try this with a Tumbleweed Live medium or 15.4 alpha, which might have newer tools.
I did bring up a live version of OpenSuSE Leap 15.2 and tried to use the YaST partitioner to set up partitions for Linux. But was unable to do so, the YaST partitioner got told that the SSD drive was "busy" and wouldn't touch it. I will try it again with 15.4 alpha or TW as you suggest and report back soon.
I'm not sure that Windows is not just hibernated, hence the drive is marked as open. To do a full shutdown in Windows, use this from an elevated command or powershell prompt: shutdown /s /f /t 0 . Then try to mount/access the drive from the live Leap USB.
If your linux experience is low, then using VirtualBox as a Hypervisor and running openSUSE virtualized, as also suggested by Larry, might be a solution which gets you the best of both worlds.
Like I said to Larry, this sounds like an interesting approach. Will that affect my ability to run email, Apache and other servers? Will I still be able to support autofs, samba, and nfs services. I use all of these all the time.
It was also suggested that I submit a bug report but I will hold off until someone on this list tells me to do so.... I will much appreciate it if some of you can take the time to read this thread and get back to me with your thoughts, ideas, and suggestions.
A bug report is still a good idea, because the gained insights and data are later available for the development team in a condensed form. Nobody is later going to read the mailing list threads (I did not even dare to read the opensuse-users thread until its bitter end after lots of nonsense was suggested in there ;-))
OK I will try and submit one.....
Good luck! And have a lot of fun...
Oh yeah, solving computer issues is definitely my idea of having lots of fun! LOL Thanks so much for your thoughts, Marc
Stefan
--
*"The Truth is out there" - Spooky*
*_ _ . . . . . . _ _ . _ _ _ _ . . . . _ . . . . _ _ . _ _ _ . . . . _ _ . _ . . _ . _ _ _ _ . _ . _ . _ . _ . *
Computers: the final frontier. These are the voyages of the user Marc. His mission: to explore strange new hardware. To seek out new software and new applications. To boldly go where no Marc has gone before!
(/This email is digitally signed and the electronic signature is attached. If you know how, you can use my public key to prove this email indeed came from me and has not been modified in transit. My public key, which can be used for sending encrypted email to me also, can be found at - https://keys.openpgp.org/search?q=marc@marcchamberlin.com or just ask me for it and I will send it to you as an attachment. If you don't understand all this geek speak, no worries, just ignore this explanation and ignore the signature key attached to this email (it will look like gibberish if you open it) and/or ask me to explain it further if you like./)
On 12/31/21 11:55, Marc Chamberlin wrote:
Like I said to Larry, this sounds like an interesting approach. Will that affect my ability to run email, Apache and other servers? Will I still be able to support autofs, samba, and nfs services. I use all of these all the time.
Marc, This is yet another Larry. I am the maintainer of VirtualBox (VB) for openSUSE. Given your workload, you will need to take some care in how the VM connects to the network. The default installation of VB does a ]NAT connection to get access to the Internet. Connections can be made from the VM to the LAN, but not from the LAN to the VM. That will not work for servers running on the VM. For this functionality, you will need to select the "Bridged Adapter Mode" in the "Attached to" section of the Network setup in the GUI. With this setting, the VM appears on your LAN as though it were a separate computer. The bridging is done for a single network interface, which can be an inconvenience for laptops that work part of the time connected to a wire, and other times working with wireless. When you switch the host adapter connected to the network you will need to change the connected adapter using the VB GUI. It is not necessary to shut down the host or the guest VM. One caveat: I have only tested this with a Linux host. I have no machines running Windows other than some Windows VMs. The VirtualBox User Manual does not explain the choice of network connections very well. A network search for "VirtualBox network setup" will give you some better options. I find https://www.nakivo.com/blog/virtualbox-network-setting-guide/ to be useful in explaining the topic. Larry
On Fri, 2021-12-31 at 09:48 +0100, Stefan Seyfried wrote:
Most likely Larry Len Rainey's diagnosis is right: Firmware setting the controller into "raid" mode, causing it to be not easily manageable by linux. This would be probably in line with the error you are seeing about a "MdContainer" not being deleted. This is likely not about a docker container -- I don't think YaST knows anything about docker yet -- but about a md as in "multiple devices" -- Linux softraid -- "container", an agglomeration of disks, as which the windows raid setup is detected, but can't be managed.
In the past, Intel RST was basically a software RAID solution with the hardware being supported either by the ahci or isci SATA drivers. The RAID meta data used the intel matrix format ("imsm" in MD RAID/mdadm terms), and it used to work just nicely with this combination (iscsi or ahci + MD RAID). So here we seem to have an NVMe configuration instead. In principle, nvme + MD RAID should also work.
The "Details..." button shows - "cannot delete MdContainer" and again I can only choose "OK" to dismiss this rather useless error message. I got no idea what an MdContainer" is!
This is a bad error message, but it almost certainly means an MD-RAID (aka Linux native software RAID) array, as Stefan stated. (*) The big question is why YaST was trying to remove the MD array. This looks wrong. Perhaps Intel has changed the meta data in ways that are incompatible with what mdadm supports. This would make sense with Stefan's hypothesis that this is not a regular RAID array but rather a caching solution of some sort. Under Linux, cache layering would normally be represented using lvmcache or bcache, not MD RAID. Still, even if the RAID meta data is unsupported, YaST shouldn't try to remove it. Doing so successfully would have destroyed your Windows installation with high likelihood. To figure out what's going on, you need to open a bug and attach full yast and boot logs of your installation attempt. If you do, feel free to set me on CC. You may also want to switch to a terminal (ctrl-alt- f2) and run "mdadm --examine /dev/nvme0n1" (or whatever your disk device is). If the Windows installation is actually using a combination of firmware setting and Windows (RST?) driver that's incompatible with Linux MD- RAID, you will have to change the firmware settings and reinstall Windows (or use virtualization, Larry suggested). Martin (*) It's called a "container" because the IMSM meta data can hold definitions of multiple RAID arrays. In practice, that matters only on servers with many disks. On laptops there's usually just one array, but MD-RAID still needs to represent the "container" and the actual array separately.
Hi Martin and Thanks much for responding. I am not grokking everything you said, but enough to get the gist of it. I have not used RAID systems much so my understanding of them is very limited. I finally managed to gather the information that you wanted me to submit in a bug report https://bugzilla.opensuse.org/show_bug.cgi?id=1194355 This was a lot more difficult to do than I first imagined. I first created a Live Tumbleweed USB stick and had to baby sit it through over 800 updates! Seems like the Tumbleweed live image ought to be updated a bit more often, there were almost a dozen instances where a particular update could not be done because of missing files, rpm or otherwise. That took me over a day to complete. Next I spent another day trying to figure out how to increase the font size of almost all of the tools so that I could read them. At 9pts (which is apparently a ridiculous default) most of the tools were unreadable and I had to look at the tools (such as gparted or Yast) on another system to figure out what was being said. I wish someone could write a tool that unifies font settings under one GUI, and sets a reasonable default font size. This has been a long standing (many many years) complaint of mine about OpenSuSE. Finally when I got to where Tumbleweed was usable, I tried the Yast partitioner and gparted to see if either could handle my system with it's SSD and Optane memory drives. No joy! Same error messages as always. You can find the output from the command mdadm --examine /dev/nvme0n1 and the Yast log files, boot messages etc in the attachment I added to the bug report. HTHs Marc On 1/3/22 04:36, Martin Wilck wrote:
On Fri, 2021-12-31 at 09:48 +0100, Stefan Seyfried wrote:
Most likely Larry Len Rainey's diagnosis is right: Firmware setting the controller into "raid" mode, causing it to be not easily manageable by linux. This would be probably in line with the error you are seeing about a "MdContainer" not being deleted. This is likely not about a docker container -- I don't think YaST knows anything about docker yet -- but about a md as in "multiple devices" -- Linux softraid -- "container", an agglomeration of disks, as which the windows raid setup is detected, but can't be managed. In the past, Intel RST was basically a software RAID solution with the hardware being supported either by the ahci or isci SATA drivers. The RAID meta data used the intel matrix format ("imsm" in MD RAID/mdadm terms), and it used to work just nicely with this combination (iscsi or ahci + MD RAID). So here we seem to have an NVMe configuration instead. In principle, nvme + MD RAID should also work.
The "Details..." button shows - "cannot delete MdContainer" and again I can only choose "OK" to dismiss this rather useless error message. I got no idea what an MdContainer" is! This is a bad error message, but it almost certainly means an MD-RAID (aka Linux native software RAID) array, as Stefan stated. (*)
The big question is why YaST was trying to remove the MD array. This looks wrong. Perhaps Intel has changed the meta data in ways that are incompatible with what mdadm supports. This would make sense with Stefan's hypothesis that this is not a regular RAID array but rather a caching solution of some sort. Under Linux, cache layering would normally be represented using lvmcache or bcache, not MD RAID. Still, even if the RAID meta data is unsupported, YaST shouldn't try to remove it. Doing so successfully would have destroyed your Windows installation with high likelihood.
To figure out what's going on, you need to open a bug and attach full yast and boot logs of your installation attempt. If you do, feel free to set me on CC. You may also want to switch to a terminal (ctrl-alt- f2) and run "mdadm --examine /dev/nvme0n1" (or whatever your disk device is).
If the Windows installation is actually using a combination of firmware setting and Windows (RST?) driver that's incompatible with Linux MD- RAID, you will have to change the firmware settings and reinstall Windows (or use virtualization, Larry suggested).
Martin
(*) It's called a "container" because the IMSM meta data can hold definitions of multiple RAID arrays. In practice, that matters only on servers with many disks. On laptops there's usually just one array, but MD-RAID still needs to represent the "container" and the actual array separately.
-- *"The Truth is out there" - Spooky* *_ _ . . . . . . _ _ . _ _ _ _ . . . . _ . . . . _ _ . _ _ _ . . . . _ _ . _ . . _ . _ _ _ _ . _ . _ . _ . _ . * Computers: the final frontier. These are the voyages of the user Marc. His mission: to explore strange new hardware. To seek out new software and new applications. To boldly go where no Marc has gone before! (/This email is digitally signed and the electronic signature is attached. If you know how, you can use my public key to prove this email indeed came from me and has not been modified in transit. My public key, which can be used for sending encrypted email to me also, can be found at - https://keys.openpgp.org/search?q=marc@marcchamberlin.com or just ask me for it and I will send it to you as an attachment. If you don't understand all this geek speak, no worries, just ignore this explanation and ignore the signature key attached to this email (it will look like gibberish if you open it) and/or ask me to explain it further if you like./)
Hi Marc, please be so kind to configure your mail client to send email to this list in plain text format, and avoid top-posting. On Wed, 2022-01-05 at 20:13 -0800, Marc Chamberlin wrote:
Hi Martin and Thanks much for responding. I am not grokking everything you said, but enough to get the gist of it. I have not used RAID systems much so my understanding of them is very limited. I finally managed to gather the information that you wanted me to submit in a bug report https://bugzilla.opensuse.org/show_bug.cgi?id=1194355
Ok, I'll have a look.
This was a lot more difficult to do than I first imagined. I first created a Live Tumbleweed USB stick and had to baby sit it through over 800 updates! Seems like the Tumbleweed live image ought to be updated a bit more often,
The current image on mirrors is 3 days old (220103). AFAIK we build a new TW image with every snapshot. Perhaps you grabbed an older one? Of course, if you're out of luck, you could meet a point in time shortly before a larger update happens, in which case 800 packages isn't really that much for TW. The network image might have worked better.
there were almost a dozen instances where a particular update could not be done because of missing files, rpm or otherwise.
Sounds related to the late issues we were seeing with mirroring, which are currently being debugged (See "Tumbleweed repo errors" thread on this ML, and https://github.com/openSUSE/zypper/issues/399). Anyway, all this is off topic.
That took me over a day to complete. Next I spent another day trying to figure out how to increase the font size of almost all of the tools so that I could read them. At 9pts (which is apparently a ridiculous default) most of the tools were unreadable and I had to look at the tools (such as gparted or Yast) on another system to figure out what was being said. I wish someone could write a tool that unifies font settings under one GUI, and sets a reasonable default font size. This has been a long standing (many many years) complaint of mine about OpenSuSE.
Now you're really mixing issues. Please open separate thread, or create a separate bug. Personally, I have never experienced this issue. It's possibly related to your monitor reporting a wrong size or DPI value. But you didn't even say if this was under X, wayland, or the text console. And it's really OT here. I understand your frustration, but this isn't an "openSUSE sucks" thread.
Finally when I got to where Tumbleweed was usable, I tried the Yast partitioner and gparted to see if either could handle my system with it's SSD and Optane memory drives. No joy! Same error messages as always.
Ok. Will follow up on BZ. Martin
On Thu, 2022-01-06 at 12:26 +0100, Martin Wilck wrote:
On Wed, 2022-01-05 at 20:13 -0800, Marc Chamberlin wrote:
I finally managed to gather the information that you wanted me to submit in a bug report https://bugzilla.opensuse.org/show_bug.cgi?id=1194355
Ok, I'll have a look.
Summary for the ML: RST has been "enhanced" to cover caching setups in addition to RAID. To my knowledge, this storage setup is currently unsupported under Linux, and YaST failing is actually better than other Linux flavors which have been reported to corrupt existing Windows installations [1]. Intel provides a Windows utility for configuring Optane devices. With that utility, it should be possible to change the storage setup such that Linux is installable. Whether or not after that the Optane device remains usable in any form under either Linux or Windows is unknown, and probably depends on BIOS/firmware features of the system at hand. Likewise, disabling the Optane cache may necessitate a Windows reinstallation. High caution is recommended. Details on BZ. Martin [1] https://askubuntu.com/questions/1204386/windows-10-wont-boot-after-dual-boot...
On 1/6/22 03:26, Martin Wilck wrote:
Hi Marc,
please be so kind to configure your mail client to send email to this list in plain text format, and avoid top-posting. Hi Martin and again thanks for responding. Noted, I will intersperse my comments following yours and/or bottom post in the future. As for plain text formatting, Thunderbird must be mucking up, I have specifically told it to send postings to the OpenSuSE mail lists in plain text format.
On Wed, 2022-01-05 at 20:13 -0800, Marc Chamberlin wrote:
Hi Martin and Thanks much for responding. I am not grokking everything you said, but enough to get the gist of it. I have not used RAID systems much so my understanding of them is very limited. I finally managed to gather the information that you wanted me to submit in a bug report https://bugzilla.opensuse.org/show_bug.cgi?id=1194355 Ok, I'll have a look.
This was a lot more difficult to do than I first imagined. I first created a Live Tumbleweed USB stick and had to baby sit it through over 800 updates! Seems like the Tumbleweed live image ought to be updated a bit more often, The current image on mirrors is 3 days old (220103). AFAIK we build a new TW image with every snapshot. Perhaps you grabbed an older one? Of course, if you're out of luck, you could meet a point in time shortly before a larger update happens, in which case 800 packages isn't really that much for TW. The network image might have worked better.
there were almost a dozen instances where a particular update could not be done because of missing files, rpm or otherwise. Sounds related to the late issues we were seeing with mirroring, which are currently being debugged (See "Tumbleweed repo errors" thread on this ML, and https://github.com/openSUSE/zypper/issues/399). Anyway, all this is off topic. OK good to know... That took me over a day to complete. Next I spent another day trying to figure out how to increase the font size of almost all of the tools so that I could read them. At 9pts (which is apparently a ridiculous default) most of the tools were unreadable and I had to look at the tools (such as gparted or Yast) on another system to figure out what was being said. I wish someone could write a tool that unifies font settings under one GUI, and sets a reasonable default font size. This has been a long standing (many many years) complaint of mine about OpenSuSE. Now you're really mixing issues. Please open separate thread, or create a separate bug. Personally, I have never experienced this issue. It's possibly related to your monitor reporting a wrong size or DPI value. But you didn't even say if this was under X, wayland, or the text console. And it's really OT here. I understand your frustration, but this isn't an "openSUSE sucks" thread. Sorry, my intent was simply to explain why it took me so long to respond. Finally when I got to where Tumbleweed was usable, I tried the Yast partitioner and gparted to see if either could handle my system with it's SSD and Optane memory drives. No joy! Same error messages as always. Ok. Will follow up on BZ. Thanks, I will look forward to seeing what, if anything, you find! Martin
-- *"The Truth is out there" - Spooky* *_ _ . . . . . . _ _ . _ _ _ _ . . . . _ . . . . _ _ . _ _ _ . . . . _ _ . _ . . _ . _ _ _ _ . _ . _ . _ . _ . * Computers: the final frontier. These are the voyages of the user Marc. His mission: to explore strange new hardware. To seek out new software and new applications. To boldly go where no Marc has gone before! (/This email is digitally signed and the electronic signature is attached. If you know how, you can use my public key to prove this email indeed came from me and has not been modified in transit. My public key, which can be used for sending encrypted email to me also, can be found at - https://keys.openpgp.org/search?q=marc@marcchamberlin.com or just ask me for it and I will send it to you as an attachment. If you don't understand all this geek speak, no worries, just ignore this explanation and ignore the signature key attached to this email (it will look like gibberish if you open it) and/or ask me to explain it further if you like./)
Hi Marc, Am Donnerstag, 6. Januar 2022, 05:13:58 CET schrieb Marc Chamberlin: ...
That took me over a day to complete. Next I spent another day trying to figure out how to increase the font size of almost all of the tools so that I could read them.
I guess you have a hiDPI Screen? This is a known (and unsolved) issue: https://bugzilla.opensuse.org/show_bug.cgi?id=1185133 Cheers Axel
participants (11)
-
Axel Braun -
Axel Braun -
Carlos E. R. -
Jan Engelhardt -
Larry Finger -
Larry Len Rainey -
Marc Chamberlin -
Mark Petersen -
Martin Wilck -
Patrick Shanahan -
Stefan Seyfried