![]() |
|
OpenBSD Installation and Upgrading Installing and upgrading OpenBSD. |
![]() |
|
Thread Tools | Display Modes |
|
|||
![]() Quote:
|
|
|||
![]()
So Matthew, I see from Jggimi's post clarifying the 'fake' partition issue that there's mention of the OpenBSD boot loader doing a check for UEFI or what I suspect is the hybrid CSM booting mechanism. How might this be relevant to my BIOS observations on the target installation system that I noted in the 1st post of this thread. You were indicating there that I should definitely set the BIOS to the pure UEFI booting option. From jggimi's post I'm now guessing that it doesn't matter as the OpenBSD bootloader simply adapts to the BIOS whether its pure UEFI or hybrid (CSM enabled). Is this right?
|
|
||||
![]()
I'll jump in and try to answer. You should no longer need Legacy (CSM) boot configuration. You should be able to boot the installation media (once you successfully write it) with the hardware in UEFI mode.
Then, the installation script will ask if you want to create a new MBR or a new GPT on the hard drive during installation. This particular question is documented in the INSTALL.amd64 text file co-located with the install media images, kernels, and file sets on your nearby mirror. Last edited by jggimi; 18th November 2022 at 02:17 AM. Reason: typo - one too many commas |
|
|||
![]() Quote:
Now regarding the install.amd64 text file which I understand to be here, we've already seen from the section regarding "Create Installation Media" that its solely written with consideration for MBR based partitioning. Isn't this so? My question here then is, when I go to install this as a single (non dual boot) operating system on the HP laptop, do I have a choice of making it MBR?, or am I obliged to stick with the UEFI/GUID/GPT partitioning installation given the existing BIOS running in UEFI and thus what I understand must be a whole drive partitioned already in UEFI/GUID/GPT format? Because if I could change it to MBR (and thus be in an environment entirely consistent with the official instructions that are written for MBR), then wouldn't this be wiser for avoiding any future complications? Again if it complicates things unnecessarily I'd rather also stick with making this work under the UEFI as we've troubleshooted so far.. Does this make sense? ![]() |
|
||||
![]()
You should use a GPT. From that document:
Quote:
Individual disklabel partitions to carve up that space follow as the next step of the installation. I recommend using the default automatic partitioning for your first installation. |
|
||||
![]() Quote:
Quote:
In any case if you have a UEFI system then you should boot it in UEFI mode (ie, with CSM disabled) with a GPT disk. The UEFI standard does officially support booting from an MS-DOS partition table but in my experience the practical, real-world implementation of this feature is somewhat patchy, to say the least. A GUID partition table is recommended for UEFI systems, at least by me :-) |
|
|||
![]()
I suggest a UEFI install if possible, as that is the modern standard - most of my machines are MBR, & those that aren't are a faff to put into MBR state, so I just get used to using UEFI now.
If you answer with the defaults to all the asked questions, the installation should work OK.
__________________
Linux since 1999, & also a BSD user. ![]() |
|
|||
![]()
Thanks jggimi, Keith and Matthew for the replies. I've cobbled together your collective advices and attempted the CP command to create the OpenBSD install media (on Sandisk 32Gb USB) using a destination path of /dev/rdisk4 instead of the previous /dev/disk4. The outcome? The exact same error message as when I used the CP command without the r prefix..
Here's the exact steps I followed today: (1) On inserting the USB disk following yesterdays first CP attempt (done with destination /dev/disk4 instead of dev/rdisk4) I was confronted by an error msg that I actually saw just after MacOS terminal finished doing the CP task. ie. Disk unreadable with options to "ignore" "initialise" or "Eject" (I'd refer to the screenshot here but I'll make reference to it further down in an all encompassing screenshot as that will address this but also the main outcome of todays CP with "rdisk" command attempt. I have, however, attached a screenshot (9:49pm) showing the state of the USB drive following yesterdays semi successful CP attempt. Note that 695mb of the USB stick is written which suggests to me that the bulk of the transfer must have succeeded and that its MBR partitioned. (2)I loaded MacOS Diskutil to partition and reformat (Apple calls this "erase") the USB using exFAT (as before) but with GUID partitions selected instead of MBR taking into account the UEFI BIOS on the target installation system (HP laptop). I know Matthew that you've clarified this doesn't matter, but I didn't have the benefit of reading that post at this attempt, and it probably doesn't matter anyway. So, it formatted successfully per the attached screenshot 9:51pm (you might want to take a look in case there's something relevant to part 3's outcome) and so I went about the CreateInstallationMedia procedure: (3) Create Install Media - System_Profiler SCUSB command to double check its still /dev/disk4 for the USB (yes) - diskutil unmount /dev/disk4 (yes - success) - CP command with the /dev/rdisk4 destination path and boom - see the all encompassing screenshot referred to earlier. Take a look and tell me if you know whats going on. Noteworthy parts are the "unable to copy extended attributes" error (same as yesterday), and like yesterday that middle screen Apple warning message (Initialise, Eject etc.) came up straight after a long pause following the entry of the CP command. Now obviously this is getting quite frustrating as its day 3 and I'm unable to even get the installation media prepared. But I'm sure I'll sort it out.. hopefully in the not too distant future! I did some general net searching of this extended attribute error in the context of the CP command and came across this interesting article which explains that the data is still copied, but that its minus metadata that likely doesn't impact the relevant programme anyway. To trial this I went to the target installation system (HP laptop) and after F10 to set it as pure UEFI, I rebooted with the OpenBSD USB stick just created in step 3 and it skipped straight past it as if it didn't exist. In Windows, when I inserted the USB it came up immediately with a prompt about the drive being unreadable and to format it. I've not done that pending any replies to this post. So what's the solution here guys... How do I get around this extended attribute error. Is rsync the solution as proposed by one of the replies in that article of interest linked above? Or is it this -P argument on the CP command proposed by the other reply... Either way there must be a solution as I'm sure somebody has managed to successfully create OpenBSD USB install media from within OS X ! |
|
||||
![]()
I have no experience whatsoever with MacOS. A cursory Internet search found this guide for creating Linux USB images from MacOS, where they use sudo with dd. They use sudo for administrative authority, and with that authority use dd to write the drive image file sector-by-sector to the device.
https://www.howtogeek.com/741125/how...b-on-your-mac/ To misquote Gandalf, "rsync will not avail you, son of Apple." |
|
||||
![]()
This may help a little bit:
Your objective is to ignore what is currently on the disk drive, and to place each sequential block from the image file onto the physical drive, starting with its first sector, then its next sector, and so on ... building a new filing cabinet completely, from the proverbial screws, flanges, drawers, handles, ... and files that are contained within the image. Last edited by jggimi; 18th November 2022 at 05:26 PM. Reason: typo |
|
|||
![]() Quote:
Quote:
(My HP thin clients use F9)
__________________
Linux since 1999, & also a BSD user. ![]() |
|
|||
![]()
__________________
Linux since 1999, & also a BSD user. ![]() |
|
|||
![]() Quote:
|
|
|||
![]() Quote:
I've searched up this ArchWiki link in any event and it's instructions seem a bit superior to other material I've come across. Notably it gives MacOS specific instructions for creating the USB stick. The only difference I can see in the DD command line is that it omits the sudo command. So I may give that a try in case it makes any difference. Last edited by Entropic; 19th November 2022 at 01:50 AM. |
|
|||
![]() Quote:
Quote:
|
|
|||
![]()
Great news: I've got OpenBSD installed on the HP laptop and am ready to get it working properly now with browsers, VPN etc.. Thanks to all (Keith, Jggimi and Matthew) for your advice!
Now I need to see about getting a browser and VPN working (Firefox seems the only secure option so far as I can tell (no WebRTC etc.). I need this up and running so that I can go about changing the password on all my services. I plan to save the passwords in a password manager I've been subscribed to for many years now (Lastpass), and I've found a port, albeit last updated in early 2019. Should still work right? I've also been subscribed to ExpressVPN for years now running it on my Macbook and iPhones, but I don't see a port for it... Thanks again all |
|
|||
![]()
Excellent - enjoy your journey.....
![]()
__________________
Linux since 1999, & also a BSD user. ![]() |
![]() |
Thread Tools | |
Display Modes | |
|
|
![]() |
||||
Thread | Thread Starter | Forum | Replies | Last Post |
Installation on OpenBSD area | Zielonykid123 | OpenBSD Installation and Upgrading | 5 | 26th November 2021 05:23 AM |
directing DNS queries to local unbound? | 22decembre | OpenBSD Security | 16 | 28th December 2014 04:52 AM |
New OpenBSD installation on HDD with Windows | cableguy | OpenBSD Installation and Upgrading | 8 | 3rd April 2014 10:57 PM |
OpenBSD installation goes perfectly | passthejoe | OpenBSD Installation and Upgrading | 4 | 16th November 2012 02:40 AM |
Some questions about installation of OpenBSD | aleunix | OpenBSD Installation and Upgrading | 20 | 15th June 2008 11:57 AM |