HP Z640 - NVME PCI-E Drive

[JJ Fullmer]

@Sebastian-Roth @Tom-Elliott Is there a timeline on when you think this issue will be fixed?

Please and thank you

[JJ Fullmer]

@Sebastian-Roth @Tom-Elliott
I updated to the latest trunk since I noticed one of the commits said that it had adjusted the way partition numbers were found. But sadly it still isn’t working. I will try making it a raw image to see if that works as a workaround like it did for the other person with this problem with the m.2 drive. The debug information lsblk and for variable dump all still have the same information.

Thanks,
-JJ

[Sebastian Roth]

@Arrowhead-IT We are working on this and try to get this fixed for mmcblk devices as well (another thread). Unfortunately those devices are even more special than your nvme drive but pretty similar naming-wise! As we are all doing this in our free time there is no timeline I can give you. We are onto it and I hope we can give you something to test in the next couple of days. So please bare with us.

[Tom Elliott]

@Arrowhead-IT i did indeed push a commit that changes how the partitions are numbered but I have not yet put those files into the init. There is a lot of work to go through with it and I don’t want to push something that will break already working imaging right now.

[Sebastian Roth]

@Arrowhead-IT Here we go. I found some time to look into this more closely!
Download init.xz/init_32.xz test files from https://drive.google.com/folderview?id=0B-bOeHjoUmyMazJLZDhGaEl5VTQ&usp=sharing and put into place in /var/www/fog/service/ipxe/ or /var/www/html/fog/service/ipxe/ (probably a good idea to rename the original files instead of just overwriting them!)

Please test and report back if capturing/deploy is working (Image type: Multiple Partition - Single Disk, Host Primary Disk: /dev/nvme0n1).

[JJ Fullmer]

Awesome! I’m giving it a try right now. And my apologies if I seemed impatient in my asking, I was just curious and getting all excited.
I currently only have one image and it is a resizable one, we’ll see what happens.

Thanks,
-JJ

[JJ Fullmer]

@Sebastian-Roth I just tested with a resizable image to no avail. I got some error messages. It looks like some numbering didn’t quite work as expected. Mainly when it is searching for /dev/nvme0n1p instead of /dev/nvme0n1p1 and when it tries to find the image file named d1pp1.img instead of d1p1.img

My apologies that some of these pictures have duplicate information, this happened twice I was just ready with the camera the second time since it happened so fast.

I am currently uploading a non-resizable version of that image to test that idea. I should be able to post results on that test in like 10 minutes or so.

[JJ Fullmer]

@Sebastian-Roth Similar problem with a non-resizable disk, it doesn’t seem to add a 1 at the end of /dev/nvme0n1p
I’m going to do a debug session and see if the lsblk is any different from before and such. Let me know if there’s any other information you need. I’m here to help

On a side note, I did take out the primary hard disk specification and ran a hardware inventory and it found the hard drive just fine on its own. It didn’t get a harddrive manufacturer, model, or s/n but it knows it exist now. Yay progress!

[JJ Fullmer]

Here’s the information from the debug session with the new init. Looks mostly the same. I’m going to try installing windows with default partitions the old fashioned way and I’ll see if image capture works by chance.

[JJ Fullmer]

So here’s some good news, image capture works!
I did just a default multiple partition install of windows and it uploaded no problem.
I had still set the host primary disk to /dev/nvme0n1 in the gui.
I’m going to try re-downloading the image to the same computer
Then I’ll try to downloading the image I actually need on it that is single disk resizable.

Thanks for all the help thus far.

[JJ Fullmer]

So downloading a multi-partition image that was uploaded from an nvme drive works on an nvme drive. But I can’t download one that was made before I updated the init
I’m going to try uploading and downloading a resizable image and then I’ll also try reuploading my main image to see if uploading it using the new init makes some sort of difference.

[Sebastian Roth]

@Arrowhead-IT Thanks for testing and reporting. Tom has been working on this as well and I think that he’s going to update the official init files pretty soon.

[JJ Fullmer]

@Sebastian-Roth No problem. I’m happy to help.
So I can’t upload a resizable image. It says
“Problem opening /dev/nvme0n1p for reading”

I’m going to try uploading my base image as a multi-partition and see if I can download that one.

[JJ Fullmer]

Well it only seems to work when it was uploaded from nvme drive. That’s kind of odd.

[Tom Elliott]

@Arrowhead-IT said:

/dev/nvme0n1p

It would seem, to me, that the problem is the file @Sebastian-Roth used was not the most current. I’m going to rebuild the init’s with the edit’s I’ve done so far (along with what @Sebastian-Roth has done) and ask if you could test them?

I’ll have them uploaded to my website so there’s less confusion.

[Tom Elliott]

32 bit Init: https://mastacontrola.com/init_32.xz
64 bit Init: https://mastacontrola.com/init.xz

[JJ Fullmer]

@Tom-Elliott I will gladly test them!
I’m currently trying the inits that came from the most recent btsync and they seem to be working too for the hardware inventory task anyway, haven’t tried an image yet.
Where are the new inits to test? Oh wait you posted while I was typing this, I’ll download those and give them a go right now.

[JJ Fullmer]

@Tom-Elliott The first test I did was with a multi-disk image that I made from a vm and it went back to the behavior it did before where it acted like it was working but never gets into partclone.
I’ll do a debug session for that one since it happened to fast to snap a picture.

[Tom Elliott]

@Arrowhead-IT I’ve updated the inits on my server again.

I have not worked on the fog.upload script yet bet have updated the fog.download. I think we’re getting that much closer now as the download will no longer assume the partitions are in the form of disk + partition number. While disk + partition is typically viable for our /dev/sd* and /dev/hd* device locations it would not work so well on disks that end in numbers already ( such as nvme0n1 or what have you )

[JJ Fullmer]

@Tom-Elliott I’m giving it a go now!