XCP-ng 8.3 public alpha 🚀

olivierlambert

Might worth a contact to the plugin authors to put it in a authorized list (or avoid to put it in a block list)

apz

Is this release technically the point where XCP-ng starts to veer off more from the XS source as to my understanding they're already keeping some parts out of public repos nowdays?

It'd be interesting to see newer Xen for example in the future and the SMAPIv3!

olivierlambert

We can't answer since it's not only on us. But on our side, we have more knowledge and more capabilities than previously to make our own moves. However, we want to keep our "upstream first" philosophy and convince XS to stay close to us. That's why we spend some times to get decent pull requests and some patience to get them merged

We truly hope to share a common future where we wouldn't have to work twice on the same problems, but instead share our man power to be more efficient together. The rest is on XS team

Also, as said before, in increasing order of complexity:

Get a more recent Xen version
Get a more recent Linux kernel version
Get a more recent platform (beyond CentOS 7)

Anonabhar

Hi Everyone,

I was wondering if this is a bug or if it is something specific to my setup. But.. I just noticed that a 'rescan' seems to be happening on my lab server every 30 seconds..

I noticed this because I was making a new template for XOCE and was deleting snapshots (viewing the /var/log/SMlog to wait until the coalescing was complete) and noticed that this was happening.

I will include the log file for viewing.. The log file is trimmed from 0730 -> present so it would fit in the posting limits.. But this was happening from the begging of the rotated logfile..

smlog.txt

~Peg

olivierlambert

This is normal Adding @ronan-a in the loop to get a confirmation that SMAPIv1 is scanning every 30 secs.

ronan-a

@Anonabhar Could you upload the other logs (xensource.log, daemon.log, etc)? There is no valid reason to have a call to cleanup.py (not a SR scan) every 30s if there is nothing to coalesce.

Anonabhar

@ronan-a Sure.. I would be happy to... I just had to append a .txt to the end of the file in order to upload it.. Please remove .txt and decompress

logs2.tar.gz.txt

Fungusware

So I read about the 'VM snapshot with disk exclusion' feature.

Is this supported at the XAPI level now? If so, how would one go about using it?

Regards

BenjiReis

@Fungusware Hi!

Yes this feature is supported at XAPI level, the VM.snapshot method now has a ignore_vdis field which is a list of VDI ref to not include in the snapshot.
This also available through xe:

xe vm-snapshot vm=... ignore-vdi-uuids=uuid1,uuid2...

Regards

olivierlambert

@Fungusware Yes, @BenjiReis spent some time to make that contribution to XAPI upstream, see https://github.com/xapi-project/xen-api/pull/4563

benjamreis opened this pull request in xapi-project/xen-api

closed Add `ignore_vdis` to `VM.snapshot` method #4563

olivierlambert

Note: this is already implemented in Xen Orchestra since a while (by default we try to ignore the [NOBAK] disk if XAPI is able to do it, which is now the case for 8.3)

Fungusware

@BenjiReis Thanks for the info.

I'll see if I can support into the SDK also, at least for .NET anyway.

hoerup

What is current status? What is the rough expectation of next alpha/beta release

olivierlambert

What kind of status are you expecting?

Fungusware

@olivierlambert Yes, I had been doing this also for quite some time. It was certainly one of the most requested features. It also felt a bit 'brute force' to me so would be nice to use it built in.

Anonabhar

@olivierlambert Well.. I have been doing the occasional "yum update" with no new problems. Everything is working great, but I still have that re-scanning of the FS every 30 seconds.. Not impacting anything but just a little weird...

stormi

@hoerup said in XCP-ng 8.3 public alpha :

What is current status? What is the rough expectation of next alpha/beta release

The next planned release should be the first beta. There's no precise ETA yet, but first quarter of 2023 is a likely window.

Applying updates regularly on top of the alpha will also give you the same end-result.

There are a lot of updated packages that are being prepared. They're currently in the xcp-ng-testing yum repository, because we haven't had time to test them all in our internal CI. All I could assess is that they install well (yum update --enablerepo=xcp-ng-testing) and that the server I installed them to rebooted without any visible issue. Once tested, early January, they'll be moved to the xcp-ng-base repository and offered as updates to anyone running yum update on their 8.3. If you want to give it a go ahead of time, you can install them and see if everything works well or not, so that we can quickly work on any issues early January.

Recent work on 8.3 touched:

the installer (soft RAID support improvements, IPv6 support, contribution of various small improvements to upstream repositories...)
automated installation ISO generation
adding memtest86+ to the installer, both in BIOS and UEFI modes
upgrading all packages to the same level as Xenserver 8 Stream's (that's the new name for Citrix Hypervisor 8 Cloud) preview update channels - we had 4 months of changes to catch up with.
redirecting http://ip_or_name_of_xcpng_server/ to https://ip_or_name_of_xcpng_server/
ongoing work with XenServer on the python2 to python3 transition
work on UEFI certificate handling, to reach a stage where it answers both XenServer's and our needs
and more.

jhansen

Hello,
I have some trouble with update from 8.2 to 8.3

I boot the installation from USB with Image xcp-ng-8.3.0-alpha2.
Everything works until I get to the installation. The backup of the old version is still made, then this error message appears:
Old partition layout is unsupported, run prepare_host_upgrade and try again. Reboot
The 8.2 installation is a 3 month old fresh installation, I doubt there is an old partition type in this 8.2 version.
I removed all disks except the boot disk which only contains the XCP-NG system. Still the same error.
The partition structure of the disk is as follows:

Model: ATA TS512GSSD720 (scsi)
Disk /dev/sda: 512GB
Sector size (logical/physical): 512B/512B
Partition Table: gpt
Disk Flags: pmbr_boot

Number Start End Size File system Name Flags
4 1049kB 538MB 537MB bios_grub, legacy_boot
1 538MB 4295MB 3757MB ext3
2 4296MB 8590MB 4294MB ext3
3 8591MB 512GB 504GB lvm

Does anyone have an idea what's going wrong.

I don't want to do a fresh installation because then I have to restore a lot of data. It's my test machine and I have a backup, but with 29 TB it takes days to restore and without data there is little point in testing 8.3.

regards Joerg

olivierlambert

Hi,

Can you give the output of lsblk, it's more readable to me

jhansen

@olivierlambert
Does this help?

sda                                                                                                 8:0    0   477G  0 disk
├─sda4                                                                                              8:4    0   512M  0 part
├─sda2                                                                                              8:2    0     4G  0 part
├─sda3                                                                                              8:3    0   469G  0 part
│ └─XSLocalEXT--d7a8bce5--29b3--5d83--f41f--5307f2e1b1bb-d7a8bce5--29b3--5d83--f41f--5307f2e1b1bb 253:0    0 468.9G  0 lvm   /run/sr-mount/d7a8bce5-29b3-5d83-f41f-5307f2e1b1bb
└─sda1