No joy yet, the documentation is not understandable for me, how to setup a ppt device in a zone
set device
match=/dev/ppt2
and then on?
https://illumos.org/man/8/bhyve
https://man.omnios.org/7/bhyve#ppt
A blogger using #omnios passing bhyve zone commands
in a attr
https://www.cyber-tec.org/2019/05/29/using-bhyve-pci-passthrough-on-omnios/
#tribblix did not like that at all 
https://blackdot.be/2023/11/illumos-bhyve-intel-arc/
Best one yet, managed to get USB dev as a ppt dev.
#SmartOS and i did not agree with ppt_aliases or ppt_matches.
#OmniOs and pcieadm utility made it understandable on my level
prtconf -dD is ALOT
But after i have seen output of pcieadm it makes more sense.
So tomorrow i will try my favorite #tribblix to see if we can get passthrough going
People working on Linux or the BSDs (or illumos based OSes, etc), are you using two monitors? And, if so, what do you use them for?
I'm trying to understand if it makes sense to keep two monitors on my desk
Please boost
listen to Grandpa.
Get ready for the new #OmniOS release...
I took the opportunity of the #GoToSocial upgrade to 0.19.0 to also start polling the #prometheus endpoint using the #VictoriaMetrics vmagent. This lead to identifying a bug (I think) so the gentle devs now have another PR to deal with 
Regarding OSes, this is #OpenBSD polling and storing metrics from #NetBSD and rendering them using #OmniOS and #pkgsrc.
Who said I was a sane person 
Plans for tomorrow: among other things, test the refactored #BSSG on illumos-based OS.
Not sure if I can reach any #illumos #omnios #openindiana people here, but here goes: does anyone know if there is a well maintained GCP image that could be used to run some flavour of illumos? PostgreSQL uses Cirrus CI with Linux, (Free|Open|Net)BSD, Windows, macOS images. illumos is a missing target!
So results are about the same whether I ran #OmniOS, #FreeBSD or #Proxmox as a host:
- scp to the host reach ~100MB/s (NetBSD reaches 76MB/s when used as the host)
- scp to a FreeBSD VM reach ~100MB/s
- scp to a NetBSD VM reach ~36MB/s
- scp to an OpenBSD VM reach ~40MB/s
- scp to a Debian VM reach ~105MB/s
- best perfs (+10MB/s) for (Net|Open) are reached using Proxmox
I hoped that Net/Open would perform better on one of those virtualization platform 
Anyone witnesses such gaps on their lab?
Related, to whom it may concern…
The #OmniOS boo loader does not appear to support ‘unset pptdevs’ and I sure could have used that tonight.
Ping @Toasterson @SolarisDiaspora @ptribble et. al.
System Administration
Week 2, Partitions
In this video, we'll talk about how to divide a single disk -- physical or virtual -- and how the partitions relate to the physical structure of the disk. We show examples partitioning disks on #NetBSD, #OmniOS, and #Linux using the disklabel, fdisk, and format tools.
System Administration
Week 2, Storage Virtualization
In this video, we cover the concept of storage virtualization -- combining individual disks into larger storage pools and utilizing resources from such a pool. This includes a discussion of RAID and some of the different supported levels as well as Logical Volume Management (by example on #Debian #Linux). We also illustrate some of these properties by example of ZFS (on #OmniOS).
Lately, I've been very busy, both with work and personal matters. However, I did manage to do a series of experiments and tests with OmniOS, and I admit that for some workloads, it has definitely convinced me. Not to replace FreeBSD, but to complement it. Some advantages:
- The ability to run both KVM and bhyve VMs is a big advantage because, in some cases, I need to run VMs on KVM (some technicians know it and prefer it).
- By giving dedicated zones, I can assign one to each user who needs to manage their VM. This can also be done on FreeBSD by using bhyve in jails. I’ve tried it, and it works well.
- The LX zones (Linux) are more compatible compared to FreeBSD zones with the Linuxulator. Systemd works, and the zone limits are correctly reported by Linux commands such as free, etc. Sometimes, I need to give users Linux VMs, and I must use bhyve because the non-functioning systemd and incorrectly displayed limits (it always reports total RAM, not the one limited to the jail) are blocking. In this way, I can avoid double-caching, double kernel, etc.
- Some people, when they hear Illumos, think I’m talking about a new type of lamps. When I talk about *BSD, they think I’m talking about "nerdy" stuff. But as soon as I mention Solaris, their eyes widen because, in their imagination, it represents stability and security par excellence. So welcome OmniOS!
However, I have a series of issues that are limiting my testing/deployment:
1) No problem on Hetzner physical hosts – both in ipv4 and ipv6. I just need to configure the global zone with a /72 IP (instead of /64) and assign another /72 to the interface that will be assigned to the zones, enabling routing. It works; Hetzner accepts this routed setup, a dynamic I have already encountered and tested successfully on FreeBSD.
2) Hetzner VPS "cloud" servers: The Virtio-SCSI supported by Illumos is 1af4:1004, while the Virtio SCSI controller at Hetzner is: Red Hat, Inc. Virtio 1.0 SCSI [1af4:1048] (rev 01). This means that OmniOS VMs cannot boot. I reported this in the OmniOS IRC channel, and they confirmed the incompatibility. It shouldn’t be hard to adapt/create the driver, but personally, I am not able to, so I’m stuck.
3) It runs well on Netcup, and everything works fine in ipv4, but I am encountering issues (the same ones I had with FreeBSD) with ipv6. It seems that Netcup doesn’t like that type of routing, so I have to assign ipv6 to the main interface, using shared routing. It doesn’t make me jump for joy. Also, in this case, there seem to be problems (no issues from the global zone, but from other zones, I can ping Google, but there are reachability issues with other hosts. I am investigating; I remember having the same issue on FreeBSD, but I don’t remember if/how I solved it).
The experiments will continue in the coming days.
