Hi, a question: I have seen my zones have many BEs. I am wondering, how am I supposed to delete the old BE's of my zones. Can you point me out?

It seems I can run beadm inside the NGZ, but it lists the older zones as "xb" in the "Active" field. I am looking at the manpage right now to understand what that means.

do you have the same be BE names in sync on the global zone?

tomww: no, they aren't in sync: pastebin.com/raw/M6FuMBPp

the top one is an NGZ, the bottom one is the GZ.

They will be linked though - deleting a GZ BE should also remove one of the NGZ ones.

I see ,thanks for the explanation. I just have deleted one BE in GZ, and check if it makes any difference in the zone.

yes, it made a change, ok.

Now a next question: I see in the beadm manpage at the destroy section: "Use this command with caution." I understand this whole command can be dangerous, but what harm can I cause with destroying the automatically created BEs?

Like I have the BE of the original install. IS this some kind of holy cow, or can I simply delete it? Is there any chainig here, are they somehow depending from eachother, so can I break my system with deleting a non-active BE in the GZ?

That note does seem a bit dramatic. The system won't allow you to delete any which are in use.

Deleting old ones just removes the ability to roll back or recover files from it.

andyf: Thanks for the confirmation. I'll then proceed to delete all old BEs except the last 2 or 3.

It's definitely prudent to keep a couple of old ones around.

I am mostly happy with my OmniOS install which I daily use for general NAS / VM platform since january. The main issue for me is the USB 3 support. I got an NVME disk in a USB enclosure attached via USB3. I use this disk explicitly as a scratchpad for a torrent client as I don't want to excercise unnecessary the main SSDs in the machine. Downloading linux isos can cause high data-transfers, and the

controller in the enclosure can "hang". It doesn't dissapear, but it won't react to any command, the related zones will also hang. In this state the only way to restore the working condition is rebooting the whole machine. I have defined a max download speed in the torrent client, which seemingly solved the issue.

I am not sure if the issue is with the Illumos USB3 support, or with the controller in the encolusre, or with the SSD itself (cooling is not exceptional).

Maybe there is a command to reboot only the USB subsystem, If so, I havent tried that.

There is some work I saw around improving the ability to reset things - illumos.topicbox.com/groups/develop…l-you-disconnect-and-reconnect-them

And there are commands in there to do port resets that might help

Also I noticed a strange thing, I have a sparse zone, running ZNC + irssi in it. I have configured SSH with key based auth. I can simply connect via the LAN to the zone and start irssi. It works just fine, this is the way how I am comminucating with you right now. However most of the time I am messing with my daily job and if i return to the terminal connected to the zone running irssi i frequently find

the SSH client is disconnected. I have no idea why it happens.

Not sure. I have `ServerAliveInteral 120` and `ServerAliveCountMax 2` set in my ~/.ssh/config for the "Host *" entry so perhaps enabling keepalives like that will help?

Like this: paste.omnios.org/?773c162d27e5abfd#…zHr6V5JZG9X26oRLkt1H1kFaJDTbKYg2Yhf

szilard: do you have a firewall / nat on anything in between (and including) then ssh client machine and the sparse zone.

if there's no data then ssh seems to be very quite so firewalls/nat can time out the tcp session.

ssh Keepalives (as andyf suggests) can solve that, or ensuring there's something regularly updating on the screen (I run irssi inside a screen session, and screen is configured to show a bar with open windows, time and load average on the bottom)

m1ari: I have firewall running in the GZ

I'll test first with the keepalive

O, i already have them: ServerAliveInterval 60 , and ServerAliveCountMax 10 under Host *

This is on the host side (a windows machine).

I am thinking about if I noticed similar behaviour with OpenBSD, but I can't recall it...

Whatever, I need to go.

Thanks for the support and have a great weekend guys!