Search results

Why do you think it is a problem? Esthetics? If it isn't reading/writing to the swap all the time it really doesn't matter.

Probably it isn't writing to disk but you can check with htop or iostat. What often happens is that Linux swaps out pages that aren't used very much in order to have more cache or to make room to defragment memory. Since you have a relatively small swap for the amount of RAM you have it isn't...

Did you maybe run a script to, say, remove the "not subscribed" pop-up?

PVE won't emulate features that the underlying CPU does not have, it only hides/reveals them from the VM. The use-case is for when you have different CPU's of the same family in a cluster and you want to be able to migrate. If you have an Intel CPU you can't expect that you can pretend it is an...

You need to set up an API token and write a script. Search the forum, this question has been asked many times.

What did you modify? That would be the place to look.

https://wiki.debian.org/WiFi/HowToUse#Using_ifupdown_and_wireless-tools PVE is Debian so you should be able to follow the steps in the Debian wiki section on ifupdown and wpasupplicant. You should probably configure a different IP address for the wifi. There is no need to tie it to a bridge or...

Wouldn't it be easier to put such large notes on a wiki or something and put a link in the exisiting notes? The PVE GUI isn't really intended for writing novels.

The configuration files for the installed kernels are in /boot.

There are ways to create a parallel-computing cluster with Linux. Most supercomputers these days are built that way. The thing is, your applications need to be written to take advantage of it by using a library like MPI to distribute the work. Most off-the-shelf software can't take advantage of...

/dev/shm is actually intended for inter-process communication, in particular POSIX shared memory (see "man shm_open"). It isn't really intended to be a dumping ground for all of your temporary files. In fact /dev/shm may not even exist if the kernel was compiled without POSIX shared memory...

One broken wire in the cable will force the link to 100 Mbit/s.

The link is not the hardware config screen nor the options screen.

You need to post some info about your system and the configuration of the VMs if you expect any help.

Once more, RAM that is owned by a VM is not "available" to the hypervisor. It is only "available" within the VM. Outside of the special case of KSM memory is not shared between VM's. If you want that, use a container. Why is this so hard for people to grasp?

That would be convenient. It would also do a lot of DNS lookups.

What kind of IPv6 address? A link-local (fe80::<something>)? Or a real one? Did you check that the setting is really in /etc/sysctl.conf? What OS are you configuring?

That only works because they aren't actually using all of their RAM (yet). When they do you will either need to have swap space available or the OOM-killer will zap one. If they don't ever use all of their RAM and everything is fine forever, that just says you gave them too much RAM.

I'm sorry but your question makes no sense to me. Can you rephrase?

There is no magic. You can't use more resources than you have. PVE runs fine on mini-PC's. Lots of people here do that. You just can't use more resources than you have. Do you really need all of these VM's running at the same time? There's no issue with setting up 50 VM's that need 16 GB each...