We Want to Hear Your Ideas

[JustinClift]

I had it set to use default encryption for transmission. By the time I remembered that trick it was already 24hrs+ into the transfer so I wasn't going to change it then.

Definitely makes sense. :)

 

[wblock]

Doc bug: I'm still on 9.10.2-U5 and in the documentation for this version, as well as the docs for 11.1-U2, it stats: "When the OK button is clicked, a snapshot is taken and the task will be repeated according to your settings." This isn't true, at least for my version. I haven't tested on 11.1, but I'd guess it's also not true. A snapshot is not taken when I click OK. I've confirmed with "zfs list -t snapshot" and no snapshot is listed after task creation. If somebody would like to confirm with 11.1, I'll file a bug report to update the docs.

Verified, and fixed in https://github.com/freenas/freenas-docs/commit/7441c502d2813ea68d935de7dcb1b86c79f2ad81
But please, do enter bugs for this type of thing when found. It would have been easy to miss this one.

 

[Arwen]

I would like to see a virtual interface that can be composed of a 1Gb and then a 10Gb interfaces. When the 10Gb interface is up use it , if its down use the 1Gb.

Solaris has a feature like this that is NOT LACP or Teaming, or whatever. Solaris IP MultiPathing does not require any switch support. And later versions of the OS allow active / active configurations, meaning multiple IPs in the same subnet, but different interfaces. Which if outside an IPMP group is not legal, (or does not work symetrically).

 

[wblock]

I would like to see a virtual interface that can be composed of a 1Gb and then a 10Gb interfaces. When the 10Gb interface is up use it , if its down use the 1Gb.

Can't you do this already with a failover LAGG?

 

[majerus]

Can't you do this already with a failover LAGG?

Tried testing this was having issues however with my test Vm networking. Ill try sometime tomorrow and report back.

 

[IceBoosteR]

So after reading through all the posts and liking the ones which I think are important, I add some points:

Starting with an Indexing Service for SMB. I have created a bugticket moons ago:
https://redmine.ixsystems.com/issues/24353#change-162515

This would enable (and I am thinking about businesses) the user to find files quicklyl, even if they have stored 10k of Excel-Sheets...

 

[IceBoosteR]

Mounting local storage to VMs (and for Iocage also in the GUI)

 

[IceBoosteR]

OpenZFS problem:

Make deduplication more efficient. I am thinking like NetApp implementing dedup. Running the dedup process for example at night, so no impact for the users with less overhead.
A simple FAS2520 can handle up tp 840TB!!! with only 36GB of ECC RAM. And it works also.

 

[IceBoosteR]

Prioritize jails and/or VMs - so with limited CPU ressources you can prioritize. Should be easy to implement.

 

[IceBoosteR]

FreeBSD issue: Don't allocate the VMs specified MB for RAM, use it as hard limit, and only use what is acutally needed.

 

[IceBoosteR]

Allocate different services to specific interfaces.
Like having a specific 10G interface for iSCSI, another one for SMB for example, for users without LAGG

 

[IceBoosteR]

Add support for LSOF https://www.freebsd.org/cgi/man.cgi?query=lsof&manpath=FreeBSD+9.0-RELEASE+and+Ports&format=html
This would allow the admin to have a deeper dive into the open files and solving problems.

 

[IceBoosteR]

I ma not 100% sure if these points were covered before:

1. Implement a mobile application. I know there is a really good one out there, handmade by one developer. But apps for smartphones areimportant these days. You can view the health status, and get push notifications if something went wrong. I would like to see, if you want to implement something, that you will get in touch with the developer, as he has already made a great piece of software!
2. Add Smart and a general HDD overview/health status at on the status page
3. Add HA-Cluster option like Gluster FS
4. or add a Master/Slave where you can synchronize datasets like with snapshot replication, but without errors/warning if the slave is not reachable - or something like that
   
5. Add option to disable/enable snapshot replication with just one click for multiple jobs
6. Document how we can handle that FreeNAS will not be able to connect to the internet, any VM can.

And last but not least: The new GUI. Seriously, this will be an enterprise product. What the new GUI offers is a lot of gadgetry where its not needed in any case. Look at old GUI. It it offering a great overview, no wasted space, not that much of all the white spaces, and no need to scroll. I had a look to the beta UI on my testservers and ouch - the overview for filling informations is - sorry- a piece of shizzl. You cannot really see where a field is optional or not, and also not where it ends. I mean in the old GUI you had a good look/overview which fields are required, what to add and so on. Did you ever have a look at Enterprise GUIs? There are specific needs for Enterprise customers, and one of this are no animations, buttons which could lead to missclicks and it needs to be intuitiv. No need for fiddling around with hidden features or not identifyable fields. There is a reason why Qnap and Synology can sell so much units, because the GUI is great.
But this is another market, just consumer.
FreeNAS will ever be a product for people like me, understanding whats happening in the background, pay some more expensive enterprise gear, but have great features.
But what I sometimes really miss - and yes this is a little bit whiny - the easyness of adding more features. Just add Emby without finding out that you are running on FN 11.1 but warden is on 11.0, which realates to problems. Then making the step to iocage. Again reading documentation, trying things, and get other issues. This is sometimes frustrating, but this is something I knew before I made the step to FreeNAS. Now VMs can solve some problems...
Just want to mention that with a more intuitiv GUI could lead to have more people who are not that "tekky".
- Justmytwocents

 

[Ericloewe]

FreeBSD issue: Don't allocate the VMs specified MB for RAM, use it as hard limit, and only use what is acutally needed.

That's not possible, unless the guest OS cooperates. The host has no way of knowing if a given page is being used, if it's cache, or empty. Well, empty detection may be possible with nested page tables, but it's going to be expensive as hell.

 

[JustinClift]

That's not possible, unless the guest OS cooperates. The host has no way of knowing if a given page is being used, if it's cache, or empty. Well, empty detection may be possible with nested page tables, but it's going to be expensive as hell.

With Linux KVM virtualisation, that's handled by the VirtIO balloon driver for Windows VM's. It gets installed in the guest with the VirtIO tools package.

Doing some looking online now, there doesn't seem to be much mention of balloon support in Bhyve. Some mentions from 2016 that Bhyve doesn't support ballooning at that time. Not seeing any other mentions since, so at a guess it's probably still not supported.

 

[JustinClift]

I ma not 100% sure if these points were covered before:
5. Add option to disable/enable snapshot replication with just one click for multiple jobs

That's the third time it's been requested in this thread. ;)

 

[JustinClift]

Allocate different services to specific interfaces.
Like having a specific 10G interface for iSCSI, another one for SMB for example, for users without LAGG

Hmmm... isn't that already do-able, by configuring which adapter/IP gets used for each service?

eg when looking in the iSCSI section of my FreeNAS box here, it shows each of my network adapters (a mix of 1 and 10GbE), allowing me to disable traffic for that protocol on any of them. Sounds like what're you're meaning?

 

[Eluminare]

NFS4.1 would be great for using Multipath

 

[Eluminare]

Telegram Bot for sending Notifications

 

[Eluminare]

Support for Docker-Compose Files, for easier deployment of different containers