FreeNAS 11.2-Beta2 is Now Available!

[KrisBee]

Only issue I've come across so far is when installing a plugin and choosing the netmask the netmask gets added to the management IP address. Eg 10.0.0.10/24:32400/web

Sent from my SM-G930F using Tapatalk

Install plugin with DHCP and a bridge and vnet interface are created, install with a static ip and only a "shared ip" is created. Not mentioned in Beta2 guide - what are the consequences? Plus phantom controls exist for a selected plugin on UI above "plugin>installed" list. Actions on installed plugins split between "plugin>installed" page and "jails" page.

 

[jjstecchino]

I am getting this error in the middleware log, and occasional email alert notifications that this pool (zr1) status is UNKNOWN. zr1 pool is a newly created one using the new beta interface.

Code:

[2018/08/05 11:59:13] (WARNING) middlewared.devd_listen():93 - Failed to parse devd message: !system=ZFS subsystem=ZFS type=misc.fs.zfs.history_event  pool_name=zr1 pool_guid=18288707863560879808 history_hostname=XXXXX history_internal_str=pool version 5000; software version 5000/5; uts XXXXX 11.2-STABLE 1102500 amd64 history_internal_name=create history_txg=5 history_time=1533470352


[2018/08/05 11:59:13] (WARNING) middlewared.devd_listen():93 - Failed to parse devd message: !system=ZFS subsystem=ZFS type=misc.fs.zfs.history_event  pool_name=zr1 pool_guid=18288707863560879808 history_hostname=XXXXX history_dsname=zr1 history_internal_str=compression=15 history_internal_name=set history_dsid=21 history_txg=5 history_time=1533470352


[2018/08/05 11:59:13] (WARNING) middlewared.devd_listen():93 - Failed to parse devd message: !system=ZFS subsystem=ZFS type=misc.fs.zfs.history_event  pool_name=zr1 pool_guid=18288707863560879808 history_hostname=XXXXX history_dsname=zr1 history_internal_str=aclmode=3 history_internal_name=set history_dsid=21 history_txg=5 history_time=1533470352


[2018/08/05 11:59:13] (WARNING) middlewared.devd_listen():93 - Failed to parse devd message: !system=ZFS subsystem=ZFS type=misc.fs.zfs.history_event  pool_name=zr1 pool_guid=18288707863560879808 history_hostname=XXXXX history_dsname=zr1 history_internal_str=aclinherit=3 history_internal_name=set history_dsid=21 history_txg=5 history_time=1533470352


[2018/08/05 11:59:13] (WARNING) middlewared.devd_listen():93 - Failed to parse devd message: !system=ZFS subsystem=ZFS type=misc.fs.zfs.history_event  pool_name=zr1 pool_guid=18288707863560879808 history_hostname=XXXXXX history_dsname=zr1 history_internal_str=mountpoint=/zr1 history_internal_name=set history_dsid=21 history_txg=5 history_time=1533470352

Should I file a bug?

 

[captainwtf]

I've been having a few various issues.

Installing plugins, Installing Sonarr actually installs Radarr (I tried multiple times and this issue persists/yields the same results)

When clicking on the options to browse to management of installed plugins it goes to the IP of the jail/subnetmask:port(ex: 10.10.0.6/24:8112) which doesn't work so it just redirects to the FreeNAS logon page.

Cannot install QBittorrent, errors during the install.

Aside from that everything seems to work. There is no dropdown when creating SMB shares to select directory on a dataset.

 

[zambanini]

is the new user interface a punishment? it is hardly useable, who did the userbility case study?

 

[Apollo]

is the new user interface a punishment? it is hardly useable, who did the userbility case study?

Totally agree.
On small screens, most of the space is wasted by the windows.
I find dealing with dataset extremely inefficient. Too many steps back, scrolling and clicking.
Having to play around creating nested sets of datasets, I must scroll so many times, finding where my volulume is, then expanding to get the list of my datasets, then looking for the dataset I was last in to then add another dataset by clicking the three dots, then scrolling some more.
What takes seconds and only a handfull of clicks now takes overly excessive amount of time and mouse interactions.
The interface is also not too speedy, at least not real time because made available on a as needed basis.

 

[appoli]

@zambanini @Apollo - agreed & I made that known when IX created the feedback thread for the new UI which was probably almost a year ago at this point. I almost always use my iPhone/iPad to access the FreeNAS UI (except when I need to access the shell of a jail or I’m doing something particularly detailed) and when I tried the new UI back then it was basically impossible to do anything with it unless you were using a full size machine.

I haven’t tried FN11.2, but from your feedback it sounds like this is still a pretty serious issue (although I don’t know if you mean an iOS/Android device or just a machine with a small display). The feedback we got before was that the new UI wouldn’t go into full effect until it was fully feature equivalent to the old UI & to me useability from portable devices (and being able to copy messages as someone else pointed out) is part of that.

Now the new UI is the default, although we can select the old UI at the login. However, the new UI must be used to take advantage of the new features which throws a wrench in comparability.
Since 11.2 is still in beta this is fine, but I think that an RC shouldn’t be declared until the new UI can be used from a mobile/small screen device to the same degree as the old UI could be, or at least to a fairly similar degree.
This might be a PIA, but due to the new changes I think this is kind of necessary if the new UI is going to go into production... hopefully others think so as well.

 

[Apollo]

@zambanini @Apollo - agreed & I made that known when IX created the feedback thread for the new UI which was probably almost a year ago at this point. I almost always use my iPhone/iPad to access the FreeNAS UI (except when I need to access the shell of a jail or I’m doing something particularly detailed) and when I tried the new UI back then it was basically impossible to do anything with it unless you were using a full size machine.

I haven’t tried FN11.2, but from your feedback it sounds like this is still a pretty serious issue (although I don’t know if you mean an iOS/Android device or just a machine with a small display). The feedback we got before was that the new UI wouldn’t go into full effect until it was fully feature equivalent to the old UI & to me useability from portable devices (and being able to copy messages as someone else pointed out) is part of that.

Now the new UI is the default, although we can select the old UI at the login. However, the new UI must be used to take advantage of the new features which throws a wrench in comparability.
Since 11.2 is still in beta this is fine, but I think that an RC shouldn’t be declared until the new UI can be used from a mobile/small screen device to the same degree as the old UI could be, or at least to a fairly similar degree.
This might be a PIA, but due to the new changes I think this is kind of necessary if the new UI is going to go into production... hopefully others think so as well.

Hi Appoli,

I have been experimenting with Beta from a laptop connectd to a second network where my normal server resides. my laptop is roughly 768 pixel high (Windows 10).
Beside this, I have my main desktop with A 4K monitor and even then, there is no improvement or benefit when it comes to window size and content. I still have to scroll back and forth.
I did seem to recall I have pointed this out when Corral was being released.
In my opinion, Corral and Freenas 11-2 are just using the new UI because it seems it looks cool. It wouldn't hurt if you were running the web interface on and off, mostly off, but if you need to perform serious tasks, then it is not to par.
For the most part, using the new UI is fine if you don't want to deal with much accurate details, but if you need to make serious work, then I return to the old UI.

 

[appoli]

@Apollo

Thanks for the response. The few times I've used the new UI on my MacBook the layout seems to have been fine (I don't think I had to do a lot of scrolling, but then again I have only used it briefly recently and the more in-depth usage was a while back).
This is interesting though because what I meant in regard to lack of usability from a mobile device is that you literally cannot use it. Many objects wouldn't form & when an object did show up many of the objects inside of it (i.e. links, graphs, etc.) didn't form. And if you were lucky enough to have something that you could actually see most of the time you couldn't select it.

So when I say it was unusable from a mobile device I literally meant just that - you couldn't do anything or see anything. I say your input is interesting because it sounds like even for some browsers/setups/machines/etc that can display the content, said content is being formatted poorly. And as you can probably assume this sounds like either a different side of the same issue or a different issue altogether.

I agree that if you want to do detailed work you should use the old UI. If the old and new UI accessed the same things I probably wouldn't even be saying anything, but since I use both jails & VMs this is a serious issue for me. You have to use the new UI to seen the iocage jails in the UI & from what I could tell you have to use the new UI to take advantage of changes they made to the VM infrastructure.

This is why I said that the RC should be held back until the new UI could be sorted out OR (and I would think this is much less likely) allow users to access the new features in the old UI... my two cents, but I guess it comes down to who the primary users of FreeNAS are and what their needs are.

Also, for Corrall - I actually didn't mind the UI too much from a mobile perspective. You're totally right that with a small screen you couldn't see a lot of the info, but when I had my phone in landscape orientation I could get to all the menus and see most of the info. The one main gripe I had was that whenever I clicked on a button it registered as two clicks for some reason... I'm sure you can see why this would be an issue (e.g. if you wanted to toggle something you essentially couldn't).
I did like a lot of the features of the Corrall UI. Too bad the reshuffling of the IO development team means that it is pretty much done for.

 

[MatthewSteinhoff]

@Bgadd, is iXsystems hearing our concerns with regards to the new interface?

Cheers,
Matt

 

[Bgadd]

@Bgadd, is iXsystems hearing our concerns with regards to the new interface?

Cheers,
Matt

Matt,

Yes we are listening to the feedback. We are working to address these concerns and feedback about the UI. Also, we are putting out another Beta release to address most of these concerns before the Release Candidate is released. I do not want to sound like I'm on autopilot but I do appreciate you following up, but I can reassure we will do what we can to have this polished up by the time we get the RC out.

I understand there are bugs...we are committed to making it not just a great product, but a solid one. Do remember, this is a Beta, not our final product...before 11.2 FN proper goes out, we will have about 4-5 weeks of manual testing to include internal-user acceptance testing. Keep the feedback coming and as always, submit bugs when you have issues and encourage your fellow forum members as well.

 

[Bgadd]

is the new user interface a punishment? it is hardly useable, who did the userbility case study?

Good Morning, what areas do you find hardly usable? We understand there are bugs, but are you having any unique issues that are not already mentioned? Have you submitted a bug for the issues you are having?

 

[Bgadd]

Should I file a bug?

Yes please if you have not already. Let us investigate the issue. Regardless if it has been reported or not, we can use as much information as we can get but we need those issues in our system and the more issues reported on an area in the UI, the better.

 

[emk2203]

@Bgadd: I thought I posted some feedback about the new UI, but it seems to be lost, unfortunately. I have used the new UI just for a few days, after upgrading jails to iocage, and I had to log in and out into new vs old repeatedly.

Summary: The new UI feels like trying to scrape data from PowerPoint slides, vs looking at a neat Excel sheet with the old UI. I hope this brings my point across.

The information density is very sparse, which certainly doesn't help getting an overview. For the jails, I got two pages of mount points, which would have fitted on half a page with the old UI. I had to move back and forth to see what I had.

Your products are for (semi-)professionals, who can accept less glamour for a more productive interface. One very good example for me is how the guys from MikroTik handle the UI of their network products. You can even go to the command line and do your work from there, like Corrall attempted to do:

Code:

[admin@wtp] > /interface wireless cap print
...
	   locked-caps-man-common-name: CAPsMAN-000C424C30F3

An example for their UI is at https://wiki.mikrotik.com/wiki/Manual:Winbox, this is a utility, but the WebUI is the same, and the basics concepts are explained here.

It might be too late for drastic changes, but for me, it looks like a situation similar to when Windows 8 came out. Changes were well-meant, but not usable for a lot of users. Microsoft had to cave in and improve in Windows 10.

 

[MatthewSteinhoff]

Good Morning, what areas do you find hardly usable?

How do I answer these questions using the beta UI and its new reporting system:

1. Response sucks today. There is a scrub running. It seems like it is taking a lot longer that last week. Doesn't it usually run at 3am? Is it running longer than normal?

2. The new IT Directory is worried about the reliability of the FreeNAS server. He wants to replace it with a commercial solution he used at his last company (that will cost so much as to wipe out any chance of a profit sharing bonus this quarter). The owner is a simple man so let's keep this super simple... how many reboots and how long between reboots for the last year?

3. NetOps swears they didn't mess with anything over the weekend but FreeNAS replication seems to be taking forever. It looks like we're sending far more traffic out em0 (1G) than ix0 (10G) which shouldn't happen if the routing is correct. I'd like to shove this problem back to NetOps. Can you pinpoint when traffic started flowing oddly?

All three of these (real) questions can be answered quickly and easily using the classic UI and reporting. If there is a way to answer them using the new reporting interface, I haven't found it.

Most of being a SysAdmin is knowing if what is happening now is different that what happened before and, if the answer is 'yes', figuring out when something changed. The new reporting interface only shows what is happening now. That is a failure. Without historical data, the current data is near worthless. I KNOW the system is slow NOW. What I want to know is when it started being slow and the new reports provide no help whereas the old reports gave me a year of knowledge.

Cheers,
Matt

 

[Bgadd]

Most of being a SysAdmin is knowing if what is happening now is different that what happened before and, if the answer is 'yes', figuring out when something changed. The new reporting interface only shows what is happening now. That is a failure. Without historical data, the current data is near worthless. I KNOW the system is slow NOW. What I want to know is when it started being slow and the new reports provide no help whereas the old reports gave me a year of knowledge.

Thanks Matt for this example. We have created a ticket for this request to investigate if we can go ahead and bring this into the next 11.2 release. Here is the ticket

Others, please submit any other feedback into Redmine so we can investigate and review.

 

[jclendineng]

BETA 2 is running good for me, it will restart randomly when copying large files over NFS/SMB but I have since set up a VDI server running unraid for docker/vm's, for the stability since I have to have a stable virtualization system, (highly recommend it is the best thing since sliced cheese), so the only thing my freenas is doing is handing files through nfs and acting as a cold storage/backup. It IS a nas after all, probably should keep NAS and VDI separate if possible. UI is pretty bad, looks like they went with cool looking angular over functionality, the old gui was much more useful. I find that I have to refresh a few times before certain buttons work. I did not understand the switch, as if this is aimed at being functional and stable, why try to make the gui pretty but less functional? This isn't a client facing webpage, it is a nas frontend. I am sure it will get more usable but at the moment, legacy was more to the point. This is the iOS 8 version of freenas :) overall, Im running the nightlies and no major bugs that I cannot live with while on the bleeding edge.

 

[Apollo]

I seem to be having some issues with importing encrypted volumes.
The steps are not as straight forward as seen from old GUI.
On the new GUI, step 2 (Decrypt pool) it is not possible to select the disk to decrypt, even though the help section specifies "Select disks to decrypt."
Providing the GELI file and proceeding with the import seems to never end.
Is this feature working on Beta 2?

 

[Apollo]

Issue with the encrypted volumes update:
Trying to import the volumes from Legacy, I was getting Python errors related to missing volumes. No errors wre reported with the new GUI.
The cause of the problem are the missing declared but missing volume on my system.
As a backup server, I do volume rotation, meaning that only a portion of all my volumes are present on this system. This server is intented to serve replication from my main server and also experiementing with other Freenas related stuff.
Because all the declared volume installed on the backup server cannot be pluuged in all at the same time, only the one I care to do some work with are indeed connected. The other ones are tucked away.
When I don't have the need for the volume, I just turn the server off, remove or swap the drives/volumes. I never have the need to detach and reattach the volumes, especially when encrypted. It it too troublesome and time consuming. If I have passphrase disabled, then upon powerup, the import is seamless, and the server is back up quickly. If I had to detach the volume, then everytime I would need to have them attached again, it will require the volumes to be imported when the system is already on. The side effect to this approach is that it takes a significant time before the volume becomes available.

The script that handle the import of encrypted disk/volume should be robust enough not to xrash and proceed forward with the import even though volumes are missing.

As I was writing this post, decryption of the volume was still in progress. Opening another GUI session, I was informed the volume is being resilvered.
On a RAIDZ1 made of 3 disks, one disk is showing CHKSUM errors.
This looks so much like Coral all over again.

 

[Apollo]

New GUI is real annoyance:
1) Font size between menu is not consistent. ( System=> Advance, System=> Alerts)

2) What is the purpose of the Reports button in the Dashboard section for the various volumes, Memory, CPU, CPU Temperature and Load usage if it is to only point to the first tab of the Reporting menu which is CPU load only?

3) GUI doesn't remembmer current settings:

3-1) Going to Storage=>Pools and click on the volume down arrow to show the list of datasets. Then go to a different menu such as Dashboard. Then return to Storage=>Pools, the content is hidden again.

3-2) Under Storage=>Pools, I have nested datasets, but the main down pointing arrow on the left side of the volume name only allow to expand the top level datasets. The one level and beyond deeper are not going to be shown. Clicking on the next arrows is not a fast paced process as it takes a few seconds to expand the tree. Looking for a specific dataset or simply doing dataset management is inefficient and painful. Under Corral, there was an option to manual hide certain datasets so next time the list was displayed only the selected one would be presented. I kind of like this functionality. Not sure on the long run though.

3-3) Sorting by columns (ie: Pools) doesn't work for everything.

4) "No data to display". I don't know what this means, really.
For instance, going to "Storage=> Snapshots" will show the above statement. Obviously, I know I have many snapshots, so it should show either a different message or show Freenas is processing the request, but no. And off course there is the true "No data to display", ie: under the "Tasks" menu. Yet, based on my aove statement, I don't know if I have to wait for Freenas to update the page or do I wait and feel stupid?

5) "Items per page" What is the reason for having an "Items per page" in the first place. Back to "Storage=> Snapshots", by default it will only show 10 items in the list. Of course I do not want to show only 10 items out of the remaining Gazillion ones. So with pain, I click on the "Items per page" drop down list to select the largest one. Waiting some more, or am I? After a while, the list gets populated (not really unexpected here, it has always been slow on former GUI) one of the reason I barely use this feature in the first place, but rather go through SSH to create the list and handle everything I need from there. Not the point of this post though). Anyway, once the contents gets displayed, I decide I need to go to a different menu to accomplish some more tasks, Let's say "Sharing=> Windows (SMB) Shares". There I have more than the default 10 itmes from the list. So I must go to the drop down list once more.
Then I decide to go back to my snapshots. What the heck, my "Items per page" is back to its default of 10. Let's move on.
One more shortcoming about the "Items per page" I find, is that I have to check in the lower section showing the number of items actually shown that I am actually looking at the entire picture or just a segment of it. Having a clear and distinct indication at the end of the list that there are other element that are not shown.
With this new GUI, I am spending more time trying to figure out if all I see is the complete picture.

6) This one also goes in the same sense as the "Items per page". If a column is not wide enough, it is possible to widen it by finding the tab separator, which with the default IXSystem color scheme is not visible unless you hover over it. Same issue on the "Dashboard" with the "Reports" button. When you are able to wider,l or simply adjust the width, decide to change menu and come back, this setting is lost and has to be redone once more. What a fewer clicks anyway.

7) While a volume is being Resilvered, it is nice there is a indication on the GUI top bar which is represented by two circling arrows. However, I didn't notice it right away. At first I didn't know what it was. Only when I hover over it do I notice the comment.
What really troubles me here is that the "Dashboard" shows this volume as green and Healthy. It is healthy because of the parity being intact, at list until something shows up as data corruption. However, the normal procedure and info would have been to show the volume as Degraded and better yet showing the volume in a different colors with a Resilvering element. Clicking on the two circling arrow on the top bar will show as a popup message the name and progress of the Resilvering. This type of information should really be made available as a per "Dashboard" volume details.
The same should be of the Scrubbing process. This should be integrated as part of the Volume details.

8) In the same "Dashboard" category, it doesn't make sense to show missing volumes with green colors. I also see the term "NaN GB" with the Unknown label. The same applies to encrypted volume that are either missing or "Locked". However, there should be a distinction between present but locked volume/disks and Locked but missing volume/disk. As it is, it is not possible to tell whether a locked volume is physically in the system or not. Color should reflect thhe change of status as well.

9) "Dashboard" menu: "Disks in Use" section of a volume is nice, however I would find more usefull if it was possible to show the drive configuration as well, such as stripped of RAIDZ Vdevs and which Vdevs the disk belongs. When pressing the invisible Reports button (unless you hover on top of it), it should be possible to jump to the Reporting page to only show Disk reports related to the Volume in effect.

10) "Dashboard" menu: Graphs granularity is weak and limits (vertical scale) are not easily readable and perceivable. When the scale is in percentage (%) it is somewhat fine but not quantifiable unless the mouse is being hover on the specific graph. One thing I just can't make sense of is the latest information reported by the various graphs under "Dashboard". It would seem, valid data collected less than 2-5 minutes will report as 0. This leads to indicate my CPU are running cold when in fact there are in the lower 30's, memory used peaked to an unknown maximum value and drop to 0GB only in the last 2 minutes at best. If this was real data being collected, then I would see transition from max to min at steady interval. Instead, the 0GB value occuring in the first 2 minutes will show as almost maximum after a refrsh of tghe page. In a nutsheel the graph collected and shown in the last hour will be steadily set to 29.45GB, where my maximum is 32GB.
Freenas graph are lying to us.

As Matt already mentioned a few threads ago, there doesn't happear to be historical data. The old GUI was great for that.


11) "UI Preferences": Nothing really usefull in here. Most issues stated above could find user selectable parameter to tailor my preferences and be store in here.\


In conclusion, I am terribly dissatisfied with how the interface has been implemented. I am just being frustrated with the overall quality of the user interface or the lack of.
I just can't believe iXsytems came up with this approach. Are you guys actually using this interface to do serious work?
I did complain a lot back then (Corral era) and it seems there has been no improvement in the matter.
I must say, I have rarely used the new GUI, due to its lack of performance, granularity, ease of use and data reliability (Not talking about ZFS reliability). But as the transition is being force on the new UI (Old GUI being forcefully retired), I think a larger audience is going to get back to you in an unfriendly way.


All my statement above indicates to me the GUI was never meant to be a better replacement of the old GUI. Clearly whoever designed the new GUI doesn't have an ounce of experience with a live system.
As I have stated before, it seems the primary goal is to have an interface that would appeal to the eye, and everything else is not worth having.

There really is a need for you guys to works things out, otherwise you might end up aggravating the Freenas community.

 

[Ericloewe]

Old GUI being forcefully retired

The old GUI is still there. The new GUI is just the default in 11.2.