Uploading a large TV series the file management app reported a full disc error and stopped saving,
Shortly after when switching to the netdata, the browser page wasn’t showing anything
I am cuurently boning up on how to add another 2TB disc to increas the pool (look at me Mr Windows talking Linux) and increase the NAS capacity, just trying to get my ancient brain around the phenomenon of multi disc pools and splitting the pool to accommodate raid configuration.
I dont suppose the storage issue is anything to do with the Netdata issue but one never knows with them there computers and electrickery. Any body help on this please?
Must admit, since changing the board etc, the highest cpu temp reported by Netdata was low 40’s lovely feeling when something goes right
I think, if your Rockon root is on the same pool that ran out of space, this could cause unexpected behavior. Not entirely sure that the duplicate endpoint error message is really the underlying issue
If you run
docker ps -a
do you see any duplicate entries of the netdata instance (exited or running)?
Thanks @Hooverdan
Yes, the rockon root is in the data pool (master pool)
The Android tablet I use most of the time doesn’t seem to support the system shell as it just freezes then closes after the 60 second timeout.
I will dig out the laptop and try with that at another time as we are getting ready to leave for a short break.
Think I will shut down Rockstor whilst away and make a few changes on returning by adding another 2TB to the data pool and install the new fan.
Thanks again
Having just got round to replacing the cpu fan and installing extra ssd hd, all went well, rescanned for disc changes, all OK, updated to the latest version, all OK, lovelly.
Then tried the rockons but no dice, blank error message (with nothing in it) appeared so I thought I would try Hoverdam’s suggestion (via the shell and running “docker ps -a”) deploying the W11 laptop, could I get the password correct, could I as “the dark place”.
Clearly I have not got my head round the linux protocol of root and user names (or not) with associated passwords. I think I now know that the root has no name, just a password and that the username only has a password, (there is also the host name that fits in somewhere along the way).
This is where the day turned foul.
Thinking of looking at the documentation the trouble shooting process could be used to resetting the route password, luckily (or unluckily) depending on your viewpoint, I had the usb OS loader to hand and thought it would be fairly straightforward to reboot with the usb and reset the password through the troubleshooting screen.
Could I find the splash screen, then remembering problems way back earlier in my troublesome past, it no longer exists, so, I thought a reinstall would be the correct process, which went ahead faultlessly obtaining a new password (duly recorded this time) and on accessing the web ui duly filled in the required user name and password etc, accessed the NAS when the world seemed to stop spinning as there was no pool or associated structure, but the new disk was showing.
Hopefully all is not lost as I think (seriously hope) the pool and shares can be returned (the data drives would not have been erased—would they???) except I am in limbo with how it can be done.
Screen shots of the pool and shares
Thanks @Hooverdan
Certainly the pool is now listed and also the list of shares is showing but they are all empty.
What step have i missed?
Interestingly, the uptime indicator seems to be stuck on 00.50 and the version is listed as 4.5.8-0 (opensuze leap linux 5.14.21-150400.24.46-default)
I have tried to update to the latest version by clicking the up arrow next to the version which achieves nothing.
Any ideas would be much appreciated
Thanks again.
Apologies re the update
Have now located the start update “button” and attempted an update which fails to show disc activity in the dashboard screen and there is still no data in the shares.
Think I’ll try again another day
Thanks
Decided to try a fresh install
Disconnected all drives and installed a usb into the MB internal slot (thought as it was available might as well try it out releasing another ssd for storage)
Downloaded the generic iso (leap 15.6 generic.X86_64-5.0.9-0.install.iso) and built a boot disc (USB) with balenaEtcher.
Installed the usb boot drive and attempted to install the OS onto the internal usb (250GB) unfortunately the loading system didn’t recognise the drive as it couldn’t find a drive to populate. (Interestingly the internal usb used for the OS had a full set of files when viewed after the install.
Plan B
Reinstall the OS onto the original Kingston ssd, which seemed to populate OK, (data drives still disconnected) system requested to update to the latest version which was selected. All seemed well, now running 5.0.11-0
Shut down and reconnect the data drives.
Start up and import the pool by selecting the drop down bubble against the data disc entry on the disc page of the web ui this produced an “error”.
Visited the web page suggested on the error message but it made little sense to my novice mind.
Hopefully someone has greater knowledge with these matters and can help which would be gratefully received.
Thanks in advance
Screen shots below
@Mike-B Hello again.
Just wanted to pop-in on this:
Exactly. We now no-longer default to importing the ROOT pool. It’s far better to no use the system Pool at all: but we wanted to maintain the flexibility: ergo no default import of the “ROOT” pool anymore. Makes for easier on-boarding. It can always be imported if absolutely needed. But definitely best to try and manage without it. Preserved import for backward compatibility and very small installs that have only the system drive. But if a data drive/drives are available then far better re seperation of system/data to not have the system/OS/ROOT pool ‘surfaced’ within the Web-UI. Also helps to speed-up the Web-UI as it goes :).
And now its back
Browsing around looking for possible reasons why the system is not showing share content, I wondered what the little icon did and there it was.
Getting a little frustrated with the system not functioning as it was (before the content overload) I have subscribed for the stable version (and received the enormously long key code) but cannot find an update method that downgrades from the latest testing version to the latest stable version.
Am I right in thinking its a reinstall, selecting the most appropriate version during the install process with the generic iso?
@phillxnet can probably give the right answer, but I suspect that in this case you are right, a reinstall might be the answer because of the major changes that are coming with 5.x I think in situ downgrading might cause more problems …
Now running stable version 5.0.9-0
Shares showing on share page but no evidence of content.
This notice popped up similar to the previous “error” but refers to the shares and not pools.
Any ideas as to how to resolve the situation would be very welcome.
Have looked at the web page suggested but i have insufficient understanding to follow it.
(Thought (hoping) the stable version might have corrected the problem)
Thanks
The data table error you can ignore. It’s a frontend thing. Usually a browser refresh or clearing your cache will remedy that. The front-end renovation is long overdue but not yet on the plan until the underpinnings are taken care of.
On the empty shares I am not clear why this is happening to you.
Could I clarify if the ‘empty’ here is based on the zeros ‘0’ we see in share usage. That is due to the also indicated Quota Disabled state. Which we have seen on some imports of older pools.
I.e. the Shares are not empty as-such but showing zero Usage due to the Quota Disabled state of their parent Pool.
If that’s the case we are on known ground. A Pool quota disable (possibly required via cli in this instance) then re-enabled via Web-UI should resolve this: once enough time has passed for the quota info we use for Share Usage to be fully updated.
When subscribing to the Stable channel after having already updated to a newer version. You will only pick-up the next newest stable versions. But as-yet there arn’t any as we are in Release Candidate phase. This is confusing I know. And I recently tried to clarify this via the table addition at the top of the Downloads page. We haven’t got to the next stable but the RC versions in testing is now pretty close. You should be OK to update to 5.0.12-0 by moving back to testing as you get some fixes then. And to freeze you install at that version move back to stable: which is currently an empty repo. Once we earmark our first Stable release (was to be 5.0.12-0 bar the scrub unknown we are still wondering about). we publish to the stable. Our last stable release was eons ago at 4.6.1-0 and we have re-rolled the installers to include a testing (but RC) to get folks to a far more modern base give it’s been a long road in this last testing phase.
As @Hooverdan indicated: the datatable error is to be addressed in our next testing phase where we update the front-end. A Web browser refresh button use makes it go-away and it has no known affect other than being very distracting and tricky to track down. We did have a few of these: but I think we have only one left. But again: we are doing a whole-sale update of all our librarires on the front end next: but not now. Now we are moving to the next stable for our base stuff that is mainly Python.
Thanks @phillxnet (If that’s the case we are on known ground. A Pool quota disable (possibly required via cli in this instance) then re-enabled via Web-UI should resolve this: once enough time has passed for the quota info we use for Share Usage to be fully updated.)
Not being familiar with the cli, guidance as to appropriate commands is the next stage for me, hopefully the cli commands will not be to dissimilar to basic linux.
Utube help videos will come in handy I don’t doubt.
Not being able to access the shares via windows or android (My usual methods) leads me to think the system is not recognising the sub volumes. They can’t be empty as there is plenty of activity on the web ui Dasboard page.
Thanks again for your time it’s really appreciated.
Apologies for not providing that previously. I was a bit distracted with the 5.0.13-0 release (now out).
As the root user run:
btrfs quota disable /mnt2/Master-Pool
Then re-enable the quotas via the Web-UI in the Master-Pool details view.
You may have to login (if using the Web-UI embedded shell) using your Web-UI admin user (a regular user), then su command to gain super user (root) status. Local terminal access accepts root user login just dandy without needing this run-around via the su command.
The initial command line quota disable is precautionary as we have seen imported pools end up in an intermediate quota state that confuses the Web-UI. Disabling first via CLI then re-enabling in the Web-UI seems to set things straight. Our quota support within the Web-UI is fledgling at best: but we did have an update in the latest testing release on this front (just now actually) via 5.0.13-0:
that fixes Web-UI Share Resize, as after importing a Pool all its shares end up at the size of the host Pool: which is distracting. I.e. all your shares are showing this as they are 4.55 TB. It’s currently only cosmetic but still distracting.
Once you have re-enabled quotas on the Pool it will take possibly a few minutes to re-establish what size is actually taken up by the data that we hope is still there. A page refresh or two should start to indicate a non-zero share size as it tallies up the individual usage via the quota system.
As always keep us posted and keep the feedback coming: all good.
Thanks again @phillxnet
All running like clockwork now, have resized the shares with a little extra space for future additions which has provided a spectacular (well …nice) dashboard.
Interestingly I had to replace the samba shares which immediately gave client access (except the Samsung TV which used to show Jellyfin as a source automatically, so more work to do there)
All hardware additions up and running now have 1TB extra space.
Once again a successfully completed problem, mostly by your good selves.
Thinking of installing opensuse onto my laptop as getting fedup with windows continually updating (and loosing my printer software), could use any of the many distros but as Rockstor is based on the Leap version thought it was appropriate.
Anybody any thoughts on that?
