First: Thank you all for helping me. This one is on me so if you’re busy and want to “pass” on this I’ll understand.
Second: There’s a similar issue here but there was no resolution. The pictures there is exactly what mine looks like as far as the widget is there and it’s blank.
Now for why it’s all on me. This install is in a VM, it’s me testing and so I’m pushing limits in unsupported ways. I first installed Leap 15.6, then installed Rockstor from the “Plain Vanilla” tutorial. All was good, including this widget. Then, because I’m greedy, I migrated my install to Tumbleweed using the openSUSE directions (basically change the .repo files - I did keep the three rockstor repos untouched). I did it this way because I want to run with a Desktop Environment and in my testing on VM’s I couldn’t get a DE installed and working on the Rockstor Tumbleweed .iso so I tried to come at it from this direction.
If anyone has any idea what may cause this - surely it’s a dependency - I’m all ears. Everything else I’ve been able to test and it has worked fine (note: I can’t test pools as I don’t have any drives passed through to the VM just the VM boot drive). I am doing this in an unsupported way I know so I don’t wanna waste time but if there’s some guru in here who gets his kicks troubleshooting I’m all ears. Also note I did a zypper install --force rockstor to force a reinstall in case there was some file that got overwritten and the issue is still there.
For now my main machine that I’m setting up bare metal I did the Leap 15.6 install then I installed Rockstor and it appears to be working well. I will not migrate it to Tumbleweed unless I can sort this as I am okay with the Leap base but there are a few minor annoyances (not Rockstor related) that i’ve ran into and would really love to be homogeneous on Tumbleweed if possible. Again, I know this is on me. I’ll keep digging and testing.
Just to clarify. You installed Leap15.6 (desktop) put Rockstor on top and then upgraded Leap to TW desktop, and that’s when Rockstor broke (and a forced Rockstor reinstall didn’t do anything either)?
If so, did you try to install the TW deskop and put Rockstor on it as well?
in the Rockstor logs do can you possibly see some error messages that might give a hint on the failure?
the post you’re referencing is somewhat older, so I think that was still on the CentOS vs. OpenSUSE legs.
When you check the directory /proc/diskstats is there anything in there, since the data collector pulls from there (according to that post you linked to).
Yes, you have the order of what I did correctly, and yes all was working well at the Leap15.6 (desktop) + desktop stage. In fact, my bare-metal install is in this state and I’ve been messing around and it’s great. I actually needed to use the DE to fix an issue where when I deleted the pool I just made those drives were not available to add back. I didn’t post about it because I’m playing fast and loose now since there’s no data trying to figure things out. Anyway I was able to fix it by using the DE and going into Yast and deleting the btrfs partition (which included all 6 drives). Then after a refresh in Rockstor the drives showed up and I was able to work them from there
I did try to install Rockstor on a Tumbleweed install but there was no rockstor package in the repos so it stopped there. Rockstor is in the Leap repo so that is why I went the long way around. I did not check opi or anything like that just the standard installed repos. If you know a way to install it on plain Tumbleweed I’d love that as I don’t mind “keeping Tumbleweed going”, I really just want a tool to help me maintain the RAID array.
One thing I am thankful for, and thank you @Hooverdan as you’ve answered a lot of my questions, is the log system. I struggle to remember where all the logs are and going to System → Logs I got the required log and the .tar.gz file had the path in case I want to use lnav to view them which is my preferred method to view logs. Anyway, there are no errors at all in there, nothing red but mostly a bunch of these INFO notes:
[09/Mar/2025 00:02:57] INFO [system.osi:1766] -- /dev/disk/by-id missing. See 'Minimum system requirements' in docs. --
[09/Mar/2025 00:02:57] INFO [storageadmin.views.disk:150] Deleting duplicate or fake (by serial) disk DB entry. Serial = (fake-serial-167643fa-dd50-412e-bb73-0a3fdb1a810c).
I am pretty sure when it was just Leap15.6+Rockstor I still had those notes but I still did get the “pretty graph”. Right now, that’s the only thing I cannot see. On this system /proc/diskstats is a file, not a folder. It’s content is as follows:
I just checked my bare metal Leap15.6+Rockstor and this is a file, not a folder, as well and it’s web UI displays the disk stats fine. In that other older thread I did run some of the lsblk commands and all seemed well. I tried an incognito browser to make sure there wasn’t something cached. In the last couple months I’ve spun up a dozen different ways to skin this cat, and this is the second time I ran into this on the Leap+Rockstor–>Tumbleweed attempt so I think it’s repeatable.
That should give you the installation when then executing:
zypper in --no-recommends rockstor-15.0.5-0
systemctl enable --now rockstor-bootstrap
you’re right, sorry about the mix up. At least the info is there, though still not explaining why it’s not picked up.
The /dev/disk/by-id missing might be a clue why it’s not reading it for the widget, but at this point I am not sure what the root cause would be … you said, you think these messages were also present in the 15.6 install?
Well, you’ve just given me a project. I didn’t realize there were TW specific repo’s (I didn’t see anything online, but I also didn’t ask). Since I left the THREE Rockstor repositories the same as setup during the "Plain vanilla install) which means they are still Leap15.6. I may go ahead and try to change them to the ones you shared above, do a zypper dup and see if it straightens itself out first. If not then I’ll do the new TW install (With a DE) and add Rockstor exactly as outlined here and see how it goes. If I can get TW working I’ll migrate my bare metal install over as right now it’s setup and running but no data moved over so no harm in changing things around.
I don’t think so, but I’m not sure. Everything seemed to work as expected on 15.6, and I have my bare metal install on the 15.6 with Plasma and everything I have checked is working fine including disk management, setup of Pool’s, etc. Honestly now that you’ve enlightened me to there being TW specific repo’s for Rockstor I feel like an idiot and wish I didn’t have an event this afternoon because I’d spend the rest of the day messing with this.
@JBinFla Thanks for your reporting here, all appreciated as usual. Some repeat info here but just wanted to chip-in a little with your adventures.
and:
It should be possible, but we are way cut-down in comparison to a regular Tumbleweed. So you would need a lot of additional packages. But what-ever you can still take the Tumbleweed generic install and do the DIY Rockstor package/repos addition. However the dedicated nature of our appliace approach may mean you are swimming against the tide. But maybe all will be OK for your purpose; in which case great.
Nice.
You may have miss-understood /miss-read the Install on Vanilla openSUSE/SuSE SLES here, or more likely we need a tweak on that doc.
We have had for a goodly number of year a Tumblweed repo, but the url is a little different to the Leap ones as Leap has another subdirectory to cover 15.1 → 15.6 where as Tumblweed does not as it’s rolling. I.e. we have repos for each Leap version individually. But only one repo for Tumblweed. All recent repos are also multi-arch. I see @Hooverdan has already detailed this however.
The main kiwi-ng config file rockstor.kiwi in our rockstor-installer GitHUb repo will indicate all repos used for all current profiles, along with the urls used. We also build every rpm on the target repo’s OS.arch profile and test install each on its intended OS.arch before publishing any rpm.
Incidentally, if you are on x86_64 you may want to consider Slowroll in the near future? Our next rpm should be compatible with this emerging (beta currently) Tumbleweed variant. But as per Tumblweed we have these more cutting edge openSUSE variants indicated on the Downloads page as:
Development/Advanced-user/Rescue use only
And given Slowroll is a less frequently changing Tumbleweed variant, we simply use our Tumbleweed repository. But again, this is only for the next rpm release but the initial incompatibilities have now been addressed in:
You can add virtual devices, i.e. if you are using something like VMM Virtual Machine Manager it can help with this. Just make sure to either select SATA as type/interface as it then gets a serial number automatically allocated, and fake SMART but only for show. Or if you are using virtio, which is likely faster, be sure to add unique serial numbers to them all otherwise you will see:
and big red errors in the disk page. But virtio disks do not have a SMART interface so that will all be greyed out.
So in short you can play with disk/pool management by creating a bunch of fake drives for the VM you are using. But be sure they are > 5 GB each otherwise they are ignored. Plus any smaller and you bump into btrfs small-drive limitations. 10 GB each is safer if your want to experiment with for example live btrs-raid transitions - what we call Re-Raid. I’ve had to reproduce some large disk count errors this way by creating for example 26 virtio drives way back when we had a problem at around that number. Although I would have preferred to have had real hardware with that capability of course .
Just wanted to chip-in with this one as you similarly post:
Which are virtio devices: so will by default have no serial numbers assigned, in contrast to real drives that always have serial number accessible unless they are in a poorly designed enclosure.
But as you have stated an awareness of we don’t support this type of virtual drive arrangement as it is not real redundancy, and introduces layers we try to avoid.
Thanks for all the tips. For clarity, the testing VM where I’m fussing with the non-standard install, upgrade from Leap to Tumbleweed and so on is just for play. The end result will be a bare metal install not ran in a VM, and in fact the new machine I built with all the drives has Leap + Rockstor now I’m just trying to sort this little glitch in the UI that I ran into on the VM before I make the change. Since I have no data on the drives yet, it’s somewhat trivial to reinstall everything and that’s kind of my plan once I figure out the best way. I’m happy there are Tumbleweed repo’s, what I read online was something like “replace the 15.x with your version”, so I just assumed (cough, cough) that there weren’t TW repos.
As a side note I did take that existing broken Leap 15.6, add Rockstor, upgrade to TW machine and changed the Rockstor repo’s over to what was mentioned. I did a zypper dup and even a zypper install --force rockstor to reinstall but the drive widget was still blank. I haven’t had time to do the Tumbleweed + Rockstor straight shot yet but plan to when I get a few hours in a row.
1st let me say I know I’m banging my head in an unsupported manner. I really don’t expect people to keep helping me, and really I think I’m now figuring out this isn’t the way to go but I’m stubborn so I’m gonna keep playing with the VM (left Leap on the bare metal install which is working well).
So, added the repos, and after getting in and going to add the update repo got the below. Interestingly, the drive widget still doesn’t work so I’m sure the leap → tumbleweed upgrade isn’t the cause (those machines do_not get this error they work otherwise). I did confirm yum was installed as I saw that in the error and just wanted to make sure my little knowledge couldn’t fix it with a dependency install. You know what, I’m gonna do a zypper inr since it’s a VM if it crashes and burns nothing lost. I’m sick as a dog so if I feel well enough this weekend I’m going to get the 5gbe USB installed on my Synology and the Mikrotik 10gbe router setup so I can start copying data. I’m not sure what the best tool for that is. I’d like something that just makes a “copy” of the source data and uses BTRFS snapshots to keep versions. That’s my goal anyway.
Traceback (most recent call last):
File "/opt/rockstor/src/rockstor/system/pkg_mgmt.py", line 109, in rpm_build_info
o, e, rc = run_command([YUM, "info", "installed", "-v", pkg])
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/opt/rockstor/src/rockstor/system/osi.py", line 289, in run_command
raise CommandException(cmd, out, err, rc)
system.exceptions.CommandException: Error running a command. cmd = /usr/bin/yum info installed -v rockstor*x86_64. rc = -11. stdout = ['']. stderr = ['']
During handling of the above exception, another exception occurred:
Traceback (most recent call last):
File "/opt/rockstor/src/rockstor/storageadmin/views/command.py", line 262, in post
return Response(rockstor_pkg_update_check(subscription=subo))
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/opt/rockstor/src/rockstor/system/pkg_mgmt.py", line 325, in rockstor_pkg_update_check
version, date = rpm_build_info(pkg)
^^^^^^^^^^^^^^^^^^^
File "/opt/rockstor/src/rockstor/system/pkg_mgmt.py", line 116, in rpm_build_info
if e.err[0] == emsg or e.err[-2] == emsg:
~~~~~^^^^
IndexError: list index out of range