Thank you for the fast turn-around @awalkerix and the iX engineering team! And a special thank you to the community testers who helped us validate the fix was working properly in advance.
yeah also noticed it, itās kind of annoying. If you update from 24.04 to 24.10 i can understand the need to manually install the driver. But when installing a point release it seems unnessesary to always install the driver new. Canāt we get a promt before pressing the update button to ask if we want to keep the current nvidia driver?
It seems like the āUpdatesā are practically a complete install. The file size of the ISO vs the Update is pretty similar. Is this really a patch upgrade. If it is, then the majority of the software had been touched.
Way back when FreeNAS 9.3 came out and changed to using a live ZFS boot pool (vs. the run-from-RAMdisk that had been done through the 9.2 series), it was suggested this could result in much smaller downloads for updates, and possibly even not needing to reboot for every update. Obviously nothing came of either of those possibilities.
Omg I got crazy today, because since 24.10.0.1 Almost every file I copied from TrueNas to Windows over SMB is corrupted and on every copy the same file had a different checksum.
I thought at first it was the network card until I have realized that it had data errors over the 1Gb-NIC and on my second 10gb-NIC.
But I had no errors on my linux clients I guess.
Also files from MinoIO had no corruption.
I just installed 24.10.0.2 and the corruption seems to be goneā¦
Can someone confirm that this bug is fixed now?
Iām afraid to use SMB again.
Btw, the software Restic should not be used on linux with smb, because there was a bug in an older Golang version I guess?
This is something we still dream of doing, perhaps using apt to do much smaller updates in the future. But right now we donāt want to touch the updater, because if it aint broke, donāt fix it
My update was yet again, un-eventful. Couldnāt ask for much more.
Now to mention something that wasnāt really completely fixed:
SMART for NVMe. What looks to be fixed was the ability to āscheduleā a SMART test for NVMe drives. Now the other foot dropsā¦ but no SMART test is actually run. This gives a person false hope that a SMART test will be run on the NVMe drive. So I would have to say that the Release Notes are not clear on the status of SMART for NVMe.
HOWEVER! you can manually commence a test via the GUI. So part of it is functional.
I would however it will get closed immediately because smartmontools 7.5 is needed for a complete fix. I have built 7.5 smartctl, however smartd is still pending being fixed. The 7.5 smartctl work manually just fine. If people want to ensure NVMe drives are tested, they can run my little script or send ānvmecontrolā or ānvmeā commands in a Cron Job. Those two commands work flawlessly.
Once smartmontools 7.5 comes out and if the problem is not fixed, you can count on me to file another bug report. I need to look through the TrueNAS python code and also provide a solution that can just be rolled into TrueNAS SCALE (I learn this way). Since I do not expect CORE to get an update, folks are likely better off using my little script. I donāt like peddling this but it is an option for those folks who do have NVMe drives right now.
About 8,000 systems were updated to 24.10.0.2 on its first day with no particular issues. It has passed its first field safety test. Thanks for all the testing and feedback.
The next expected Electric Eel update is 24.10.1 - planned for December.
Why would any drivers need to be āinstalledā, independent of the NAS appliance itself?
How does this work with TrueNAS SCALE? Is there a āDriversā menu that allows you to install the Nvidia proprietary drivers, which donāt exist in the default base installation?
If so, that makes TrueNAS more modular than I had assumed.
Thereās a checkbox you can tick in the apps menu to download and install the drivers. If itās not checked, no driver gets installed and by default EE installs without a driver
My nvme doesnāt seem to be tested with your tool. It reports āLast Test Type (time conducted)ā as 76 (13127 hrs). Iāve tried running the GUI SMART test and it fails too. Weird. Any clues? It is a Crucial CT500P1SSD8 P1 500GB 3D NAND NVM PCIe M.2 (model number CT500P1SSD8) with serial number 2031E2BD763F.
With the GUI it reports the following and never progresses beyond 0.00%:
