New installation, Storage not visible, folders unreachable.

[DarkCorner]

I had already opened this thread, for CPU always at 100% and errors on a disk, but I have now closed it because other problems arose with a new installation.

Being a backup server, I was able to delete all content, remove shared folders, datasets and pools.
Then I did a fresh install with the latest available ISO (TrueNAS-SCALE-22.12.3.3.iso).

Now, the CPU always has the wrong number of cores, but this also happens on 3 other NAS with the same motherboard and the same type of CPU.
I opened a separate thread for this issue.

At this point, after having recreated pools, datasets and shared folders, I found the CPU with an average load of 4% and the memory almost entirely free.
I then activated hourly and daily sbapshot procedures.
I haven't installed any apps.

Finding confirmation that after a few days everything was still fine, last night I launched some backups and moved the previous zip files to a shared folder.

At 3am everything seems OK.
This morning at 8am:

• The CPU has an average load of 55%, never drops below 50% and with peaks at 97%.
• The 9.2GB memory is divided with:
  • Free: 1.4GB
  • ZFS Cache: 4.6GB (it was only 1.8GB in the old installation with the most occupied storage; now there are about 200 zip files, I don't have precise data because it is now inaccessible; see notes below).
  • Services: 3.2GB
• In the dasboard, the space used in the pool is only 9%.
• If I select the Pool menu, or I see the blue wheel spinning continuously or with other attempts I don't even see a pool or even a dataset.
• If I try to view the discs I see the blue wheel spinning continuously. Same with shared folders and services.
• The shell is totally black.
• To see the services I have to reboot and immediately enter the menu (SMB is enabled). If I enter the shell from here, I only see the introductory text and after "Welcome to TrueNAS" everything is blocked.
• The last two backups failed because the shared folder was missing.
• If from Windows I type \\IP-Address it doesn't even find the server, but it responds to the ping.
• Unlike the previous installation, I don't see any error messages in the dashboard.

This is the NAS configuration:
For those who are ready to object that it is an old configuration, I would like to point out that:

1. the NAS was built when 2 cores was the standard;
2. which was rarely used until a few months ago (and almost it was always turned off);
3. that the office sees only 4 users who use the NAS only for backup at different times, with little overlap.

• TrueNAS version: TrueNAS-SCALE-22.12.3.3
• Motherboard: ASRock FM2A85X-ITX
• CPU: AMD A4-5300 APU with Radeon(tm) HD Graphics
• Memory: 9.2GB DDR3
• 3 work HDD WD Red 3TB (WDC_WD30EFRX) with 1 Pool RAIDZ
• 1 Boot SSD Crucial CT120BX500SSD1
• 1 Apps SSD Crucial CT120BX500SSD1 (No apps installed)

The installation is the same as the previous one; the difference is that in that case there was an update to the new version while now it is a clean install.
I am surprised by a completely different situation, without errors, but with the NAS completely blocked.

Thanks in advance for the advice, but without being able to access the menu and the shell I fear I will have to do a new installation and I don't think adding a little RAM makes a difference.

 

[Patrick M. Hausen]

If you don't run apps and essentially want a storage appliance, have you considered running TrueNAS CORE?

 

[DarkCorner]

@Patrick M. Hausen
Initially they were all "Core", then I migrated to SCALE because it seemed like it was the best choice.
I don't have any apps because in the previous installation it was impossible to configure a simple app to do synchronization or backups.

The point, however, is another.
At 1.45pm I connect and now the CPU varies between 5% and 13%; pools and datasets are visible again, shared folders are still accessible.
Furthermore, ZFS Cache has shrunk to 1.3GB.
Without me having done anything.

What could have blocked TrueNAS for hours?

Now that the shared folders are accessible, 1517 files have been created in 51 folders for a total of 608GB (500GB on disk).
If it takes so little to make it unmanageable we are truly in bad shape.

 

[Patrick M. Hausen]

I have no idea. It's SCALE. It's experimental pre-release software. With known memory management issues. If I wanted to build a storage appliance I would use CORE.

Software Status - TrueNAS Roadmap - Open Source NAS Software

Get up-to-date insight on the TrueNAS roadmap. Explore the TrueNAS project timeline and stay informed with the latest updates.

www.truenas.com

 

[sfatula]

The point, however, is another.
At 1.45pm I connect and now the CPU varies between 5% and 13%; pools and datasets are visible again, shared folders are still accessible.
Furthermore, ZFS Cache has shrunk to 1.3GB.
Without me having done anything.

What could have blocked TrueNAS for hours?

You have minimal ram, bare minimum, and very minimal CPU. Didn't know it even ran with < 16GB ram (ok, min specified in system requirements is 8GB but I can't imagine that working very well) zfs consumes 50% of ram by default on Scale, it will adjust usage as memory pressure builds or reduces as needed (though not perfectly on Linux and with a 2 core CPU, I can imagine this adjustment falling behind), this is normal. I wouldn't really run Scale on such small memory and 2 core though it does barely meet the minimum requirements. I would expect spotty performance based on whatever you are doing at the time. If you are trying to access the UI after a reboot, I would expect terrible performance with a 2 core processor and such limited memory as many things happen on startup.

You might be able to squeeze by for just backups but this is a very minimal setup, I'd certainly disable as many services as possible that you don't need. To help figure out what happened, you should be scouring the various logs in /var/log to see what details you can find there, there may be some hints during the minimum response times period you noted.

I do think Core would work better for such a minimal system. I find Scale to be super reliable, but I am not running a minimal system.

 

[ChrisRJ]

Initially they were all "Core", then I migrated to SCALE because it seemed like it was the best choice.

Why?