FreeNAS suddenly won't boot anymore after 2 years

[hhornet]

Hi,

Although I have been using FreeNAS for about 2 years, I am still a newbie. I installed it according to the online documentation and it has been working ever since. But after not using it for few weeks, all of a sudden, the FreeNAS doesn't boot anymore (I always shut it down during the night). The boot sequence gets stuck at Trying to mount root from zfs:freenas-boot/ROOT/default []... and that's it. I have no idea what could be the issue and how to fix it.

The OS is installed on a 120GB SSD. I don't remember what version of FreeNAS I have, nor do I know how to find out.

Any help is much appreciated. Thanks

 

[hhornet]

Some more details about the issue: after enabling the verbose option, the last 4 lines shown on the console are:
(da0:umass-sim0:0:0:0): PREVENT ALLOW MEDIUM REMOVAL not supported.
(da1:umass-sim0:0:0:1): PREVENT ALLOW MEDIUM REMOVAL not supported.
(da2:umass-sim0:0:0:2): PREVENT ALLOW MEDIUM REMOVAL not supported.
(da3:umass-sim0:0:0:3): PREVENT ALLOW MEDIUM REMOVAL not supported.

Also updated the hardware details in my signature.

 

[Arwen]

It's likely that the 120GB SSD used for boot experienced a serious enough failure that it can't boot. Simply replace it and restore your configuration. That involves using a USB drive with an TrueNAS image on it, which you boot from. Then install to the new boot device. Whence you have the server booting, you restore the configuration.

You may be asked / shown that you have an older version of ZFS on your data pool.Don't upgrade your data pool's version, unless you want the extra features. Whence a pool is upgraded, it can't go back. And can't be used by older software.

Note that if you did not save / backup your configuration elsewhere, you would likely have to re-create it. But, unless your data pool disk, the 2TB one, also has serious errors, you should be able to import it too.

However, using a single disk in a ZFS pool is not recommended as it does not include redundancy. Further, WD Blue drives are also not recommended for NAS usage. But, each person makes their own choices.

 

[hhornet]

It's likely that the 120GB SSD used for boot experienced a serious enough failure that it can't boot. Simply replace it and restore your configuration. That involves using a USB drive with an TrueNAS image on it, which you boot from. Then install to the new boot device. Whence you have the server booting, you restore the configuration.

Thanks Arwen for your reply. I was thinking about the SSD failure, though it is one aprox. 1 year old and barely used - I only start the NAS when I want to watch a movie then shut it down for the night. Also I moved the SSD to my Win10 PC and checked it with SSDLife Pro and Hard Disk Sentinel and they both reported it as perfect condition (though TBH I don't know what kind of tests they did, as the results were reported instantaneous).

If the SSD is damaged, it is still under warranty and can take it back to vendor. But what reason do I give? Is there a way to surely determine the SSD state?

 

[NugentS]

If you have tested it and its OK - then just rebuild the NAS, Import the pool and apply the configuration.

Oh and WD have dedicated diagnostics for their SSD's that you can download. If they say its OK - then it probably is

 

[Arwen]

@hhornet In general, if a TrueNAS boot media starts having trouble, you would normally see it via zpool status, and an E-Mail warning.

However, if the part of the lost blocks are related to booting, then you can't get it to boot, to run the zpool status.


ZFS was specifically designed to handle un-expected power offs, with only the files, (or updates to files), in flight potentially being lost. However, some devices have a write cache that can lie to ZFS. So if ZFS thinks something is written and then writes more based on that assumption, with the device re-ordering the writes and prioritizing the new data over the old, ZFS pool corruption can occur. For those devices it's very good idea to disable any write cache.

That's unlikely to be your cause, but it happens with storage device firmware that does not follow standard practices. (Aka write barriers, write cache flush, etc...).


So, I too would try what @NugentS suggested. What is meant by "rebuild the NAS", is reload the OS, NOT the data disks. Simply import the data pool after re-loading the TrueNAS / OS software.

 

[hhornet]

If you have tested it and its OK - then just rebuild the NAS, Import the pool and apply the configuration.

Oh and WD have dedicated diagnostics for their SSD's that you can download. If they say its OK - then it probably is

I ran the WD Dashboard test as well and the report is all green, showing 100% life remaining. I did a full erase with the tool and will try to reinstall FreeNAS on it. In the meantime, since I wanted to check that the SATA controller on my MB is OK, I reinstalled FreeNAS on a USB stick and everything went smooth. I even reimported the data pool from the 2nd HDD and my NAS is back. So the MB is fine and everything seems to point to the SSD.

Thanks for suggestions

 

[hhornet]

Small update: finally my FreeNAS system is back online. After erasing the SSD I managed to reinstall the system on it, and imported the data pool.

Thanks for helping!