Problem after updating to 9.2.0 x86

[ManolisK74]

Hello everybody,

I just updated my system from 9.1.1 to 9.2.0 and even though it used to work fine, now it won't boot.

On the console I get:

Mounting local file systems: .

Fatal trap 12: page fault while in kernel mode
cpuid = 0: apic id = 00
fault virtual address = 0x11
fault code = supervisor read, page not present
instruction pointer = 0x20:0x88a195b4
stack pointer = 0x28:0xc6aa6a60
frame pointer = 0x28:0xc6aa6a84
code segment = base 0x0, limit 0xfffff, type 0x1b
= DPL 0, pres 1, def32 1, gran 1
processor eflags = interrupt enabled, resume, IOPL = 0
current process = 315 (txg_thread_enter)
[ thread pid 315 tid 100529 ]
Stopped at range_tree_vacate+0x24: calll *0x10 (%edx)

any ideas ? ?

 

[Yatti420]

What are your specs? You sure the USB install is good?

 

[ManolisK74]

It was working fine before the update, on version 8 initialy and still fine with 9.1.1, although the machine is quite old and rather 'weak' .. running with an AMD Sempron processor with four disks on the onboard controller and four more on a 3ware sata controller.

 

[Yatti420]

What motherboard / ram setup etc.. Using UFS or ZFS?

 

[ManolisK74]

Using ZFS, Motherboard is ASUS A8R MVP with 2 x 512 RAM

 

[gpsguy]

RTFM .

ZFS on FreeNAS requires a minimum of 8Gb of RAM.

We've had a couple of use's (just this week) loose their pool due to running it with too little memory.


Sent from my phone

 

[cyberjock]

I believe the current number is 3 that I have personally diagnosed.

 

[Mguilicutty]

Beyond your obvious lack of RAM, you would need someone wiser than I to interpret your error message. However, if it were me I would try doing a fresh install on a new usb stick with all the drives disconnected. If you can boot into that then hook all your drives back up and try to import your pools back in. Read up on auto import in the manual and here in the forums. You'll lose your configuration this way but I'm guessing your setup isn't too complex. You may have issues importing the pools since they were not exported properly (though maybe a shutdown does this anyway, so it may not be an issue) but you can read up on that here as well.

 

[cyberjock]

I believe the current number is 3 that I have personally diagnosed.

And now that I see he has 1GB of RAM, I'm thinking we're at #4 now...

 

[ManolisK74]

What if I do a fresh install using the previously working version (9.1.1) ? Do you think it will work ? I don't care about the configuration that much anyway

 

[Yatti420]

I would put in atleast 4gb of tested ram.. Whatever the max is first.. Use 9.1.1 or whatever was crashing.. You will want to get the data off and run UFS (which requires a minimum of 2GB ram) from now on or do a big hardware upgrade.. Good luck!..

 

[ManolisK74]

I got it to 3gb (thats all I could find lying arround)
the original usb stick still gives the same arror
I created a new with a clean install and it stops at:
Waiting up to 5 seconds for ixdiagnose to finish... done.

 

[ManolisK74]

I also tried removing the disks and booting with the old stick, and it gets past the inital error, but also stops at:
Waiting up to 5 seconds for ixdiagnose to finish... done.

 

[Yatti420]

The ixdiagnose takes awhile on first boot.. Are you sure it's frozen there? The ram is tested right?

 

[Mguilicutty]

I believe ixdiagnose sends crash reports to the devs, it is not something that should be dying. Anybody know what comes directly after this? At this point, I'm thinking hardware issue. Try another fresh install without the original memory inserted, perhaps that is the issue. If they updated any ZFS features and you upgraded your pools when you went to 9.2 then 9.1 will no longer mount them.

In any case, a new server should be in your plans if your data is important to you and you like having a server.

 

[Yatti420]

If you do manage to recover get everything off.. Make sure you setup new UFS volumes/partitions instead of using ZFS..

 

[ManolisK74]

First of all, thanks for all the help,

It did boot after a while (using a clean 9.1.1 x64 install) and I was able to automaticaly import the volume with all data intact.
I just reconfigured the network and shares and everything looks fine.

This server is used just for backing up workstations data so it is not that crucial (at least as long as it doesn't die same time with a workstation !)

So I guess I'll be leaving it as is (except that it now has 3 instead of 1 gb of ram ...) the disk capacity isn't that large anyway, it has 4x500+5x750 disks.

 

[Knowltey]

Hon, x86 doesn't meet the minimum requirements for ZFS.

 

[cyberjock]

You should seriously consider upgrading RAM to 8GB. I've already worked with 4 people this week that have lost their pools from <8GB of RAM with ZFS. You are treading on thin ice using FreeNAS and ZFS in the fashion in which you are.

 

[Knowltey]

You should seriously consider upgrading RAM to 8GB. I've already worked with 4 people this week that have lost their pools from <8GB of RAM with ZFS. You are treading on thin ice using FreeNAS and ZFS in the fashion in which you are.

Well he can't unfortunately since he's running 32-bit

just for backing up

not that crucial

Wow, contradiction of the year.

at least as long as it doesn't die same time with a workstation !

This happens, trust me. I have had a backup server (hard drive in one rather) die during the process of pulling the backups off of the server. Only reason I managed to recover them was I had to stick a heat sink and fan on the drive while pulling only the most important data since it was a severe overheating issue.