[solved] FreeNAS crashing ... whats the Problem?

[Moe5k]

OK, i banned the Fasttrack-controller and now i started a 48 hour stresstest while permanent read / write data to the System.
I do only use the onboard-controller of the S5000VSA-Board. This controller has only 6 ports.
If the actual stresstest runs without problems i do plan to buy a M1015 controller to increase the nummer of sata-ports.
Does anybody know about a mix of these controllertypes?

 

[Moe5k]

OK, the system crashed again while the stresstest with ONLY the onboard-controller
This is an Intel® 6321 ESB I/O Controller.
Are there any experiences with that controller or that S5000VSA Mainboard?
In this case i copied a lot of single files and directorys ( about 2 mio.)

 

[Moe5k]

Now the system is not rebooting ... it seems like the pendrive with the system is dead ...
Are the generally problems with FreeNAS on pendrives?
What i mean is ... maybe the problems with system-crashing is the pendrive`?

 

[Moe5k]

It seems the pendrive was the problem...
Now i installaed FreeNAS on a usb-hdd and from this moment
i had no more crashes ...and i made a lot of stresstests.

Are there any know problems wiht installations on usb?
Because before that another pendrive-installation (on another pendrive)
gone also faulty...

 

[danb35]

On a USB stick is the recommended way to install FreeNAS. They do eventually wear out, particularly with a lot of write activity. However, in ordinary use, FreeNAS doesn't write to the boot device very often at all, so it's not likely to be a problem provided a good-quality USB stick is used.

 

[Moe5k]

it was a "Verbatim Store n Go" with 8GB. Its not the fastest one but i think not a so bad quality... Two Sticks of these seems to made problems...

 

[Moe5k]

Hello again,

my server now ran for a week perfect, stable and fine... i transfered 5 or 6 TB over the last days without any problems or failure-messages.

But today my iSCSI device (mounted on a Win2008Server) disappeared.
FreeNAS answered to my ping but couldn´t log-in via ssh (typed username und password and nothing happend).
So i gone to the physical machine and saw the following:

adaasync: Unable to attach to new device due to status 0x6
cam_periph_alloc: attempt to re-allocate valid device ada3 rejected flags 0x118
adaasync: Unable to attach to new device due to status 0x6
cam_periph_alloc: attempt to re-allocate valid device ada4 rejected flags 0x118
adaasync: Unable to attach to new device due to status 0x6
cam_periph_alloc: attempt to re-allocate valid device ada5 rejected flags 0x118

these messages repeated sometimes...

I´ve tried to restart the machine (systems reaction was extremly slow)
it seems to made a reboot but i doesn´t ... so after 3 tries i made a cold shutdown.

The system booted fine after this, but when normally comes the menu after reboot it
displays the same failure-messages...

What could be the problem? What can i do to fix it?

thx moe

 

[Moe5k]

Okay, i fixed it:
1. switched the machine off again
2. removed the three hardsdisks which has the failures
3. booted the machine and hotplugged the 3 harddisks
4. removed the old volume (which displayed as "Unknown") WITHOUT deleting the file
5. autoimported my old volume
6. restarted iSCSI service

System runs now fine again...

But what could caused the problem before?

 

[jgreco]

Bad cabling? Problems with the controller? Hard to know. It suggests a hardware issue. The time to fix issues like this is during build and burn-in, not once you've got production data on the system, because many of the things that you'd want to do to help identify the issue are incompatible with a running pool.

 

[Moe5k]

thx for your post!
the 3 harddisks regarding this problem are connected to the onboard-controller but there are
also other 3 hdds conntected to the same controller without problems.
i saw that the 3 problem-hdds are connected the same 3-bay backplane. the other 3 hdds are
connected to another backplane ... so i think its a good idea to change the backplane...
btw: no critical data are in this pool at this time
i will report ...next week

 

[jgreco]

Check the power cabling to the backplane.

 

[mjws00]

Is that Intel® 6321 ESB I/O Controller fake raid? I wouldn't rule it out either... maybe someone loves it under bsd (it is on an intel board). But if even a firmware driver version issue can and does drop drives. It makes me wonder. I'd have a 1015 swapped in on the bench to test with before I ordered a backplane. Unless they are sitting in a cheap hotwap cage/case.

Hopefully it is just a cable or something. Good luck.

 

[Moe5k]

Thx for your answers ...
5 hours later the same problem occurred, so i changed the -possibly faulty- backplane and removed a Y-cable for the power-connection to the backplane ...
Now i will test the system again ...

 

[Moe5k]

so, now i made 14-days burn-in tests without any problems ... i think the backplane/cable made problems