XigmaNAS

I thought I had my system up and running. This morning and now tonight the NAS has become unreachable. I am not able to ping it, access the web interface or my network location. When I go to the NAS it is still on and I don't see anything wrong with it. When I restart it comes back up fine except the mount point has an error. To fix this error I run fsck which allows me to retry next to the error on my mount point and it lets me access it again. I am just trying to figure out why it crashes and if it is the cause of the mount point error. I attached the system log when it boots.

NAS4Free version 9.1.0.1
x64-embedded on AMD Phenom(tm) II X4 955 Processor 3.2 GHz, 2 GB of RAM
Raid 5 setup with 3x2TB WD Red Drives

The only services I have setup so far is CIFS/SMB and AFP.

Here's the log after being on for a little bit. I noticed some errors and core dumps but I don't really know how to read these files yet.

Hi Takachsin,

I see 2 things:

1. Looks like you are using SMB2, disable AIO if you are and see if errors stop.
2. Looks like you are not using ZFS, uncheck the snapshot option in CIFS/SMB config.

Takachsin wrote:I am just trying to figure out why it crashes and if it is the cause of the mount point error.

CIFS/SMB appears to be crashing. Yes, this crash is causing the mount point error and requiring fsck. Similar to a power outage.

Regards,
Al

al562 wrote:Hi Takachsin,

I see 2 things:

1. Looks like you are using SMB2, disable AIO if you are and see if errors stop.
2. Looks like you are not using ZFS, uncheck the snapshot option in CIFS/SMB config.

Takachsin wrote:I am just trying to figure out why it crashes and if it is the cause of the mount point error.

CIFS/SMB appears to be crashing. Yes, this crash is causing the mount point error and requiring fsck. Similar to a power outage.

Regards,
Al

It was the AIO and ZFS stuff, I read somewhere to increase performance and I obviously didn't understand it all the way. I'll just be more careful in the future. Thank you!

In the latest build (621+) AIO seems to be working again. I have put about 1Tb of data through it with no problems.



All these tests had AIO enabled.

Also this is a test of CIFS with and without AIO enabled

Just FYI, it seems to help with rewrites more than anything

I think it was a faulty samba release, because with 621 i can get MS AD to work again, too.