Disk Writes sometimes Interrupted?

[Grimm Spector]

Sometimes when writing things to disk, such as copying between locations in the ZFS volume, or moving items across the network, starting a new disk write, such as saving a new file, will interrupt other processes either locally or remotely that are also writing files, causing them to fail. If I immediately restart such processes they continue (though they start over), and finish fine. It's very odd, and I'm not sure what the cause would be. Hopefully someone can help me figure this out, as it's pretty useless if it can only do one disk write process at a time. I've got 4 disks in a zvol, nothing particularly fancy about it that I know of.

Edit:

Motherboard: Tyan S7067GM1NR-1T(BTO)
CPU: 2x Intel E5-2630 v2
RAM: 48 GB ECC
HDD: 4x SUN 900 GB SAS HUC09090CSS600 in 1 vdev with raidz1
HDD Controller: LSI onboard controller set to no RAID/control, should be pass thru only
Network Cards: Onboard 10 GbE, plus 1 4 Port Intel Gigabit LAN Server card

 

[Ericloewe]

I've got 4 disks in a zvol

That's not what "zvol" means. You have a pool, of unspecified geometry, with four disks.

As for your problem, please provide the information that the forum rules ask you to provide, since we have very little to go on based on what you've told us so far.

 

[Grimm Spector]

That's not what "zvol" means. You have a pool, of unspecified geometry, with four disks.

As for your problem, please provide the information that the forum rules ask you to provide, since we have very little to go on based on what you've told us so far.

My bad, I have a zvol with 1 vdev with 4 disks in raidz1 in it. Better? I've updated the original post with the requested information.

 

[Ericloewe]

I have a zvol

No, you have a pool. A zvol is something completely different, it's a block device backed by ZFS.

Motherboard: Tyan S7067GM1NR-1T(BTO)
CPU: 2x Intel E5-2630 v2
RAM: 48 GB ECC
HDD: 4x SUN 900 GB SAS HUC09090CSS600 in 1 vdev with raidz1
HDD Controller: LSI onboard controller set to no RAID/control, should be pass thru only
Network Cards: Onboard 10 GbE, plus 1 4 Port Intel Gigabit LAN Server card

No glaring issues there, but you forgot to mention the FreeNAS version. Assuming it's halfway recent (if it isn't, please try 11.1-U5), we've got a weird one on our hands. Do the logs say anything when you observe this behavior?

 

[Grimm Spector]

No, you have a pool. A zvol is something completely different, it's a block device backed by ZFS.


No glaring issues there, but you forgot to mention the FreeNAS version. Assuming it's halfway recent (if it isn't, please try 11.1-U5), we've got a weird one on our hands. Do the logs say anything when you observe this behavior?

Oops, sorry about that. I'm using 11.1-U4 currently. The logs say nothing unusual, when using the share I see a lot of lines just like this:

Code:

kernel: pid 97865 (smbd), uid 0: exited on signal 6 {core dumped}

Hope that's normal.

 

[Ericloewe]

That's decidedly not normal, that's Samba crashing. I'd recommend filing a bug report (please post the issue number here afterwards) - include a debug tarball, which you can do by submitting the bug from the FreeNAS GUI.

 

[Grimm Spector]

That's decidedly not normal, that's Samba crashing. I'd recommend filing a bug report (please post the issue number here afterwards) - include a debug tarball, which you can do by submitting the bug from the FreeNAS GUI.

Will do, is there a specific log file I should be including? I'm new to FreeNAS and am not sure where/what log files are for what at this stage.

 

[Redcoat]

The tarball will be created and attached automatically when you submit the bug report from the FreeNAS GUI.

 

[Grimm Spector]

The tarball will be created and attached automatically when you submit the bug report from the FreeNAS GUI.

Excellent, that saves a lot mucking around. It's been filed under Bug #34540. Hopefully it's something fixable.