XigmaNAS

Hello again! I am in a bit of a spot here with my production NAS4Free setup and need some guidance.

I have a single 6 disk ZFS Pool and one of the disks has failed. I found a walkthrough for replacing the disk, but the first step is to venture into ZFS>Pools>Tools> and "offline" the failed disk. Unfortunately there are only 4 disks listed in this screen, not six. how do I get the interface to show all the drives so I can "offline" this bad drive? Other alternatives?

I think I figured this out from the CLI...

zpool list # will give you a list of drives and zPools, in my case I had a drive listed as "ada2" that was in the "REMOVED" status.

zpool offline zPool_01 ada2 # this changed the status of my drive in zpool list from "REMOVED" to "OFFLINE". Guess I can continue with the tutorial.. Wonder why the GUI didn't know about ada2... Hmm..

Worst case just power down and swap the drive.

ZFS is truly incredible, it may just pull in the new drive and start the resilver. Or, you may need to issue the replace command.

It sure didn't feel truly incredible.. When the drive failed it made the entire iscsi target unavailable until the system was restarted. Afterward it seemed to fail the drive and function as it should, but I could have done without rushing to the site......

please post zpool status.

if you have raidZ1 , one drive fail DO NOT make innaccesible your data, only a bit slow, so I suspect that you have a simple ZFS pool wihtout redundancy.

Please post a screen capture of your detected disk and ZPool detected & current config too.

Hi there. I am sure I do NOT have a simple ZFS pool. It consists of 3 different Mirrors. I did this to achieve adequate redundancy as well as performance. See a current zpool status command below. When the failure occurred the "ada2" device was listed as "REMOVED" and the pool was in a degraded state. However the iSCSI target was completely inaccessible until the NAS4Free server was restarted. After restarting the server the "ada2" drive seemed to come back online for a short time and the resilver process completed.. Then it shortly after went into the "REMOVED" status again, however the 2nd time the iSCSI target remained available.

I shut the NAS4Free server down last night and physically replaced the "ada2" device with a new drive. Once the server came back online I did an import from the drive management screen, then ran the replace command to bring everything back online and get the "mirror-2" healthy again.

What I do not understand is why the entire target was unavailable. That shouldn't have happened. I also don't get why the drive failed. I have since put this brand new drive into a different system and it appears to run just fine. I have done quite a few tests on it, no issues yet.

Immediately after restarting the NAS4Free box the first time while the "bad" drive was still installed, the zpool status command showed a "1" value under CKSUM for "ada2".. What does this mean? This drive is a Western Digital 1.0 TB Black (WD1003FZEX).. Is this drive not compatible for some reason? It doesn't automatically spin down or something silly making the NAS think it is bad?

nas4free: ~ # zpool status
pool: zPool_01
state: ONLINE
scan: resilvered 127G in 0h49m with 0 errors on Thu Jun 25 15:27:19 2015
config:

NAME STATE READ WRITE CKSUM
zPool_01 ONLINE 0 0 0
mirror-0 ONLINE 0 0 0
ada0 ONLINE 0 0 0
ada1 ONLINE 0 0 0
mirror-1 ONLINE 0 0 0
ada3 ONLINE 0 0 0
ada4 ONLINE 0 0 0
mirror-2 ONLINE 0 0 0
ada2 ONLINE 0 0 0
ada5 ONLINE 0 0 0

errors: No known data errors

actyler1001 wrote:What I do not understand is why the entire target was unavailable. That shouldn't have happened.

Yes. I agree. But no experience here in iSCSI so I can't help , sorry.

actyler1001 wrote:However the iSCSI target was completely inaccessible until the NAS4Free server was restarted.

actyler1001 wrote:What I do not understand is why the entire target was unavailable. That shouldn't have happened. I also don't get why the drive failed.

This is not surprised thing if the drive is not respond long time.
ZFS request is locked until first is finished.
As a result upper layer cause timeout.
SCSI device on OS(you means windows?) -> iSCSI Initiator -> iSCSI Target -> ZFS -> broken physical device.

There are two timers. One is SCSI request timer, other is iSCSI request timer.

You might need MPIO and adjust timeout value.
https://technet.microsoft.com/en-us/lib ... s.10).aspx

daoyama,

Thanks for your reply. Not sure I follow your description however. In regard to MPIO, I am actually not connecting the iSCSI target to a Windows server. It is connected to multiple VMware ESXi 5.5 servers and yes, I do have multiple network interfaces with multiple paths configured to the target.

That being said, do you have any additional input on why the iSCSI target would become unavailable due to a single drive failure?