Hello people,

We have a problem and I hope someone can help us out here.

One of our iscsi stacks had a fault. So the disk was gone for a couple
of minutes.

Now everything is back up. But when we want to access the disk we see
the folowing with the format command:
We can choose the disk, but it says "not formatted"

4. c2t0690A018E0F14991125844847700D063d0 <drive not available>
   /scsi_vhci/ssd@g0690a018e0f14991125844847700d063

In messages file :
Aug  8 21:35:51 srv1 scsi: [ID 107833 kern.warning] WARNING:
/scsi_vhci/ssd@g0690a018e0f1499112
5844847700d063 (ssd0):
Aug  8 21:35:51 srv1  offline or reservation conflict

We have not rebooted the machine because this would cause a lot of
trouble for our customers, so we want to avoid this.

What is this and how can we make this disk available again ?

We tried : "devfsadm -C" but no luck.

Anyone ?

Greetings... Richard

On Aug 8, 12:41 pm, "R.Smits" <smi...@gmail.com> wrote:

> We have a problem and I hope someone can help us out here.
> One of our iscsi stacks had a fault. So the disk was gone for a couple
> of minutes.
> Now everything is back up. But when we want to access the disk we see
> the folowing with the format command:
> We can choose the disk, but it says "not formatted"
> 4. c2t0690A018E0F14991125844847700D063d0 <drive not available>
>    /scsi_vhci/ssd@g0690a018e0f14991125844847700d063
> In messages file :
> Aug  8 21:35:51 srv1 scsi: [ID 107833 kern.warning] WARNING:
> /scsi_vhci/ssd@g0690a018e0f1499112
> 5844847700d063 (ssd0):
> Aug  8 21:35:51 srv1  offline or reservation conflict
> We have not rebooted the machine because this would cause a lot of
> trouble for our customers, so we want to avoid this.

But its OK to use iSCSI ... Is it really mature enough at this point I
wonder
to dump that on a customers lap...? Perhaps you are finding out!

> What is this and how can we make this disk available again ?
> We tried : "devfsadm -C" but no luck.

Thats cleanup mode. I dont think that rebuilds anything.

Maybe devfsadm -v will do something, but I had a similar problem last
week
where FC/AL disks would simply disappear from the pool randomly.
a reboot -- -r got them back. Sometimes devfsadm was enough but not
always.
But they'd go away again later.

Make sure the /dev entry actually exists. Mine were getting wiped
occasionally.

This was on legacy equipment going through a 1 GB FC/AL switch.
I finally gave up on the switch and HBA driver and direct attached
instead. No more issues.
In my case Im not sure if it was a buggy old driver that croaks on
Solaris 10, or a misconfigured
switch or what. You may have your work cut out for you.

R.Smits wrote:
> Hello people,
> 
> We have a problem and I hope someone can help us out here.
> 
> One of our iscsi stacks had a fault. So the disk was gone for a couple
> of minutes.
> 
> Now everything is back up. But when we want to access the disk we see
> the folowing with the format command:
> We can choose the disk, but it says "not formatted"
> 
> 4. c2t0690A018E0F14991125844847700D063d0 <drive not available>
>    /scsi_vhci/ssd@g0690a018e0f14991125844847700d063
> 
> In messages file :
> Aug  8 21:35:51 srv1 scsi: [ID 107833 kern.warning] WARNING:
> /scsi_vhci/ssd@g0690a018e0f1499112
> 5844847700d063 (ssd0):
> Aug  8 21:35:51 srv1  offline or reservation conflict
> 
> We have not rebooted the machine because this would cause a lot of
> trouble for our customers, so we want to avoid this.
> 
> What is this and how can we make this disk available again ?
> 
> We tried : "devfsadm -C" but no luck.
> 
> Anyone ?
> 
> Greetings... Richard

It looks to me as if the disk has become corrupted somehow or there is 
still a hardware problem.  Can you replace the disk drive and restore 
from backup?

There is much to be said for both RAID and making regular backups. 
Both, IMHO, are essential to the integrity and availability of your data!

There is also much to be said for having spare disk drives available. 
If you have a really expensive field service contract, they can probably 
get a new drive to you in a couple of hours.  If you stock your own 
spares, that becomes a couple of minutes!

Richard B. Gilbert wrote:
> R.Smits wrote:
>> Hello people,
>>
>> We have a problem and I hope someone can help us out here.
>>
>> One of our iscsi stacks had a fault. So the disk was gone for a couple
>> of minutes.
>>
>> Now everything is back up. But when we want to access the disk we see
>> the folowing with the format command:
>> We can choose the disk, but it says "not formatted"
>>
>> 4. c2t0690A018E0F14991125844847700D063d0 <drive not available>
>>    /scsi_vhci/ssd@g0690a018e0f14991125844847700d063
>>
>> In messages file :
>> Aug  8 21:35:51 srv1 scsi: [ID 107833 kern.warning] WARNING:
>> /scsi_vhci/ssd@g0690a018e0f1499112
>> 5844847700d063 (ssd0):
>> Aug  8 21:35:51 srv1  offline or reservation conflict
>>
>> We have not rebooted the machine because this would cause a lot of
>> trouble for our customers, so we want to avoid this.
>>
>> What is this and how can we make this disk available again ?
>>
>> We tried : "devfsadm -C" but no luck.
>>
>> Anyone ?
>>
>> Greetings... Richard
> 
> It looks to me as if the disk has become corrupted somehow or there is
> still a hardware problem.  Can you replace the disk drive and restore
> from backup?
> 
> There is much to be said for both RAID and making regular backups. Both,
> IMHO, are essential to the integrity and availability of your data!
> 
> There is also much to be said for having spare disk drives available. If
> you have a really expensive field service contract, they can probably
> get a new drive to you in a couple of hours.  If you stock your own
> spares, that becomes a couple of minutes!
> 
Hello,

Yes, no problem with the backups/spares etc...
This disk was part of an active mirrorset (raid 1) so the other disk has
taken over. No one is having problems at this time. But we want to
rebuild the mirror.

Maybe a reboot is the answer, but we have to schedule this for a
maintenance window.
We will also try to try some rescan commands with iscsiadm.

Greetings.. Richard

Hi there!

Richard,

so what was the cause and the solution??I have the same problem now
and I have to fix it but have no idea how to cope with it :/ any
advices??

br
slaw


On Aug 8, 11:20 pm, "R.Smits" <smi...@gmail.com> wrote:
> Richard B. Gilbert wrote:
> > R.Smits wrote:
> >> Hello people,
>
> >> We have a problem and I hope someone can help us out here.
>
> >> One of our iscsi stacks had a fault. So the disk was gone for a couple
> >> of minutes.
>
> >> Now everything is back up. But when we want to access the disk we see
> >> the folowing with the format command:
> >> We can choose the disk, but it says "not formatted"
>
> >> 4. c2t0690A018E0F14991125844847700D063d0 <drive not available>
> >>    /scsi_vhci/ssd@g0690a018e0f14991125844847700d063
>
> >> In messages file :
> >> Aug  8 21:35:51 srv1 scsi: [ID 107833 kern.warning] WARNING:
> >> /scsi_vhci/ssd@g0690a018e0f1499112
> >> 5844847700d063 (ssd0):
> >> Aug  8 21:35:51 srv1  offline or reservation conflict
>
> >> We have not rebooted the machine because this would cause a lot of
> >> trouble for our customers, so we want to avoid this.
>
> >> What is this and how can we make this disk available again ?
>
> >> We tried : "devfsadm -C" but no luck.
>
> >> Anyone ?
>
> >> Greetings... Richard
>
> > It looks to me as if the disk has become corrupted somehow or there is
> > still a hardware problem.  Can you replace the disk drive and restore
> > from backup?
>
> > There is much to be said for both RAID and making regular backups. Both,
> > IMHO, are essential to the integrity and availability of your data!
>
> > There is also much to be said for having spare disk drives available. If
> > you have a really expensive field service contract, they can probably
> > get a new drive to you in a couple of hours.  If you stock your own
> > spares, that becomes a couple of minutes!
>
> Hello,
>
> Yes, no problem with the backups/spares etc...
> This disk was part of an active mirrorset (raid 1) so the other disk has
> taken over. No one is having problems at this time. But we want to
> rebuild the mirror.
>
> Maybe a reboot is the answer, but we have to schedule this for a
> maintenance window.
> We will also try to try some rescan commands with iscsiadm.
>
> Greetings.. Richard

In article <YvydnRaPRMfSgSfbRVnygAA@infopact.nl>,
R.Smits <smitsr@gmail.com> wrote:
>One of our iscsi stacks had a fault. So the disk was gone for a couple
>of minutes.
>
>Now everything is back up. But when we want to access the disk we see
>the folowing with the format command:
>We can choose the disk, but it says "not formatted"
>
>4. c2t0690A018E0F14991125844847700D063d0 <drive not available>
>   /scsi_vhci/ssd@g0690a018e0f14991125844847700d063
>
>In messages file :
>Aug  8 21:35:51 srv1 scsi: [ID 107833 kern.warning] WARNING:
>/scsi_vhci/ssd@g0690a018e0f1499112
>5844847700d063 (ssd0):
>Aug  8 21:35:51 srv1  offline or reservation conflict
>
>We have not rebooted the machine because this would cause a lot of
>trouble for our customers, so we want to avoid this.
>
>What is this and how can we make this disk available again ?
>
>We tried : "devfsadm -C" but no luck.

I wouldn't really expect the above to work - at least, not by itself.

After you experience a failure of the iSCSI server, you need to ensure that
the iSCSI client (in this case, the system that is displaying the "drive
not available" message) attempts to log back into the iSCSI server to
re-establish the iSCSI SAN connection. If this doesn't happen, the LUN will
remain in an offline state. If the connection is still offline (as seems to
be indicated by your log messages, above), you can run `devfsadm` until
you're blue in the face and it won't help as there are no device nodes
available to configure. You need to verify that you have actually restored
the iSCSI initiator's connection to the remote LUN (this may require a
rerunning of the remote LUN discovery - dunno for sure). Once you've
restored the connection to the LUN, things *should* go back to normal.

-- 

"You can only be -so- accurate with a claw-hammer."  --me