disk gone

disk gone

[feng D]

hi:

Some time, the online machine will appear disk gone situation, although this is not very common, but we have too many machines, so still will encounter it.

That in the case of disk gone, /sys/block/bcache0/bcache/stop is not found, this case how to delete the bcache0?

There is if it is SSD gone, and how should deal with it?

We have multiple SATA disks under an SSD.

Re: disk gone

[Eric Wheeler]

On Fri, 2 Jun 2017, feng D wrote:

> hi:
> 
> Some time, the online machine will appear disk gone situation, although 
> this is not very common, but we have too many machines, so still will 
> encounter it.
> 
> That in the case of disk gone, /sys/block/bcache0/bcache/stop is not 
> found, this case how to delete the bcache0?

An interesting question.  Is it still available from the cache device in 
/sys/fs/bcache/* somewhere?  I'm guessing not, but check.

Do you have a procedure to reproduce this problem where /sys/block/bcache0 
has been hotplugged into oblivion?


--
Eric Wheeler



> 
> There is if it is SSD gone, and how should deal with it?
> 
> We have multiple SATA disks under an SSD.--
> To unsubscribe from this list: send the line "unsubscribe linux-bcache" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> 

Re: disk gone

[Kai Krakow]

Am Fri, 9 Jun 2017 01:46:24 +0000 (UTC)
schrieb Eric Wheeler <bcache@lists.ewheeler.net>:

> On Fri, 2 Jun 2017, feng D wrote:
> 
> > hi:
> > 
> > Some time, the online machine will appear disk gone situation,
> > although this is not very common, but we have too many machines, so
> > still will encounter it.
> > 
> > That in the case of disk gone, /sys/block/bcache0/bcache/stop is
> > not found, this case how to delete the bcache0?  
> 
> An interesting question.  Is it still available from the cache device
> in /sys/fs/bcache/* somewhere?  I'm guessing not, but check.
> 
> Do you have a procedure to reproduce this problem
> where /sys/block/bcache0 has been hotplugged into oblivion?

You easily run into the same situation when you stop the device instead
of detaching it. Bcache then doesn't release its cache device and
cannot be stopped/unregistered, because every interface for shutting
bcache down has vanished: It can only be reached through the backing
devices but the bcache interface below those is gone.

I think bcache needs a way to force unregister the cache device...
Otherwise the cache device stays busy and cannot be registered, neither
can any devices be attached again in this state.

-- 
Regards,
Kai

Replies to list-only preferred.

Re: disk gone

[Eric Wheeler]

On Sat, 10 Jun 2017, Kai Krakow wrote:
> Am Fri, 9 Jun 2017 01:46:24 +0000 (UTC)
> schrieb Eric Wheeler <bcache@lists.ewheeler.net>:
> 
> > On Fri, 2 Jun 2017, feng D wrote:
> > 
> > > hi:
> > > 
> > > Some time, the online machine will appear disk gone situation,
> > > although this is not very common, but we have too many machines, so
> > > still will encounter it.
> > > 
> > > That in the case of disk gone, /sys/block/bcache0/bcache/stop is
> > > not found, this case how to delete the bcache0?  
> > 
> > An interesting question.  Is it still available from the cache device
> > in /sys/fs/bcache/* somewhere?  I'm guessing not, but check.
> > 
> > Do you have a procedure to reproduce this problem
> > where /sys/block/bcache0 has been hotplugged into oblivion?
> 
> You easily run into the same situation when you stop the device instead
> of detaching it. Bcache then doesn't release its cache device and
> cannot be stopped/unregistered, because every interface for shutting
> bcache down has vanished: It can only be reached through the backing
> devices but the bcache interface below those is gone.
> 
> I think bcache needs a way to force unregister the cache device...
> Otherwise the cache device stays busy and cannot be registered, neither
> can any devices be attached again in this state.

I'm having trouble reproducing:

]# echo /dev/centos/bcache0 > /sys/fs/bcache/register
]# echo /dev/centos/bcache0-cache > /sys/fs/bcache/register
]# 
]# echo 1 > /sys/block/bcache0/bcache/stop
]# echo 1 > /sys/fs/bcache/3e57bc2e-4f01-4164-9b6d-54ef288ae394/stop
]#
]# echo /dev/centos/bcache0 > /sys/fs/bcache/register
]# echo /dev/centos/bcache0-cache > /sys/fs/bcache/register

also tried swapping the stop order:

]# echo /dev/centos/bcache0 > /sys/fs/bcache/register
]# echo /dev/centos/bcache0-cache > /sys/fs/bcache/register
]#
]# echo 1 > /sys/fs/bcache/3e57bc2e-4f01-4164-9b6d-54ef288ae394/stop   
]# echo 1 > /sys/block/bcache0/bcache/stop 
   >>> fails, no such file, but re-registers fine below
]#
]# echo /dev/centos/bcache0 > /sys/fs/bcache/register
]# echo /dev/centos/bcache0-cache > /sys/fs/bcache/register

What am I missing here?


--
Eric Wheeler



> 
> -- 
> Regards,
> Kai
> 
> Replies to list-only preferred.
> 
> --
> To unsubscribe from this list: send the line "unsubscribe linux-bcache" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
>