[dm-devel] Device Access Question

[Edward Goggin]

I suggest you use the dm-multipath mapped device names instead of the
target device ones.  Otherwise, you can run into problems caused by the
fact that the cached partition table for each of the possibly multiple
target paths are not kept in synch with the partition table of the
multipath mapped device.
 
In this case, fdisk is failing because it attempts to read the partition
table (MBR block 0) of the disk and /dev/sdq is the passive path to the
disk.  This is correct behavior for Linux dm-multipath since it does not
attempt to provide multipathing for accesses to the target device paths,
e.g., /dev/sdq.
 
Also, be careful because this behavior can be intermittent since this
operation would have succeeded if this logical block was resident in the
page cache, which could happen if /dev/sdq was an active path to the
logical unit just a little while beforehand.


________________________________

	From: dm-devel-bounces at redhat.com
[mailto:dm-devel-bounces at redhat.com] On Behalf Of Herr, Manuel
	Sent: Thursday, September 28, 2006 6:03 AM
	To: dm-devel at redhat.com
	Subject: [dm-devel] Device Access Question
	
	

	Is it correct that if an LUN is seen over 2 Paths, 
	1 active 1 standby i cannot fdisk the standby device? 
	

	#multipath -l 
	(3600601604032190097fd5fbefd22db11) 
	[size=11 GB][features="1 queue_if_no_path"][hwhandler="1 emc"] 
	\_ round-robin 0 [active] 
	 \_ 3:0:0:7 sdi 8:128 [active][ready] 
	\_ round-robin 0 [enabled] 
	 \_ 3:0:1:7 sdq 65:0  [active][ready] 

	#fdisk -l /dev/sdi 
	Disk /dev/sdi: 11.8 GB, 11811160064 bytes 
	255 heads, 63 sectors/track, 1435 cylinders 
	Units = cylinders of 16065 * 512 = 8225280 bytes 

	   Device Boot      Start         End      Blocks   Id  System 
	/dev/sdi1               1        1435    11526606   83  Linux 

	#fdisk -l /dev/sdq 
	--> creates in /var/log/messages 
	Sep 28 12:01:22 saptf2 kernel: end_request: I/O error, dev sdq,
sector 0 
	Sep 28 12:01:22 saptf2 kernel: Buffer I/O error on device sdq,
logical block 0 

	Regards Manuel Herr 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listman.redhat.com/archives/dm-devel/attachments/20060929/512c6858/attachment.htm>