On Jun 24, 2013, at 5:56 AM, Sitsofe Wheeler wrote:
> Hi,
>
> Does anyone know if it should be possible to observe the write part of a
> COMPARE AND WRITE operation from another session when it is partially
> complete? For example imagine /dev/sdb currently consists of nulls, /dev/
> sdc is another session to the same target and a COMPARE AND WRITE is is
> issued, if the COMPARE AND WRITE has to write 255 blocks of 0xff and the
> verify succeeds, should it be possible to see only the first 128 blocks
> with 0xff and the rest with 0x00 by using hexdump /dev/sdc when using
> open-iscsi connected to an iSCSI target?
That would be a SCSI standard question. I haven't looked at that specific
command, but the normal SCSI rules is that when you overlap a read with a
write, the read may see any combination of old sectors and new sectors -- but
no sector contents other than the old or the new, i.e., no "random bits". For
example, seeing 0xff data in all odd numbered blocks and 0x00 data in all even
numbered blocks would be valid (not likely, perhaps, but legal).
paul
--
You received this message because you are subscribed to the Google Groups
"open-iscsi" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
To post to this group, send email to [email protected].
Visit this group at http://groups.google.com/group/open-iscsi.
For more options, visit https://groups.google.com/groups/opt_out.
