The process of syncing master and shadow is bidirectional. You can´t not only update the shadow from the master, you can update the master from the shadow as well. This is a really neat feature for diaster recovery.
Let´s assume, you tried a new version of your software. At first all is well, but a minute later the system is toast. Later you will find out, that there was a race condition in the new code, that only manifested on your production system. But you don´t know this now. And to add insult to injury, your face go white after looking into the directory.
# ls -l /mnt/testindex*
-rw-r--r-- 1 root root 0 Apr 25 19:40 /mnt/testindex1
-rw-r--r-- 1 root root 0 Apr 25 19:41 /mnt/testindex2
The new code killed your
testindex-files. Zero bytes. And you hear the angry guy or lady from customer support shouting your name. But you were cautious, you´ve created a point-in-time copy before updating the system.
So, calm down and recover before a customer support lynch mob reach your office with forks and torches. Leave the filesystem and unmount it.
# cd /
# umount /mnt
Now sync the master with the slave. Yes, the other way round.
# iiadm -u m /dev/rdsk/c1d1s3
Overwrite master with shadow volume? yes/no yes
Okay ... after a few moments the shell prompt appears again. Now you can mount it again.
# mount /dev/dsk/c1d0s3 /mnt
# cd /mnt
Let´s check our work and check the
testindex files.
# ls -l /mnt/testindex*
-rw------T 1 root root 1024 Apr 25 18:11 /mnt/testindex1
-rw------T 1 root root 3072 Apr 25 19:33 /mnt/testindex2
Phew ... rescued ... and the lynch mob in front of your office throws the torches out of the window, directly on the car of the CEO (of course by accident

)
Okay, the next installment of the "Less known Solaris Features" series is online. This time i will discuss the feature Point in time copies with AVS. As i have to go through some theory at first, it´s a quite long tutorial. It´s the longest so far. Point-
Tracked: Apr 27, 11:10