<div>Hi,</div>
<div> </div>
<div>Same behaviour as the one from Rafael.</div>
<div> </div>
<div>Everything is coherent as long as you use the exclusive flag from the rogue node, the locking does the job. Deactivating an already opened VG (mounted lvol) is not possible either. How could this behave in case one used raw devices instead of FS ? </div>
<div> </div>
<div>But when you come to ignore the exclusive flag on the rogue node (vgchange -a y vgXX) the locking is completely bypassed. It's definitely here that the watchdog has to be (within the tools lvchange, vgchange, or at dlm level).</div>
<div> </div>
<div> below the output of the test:</div>
<div> </div>
<div>node1 = nodeid 1<br>node2 = nodeid 2</div>
<div>
<p>node1:</p>
<p>vgchange -a ey vg11<br> 1 logical volume(s) in volume group "vg11" now active</p>
<p>[root@node1 ~]# lvs<br> LV VG Attr LSize Origin Snap% Move Log Copy% Convert<br> lvol1 vg11 -wi-a- 6.00G </p>
<p>[root@node1 ~]# ldebug</p>
<p>id nodeid remid pid xid exflags flags sts grmode rqmode time_ms r_nodeid r_len r_name<br>39a0001 0 0 434 0 1 1 2 5 -1 0 0 64 "iZ8vgn7nBm05aMSo5cfpy63rflTqL2ryr3Xrp1prEGceCkA2dhSA2ENWocunEfdf"</p>
<p>[root@node1 ~]# cdebug</p>
<p>Resource ffff81010abd6e00 Name (len=64) "iZ8vgn7nBm05aMSo5cfpy63rflTqL2ryr3Xrp1prEGceCkA2dhSA2ENWocunEfdf" <br>Master Copy<br>Granted Queue<br>039a0001 EX<br>Conversion Queue<br>Waiting Queue</p>
<p>[root@node1 ~]# mount /dev/vg11/lvol1 /mnt</p>
<p>node2:</p>
<p>[root@node2 ~]# vgchange -a ey vg11<br> Error locking on node node2: Volume is busy on another node<br> 0 logical volume(s) in volume group "vg11" now active</p>
<p>ldebug<br> nothing<br>cdebug<br> nothing</p>
<p><br>[root@node2 ~]# vgchange -a n vg11<br> Error locking on node node1: LV vg11/lvol1 in use: not deactivating<br> 0 logical volume(s) in volume group "vg11" now active</p>
<p># vg11/lvol1 is already mounted on node1 !</p>
<p> [root@node2 ~]# vgchange -a y vg11<br> 1 logical volume(s) in volume group "vg11" now active</p>
<p>[root@node2 ~]# mount /dev/vg11/lvol1 /mnt<br>success<br># ..it happens ! ;-)</p>Brem<br><br> </div>
<div><span class="gmail_quote">2009/7/31, brem belguebli <<a href="mailto:brem.belguebli@gmail.com">brem.belguebli@gmail.com</a>>:</span>
<blockquote class="gmail_quote" style="PADDING-LEFT: 1ex; MARGIN: 0px 0px 0px 0.8ex; BORDER-LEFT: #ccc 1px solid">Hi Rafael,
<div><br> </div>
<div>Good testing, it confirms that some additional barriers are necessary to prevent undesired behaviours.</div>
<div> </div>
<div>I'll test by tomorrow the same procedure at VG level.</div>
<div><br> </div>
<div><br> </div>
<div><br>
<div class="gmail_quote">2009/7/30 Rafael Micó Miranda <span dir="ltr"><<a onclick="return top.js.OpenExtLink(window,event,this)" href="mailto:rmicmirregs@gmail.com" target="_blank">rmicmirregs@gmail.com</a>></span>
<div><span class="e" id="q_122ce1dbb3e147d7_1"><br>
<blockquote class="gmail_quote" style="PADDING-LEFT: 1ex; MARGIN: 0px 0px 0px 0.8ex; BORDER-LEFT: #ccc 1px solid">Hi Brem<br><br>El jue, 30-07-2009 a las 09:15 +0200, brem belguebli escribió:<br>
<div>> Hi,<br>><br>> does it look like we're hiting some "undesired feature" ;-)<br>><br>> Concerning the 0 nodeid, I think I read that on some Redhat documents<br>> or bugzilla report, I could find it out.<br>
><br>> Brem<br>><br>><br>><br>><br> </div>
<div>> --<br>> Linux-cluster mailing list<br>> <a onclick="return top.js.OpenExtLink(window,event,this)" href="mailto:Linux-cluster@redhat.com" target="_blank">Linux-cluster@redhat.com</a><br>> <a onclick="return top.js.OpenExtLink(window,event,this)" href="https://www.redhat.com/mailman/listinfo/linux-cluster" target="_blank">https://www.redhat.com/mailman/listinfo/linux-cluster</a><br>
<br> </div>I made some test on my lab environment too, i attach the results in the<br>TXT file.<br><br>My conclusions:<br><br>1.- lovgols with exclusive flag must be used over clustered volume<br>groups (obvious and already known)<br>
2.- logvols activated with exclusive flag must be handled EXCLUSIVELY<br>with the exclusive flag<br><br>---> as part of my lvm-cluster.sh resource script, the exclusive flag is<br>part of the resource definition in cluster.conf so this is correctly<br>
handled<br><br>3.- you can activate an already active exclusive logvol on any node if<br>you dont take into accout, during the activation, the exclusive flag<br>4.- in use (opened) logvols are protected from deactivation from<br>
secondary nodes, even from main node<br>5.- after a node failure (hang-up, fencing...) logvol is not open<br>anymore, so it can be exclusively activated on a new node<br><br>All this was tested manually, but this is the expected behaviour on<br>
lvm-cluster.sh resource script.<br><br>Link to lvm-cluster.sh resource script:<br><br><a onclick="return top.js.OpenExtLink(window,event,this)" href="https://www.redhat.com/archives/cluster-devel/2009-June/msg00020.html" target="_blank">https://www.redhat.com/archives/cluster-devel/2009-June/msg00020.html</a><br>
<br>Cheers,<br><br>Rafael<br><br>--<br><font color="#888888">Rafael Micó Miranda<br></font><br>--<br>Linux-cluster mailing list<br><a onclick="return top.js.OpenExtLink(window,event,this)" href="mailto:Linux-cluster@redhat.com" target="_blank">Linux-cluster@redhat.com</a><br>
<a onclick="return top.js.OpenExtLink(window,event,this)" href="https://www.redhat.com/mailman/listinfo/linux-cluster" target="_blank">https://www.redhat.com/mailman/listinfo/linux-cluster</a><br></blockquote></span></div>
</div><br> </div></blockquote></div><br>