Dear Philip,<br><br>Thanks a lot for reviewing and telling me about the errors. I will try to modify according to your suggestion and let you know after this.<br><br>Best Regards,<br>Krishna<br><br><div class="gmail_quote">
On Mon, Nov 16, 2009 at 10:32 PM, Philipp Reisner <span dir="ltr"><<a href="mailto:philipp.reisner@linbit.com">philipp.reisner@linbit.com</a>></span> wrote:<br><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
Am Montag, 16. November 2009 17:11:00 schrieb er krishna:<br>
<div class="im">> Dear Lars,<br>
><br>
> Please tell if you like the design and you are thinking to put it on your<br>
> site. I have some detail design other documents also which I want to put on<br>
> DRBD site.<br>
><br>
> Best Regards,<br>
> Krishna<br>
><br>
<br>
<br>
</div>Hi Krishna,<br>
<br>
It is not only about the design, it is more about the content of your<br>
papers. There are a number of errors in the "ha_san_raid_lvm_aoe.pdf"<br>
document:<br>
<br>
*) Never put meta data on a loop device. It might deadlock under<br>
memory pressure.<br>
*) No need to create the file system on top of DRBD twice. Once<br>
on each node!!<br>
*) I got the impression you advise to mount an ext3 fs on the<br>
secondary (ro). This has a severe cache coherency issue, and<br>
should not be advised to anyone.<br>
<br>
The kernel patch simply drops all caches on the initiator side,<br>
which is rude in a way.<br>
<br>
Sorry, but I do not want to have this on .<a href="http://drbd.org" target="_blank">drbd.org</a> as it is.<br>
<br>
-Phil<br>
--<br>
: Dipl-Ing Philipp Reisner<br>
<div class="im">: LINBIT | Your Way to High Availability<br>
</div>: Tel: +43-1-8178292-50, Fax: +43-1-8178292-82<br>
: <a href="http://www.linbit.com" target="_blank">http://www.linbit.com</a><br>
<br>
DRBD(R) and LINBIT(R) are registered trademarks of LINBIT, Austria.<br>
</blockquote></div><br>