<font face="arial">Thank you very much Phil for your answer.<br>Sounds great.<br><br>Best Regards,<br>Ivan<br><br>At Thursday, 03/02/2011 on 10:29 Philipp Reisner wrote:<br><blockquote style="border:0;border-left: 2px solid #22437f; padding:0px; margin:0px; padding-left:5px; margin-left: 5px; ">Am Dienstag, 1. Februar 2011, um 16:49:05 schrieb Ivan Frain:<br>
> Hi all,<br>
> <br>
> I am currently evaluating DRBD as a storage candidate for highly<br>
> available storage in a virtualized environment.<br>
> It seems like a very good alternative to expensive SAN/NAS.<br>
> I was wondering how DRBD deals with the network block device deadlock<br>
> problem.<br>
> This problem (described here: <a href="http://lwn.net/Articles/195416" target="_blank" class="normal-link">http://lwn.net/Articles/195416</a>/) can be<br>
> summarized as follows: if the system runs short in memory, it will try<br>
> to write dirty page to disk in order to free memory space. if the<br>
> disk is a network block device, the dirty page write may need to<br>
> allocate some other memory pages which is not possible since the<br>
> solution to have more memory available was to write the dirty page to<br>
> disk.<br>
> <br>
<br>
This is a very well understood issue. The solution to the issue is, that<br>
all memory allocations in the write out path are backed by mempools<br>
(= Small amounts of pre allocated memory).<br>
DRBD does not deadlock in out of memory situations, at least since version<br>
8. Maybe even earlier.<br>
<br>
Best,<br>
Phil<br>
-- <br>
: Dipl-Ing Philipp Reisner<br>
: LINBIT | Your Way to High Availability<br>
: Tel: +43-1-8178292-50, Fax: +43-1-8178292-82<br>
: <a href="http://www.linbit.com" target="_blank" class="normal-link">http://www.linbit.com</a><br>
<br>
DRBD(R) and LINBIT(R) are registered trademarks of LINBIT, Austria.</blockquote></font>