<html>
<body>
<font style="font-family:Arial; font-size: 12px; color: black;">
Hi Christophe!<br>
<br>
<br>
<table height="68" border="0" cellspacing="5" cellpadding="0"><tr><td width="230" height="68" align="left" valign="middle"><img src="cid:5VfDWTiyBUOwvJT.gif@MNxAcQFeEMfDPgQ" width="230" height="68" border="0" alt="Salva Kindlustuse AS"></td><td valign="middle" align="left" height="68" style="font-family: Verdana; font-size: 10px; color: black;"><b>Viktor Larionov</b><br>IT osakonna juhataja<br>IT-osakond<br>Salva Kindlustuse AS<br>Tel: (+372) 683 0630 | GSM: (+372) 566 86811 | <a href="mailto:Viktor.Larionov@salva.ee"><font style="font-family: Verdana; font-size: 10px; color: #1A8439; text-decoration: none;">Viktor.Larionov@salva.ee</font></a> | <a href="http://www.salva.ee"><font style="font-family: Verdana; font-size: 10px; color: #1A8439; text-decoration: none;">www.salva.ee</font></a></td></tr><tr><td valign="top" width="320" colspan="2" align="left" height="68" style="font-family: Verdana; font-size: 10px; color: black;"></td></tr></table><br>
<div id="(SMX)338844"> </div>
<br>
On 18.05.2013 16:40, Christophe Varoqui wrote:<br>
<br>
>> So, we poked up with code a bit, and wrote up a custom prioritizer,<br>
>> called sg_id. (patch for the latest multipath-tools available here:<br>
>> http://viktor.ee/multipath-tools-patches/sg_id_prio.patch [1])<br>
><br>
> Would the existing weightedpath work for you ?<br>
><br>
<br>
It's a shame I missed weighted path while googling for a solution in <br>
the<br>
first place, thanks for the hint. I checked into that, and I should say<br>
it's very alike with what we have done in sg_id_prio patch. I've <br>
briefly<br>
checked into weighted path and I don't quite like the concept of the<br>
design.<br>
Weighted path as suggested wants you to pass the arguments via the same<br>
"prio" conf param which is used to indicate the prioritizer(s) to be<br>
used. I don't see much sense in that, keeping in mind that you have a<br>
"prio_args" param designed initially for that purpose.<br>
That is also the reason why weighted path needs to hack deep into the<br>
libmultipath architecture. (while the modular nature of libmultipath <br>
sees<br>
that the prioritizer is a separate shared lib and has nothing much to<br>
hack in the main core). From my point that's not quite right.<br>
When it comes to enabling multiple prioritizers in a row (our second<br>
patch) - weighted path approach makes it impossible, as it passes its <br>
own<br>
arguments on the prio line.<br>
<br>
So answering your question - I think that weighted path would not solve<br>
our problem, mainly due to its architectural approach and difficulties<br>
with multi prioritizer setup when dealing with weighted path.<br>
<br>
I also agree with Kiyoshi comment to weighted path back from 2008. The<br>
whole story of assigning a static priority based on hbtl or node name<br>
pattern from my oppinion is useless on it's own. It's a great extra, <br>
for<br>
use cases as ours, where you just need to tweak a little the priority<br>
given by a more intelligent prioritizer. (that's why multiprio)<br>
<br>
On the other hand, weighted path has a couple of neat features like<br>
setting a static priority not only based on hbtl, but also on a device<br>
node name. sg_io_prio supports hbtl only as this is less likely to<br>
change. (though not totally excluded of course). If disregard this <br>
point,<br>
and the architectural difference both modules are identical.<br>
<br>
I personally think that either weighted path should implement a more<br>
intelligent way of operation (utilize prio_args and not mess with the<br>
basic architecture of libmultipath) or our prio_sg_id should support<br>
device node names (which is a half an hour work we can do).<br>
<br>
Cheers!<br>
vik<br>
</font>
</body>
</html>