<html><head><style type='text/css'>p { margin: 0; }</style></head><body><div style='font-family: Arial; font-size: 10pt; color: #000000'>Please remember to CC the list too!<br><br><div><span name="x"></span>
<style>/* Style Definitions */p.MsoNormal, p.MsoAutoSig, li.MsoNormal, div.MsoNormal {margin:0cm; margin-bottom:.0001pt;}@page WordSection1 {size:612.0pt 792.0pt; margin:72.0pt 72.0pt 72.0pt 72.0pt; mso-header-margin:36.0pt; mso-footer-margin:36.0pt; mso-paper-source:0;}div.WordSection1 {page:WordSection1;}</style>
<div class="WordSection1">
<p class="MsoNormal">Jake<o:p></o:p></p>
<p class="MsoNormal"><o:p> </o:p></p>
</div>
<span name="x"></span><br></div><hr id="zwchr"><blockquote style="border-left: 2px solid rgb(16, 16, 255); margin-left: 5px; padding-left: 5px; color: rgb(0, 0, 0); font-weight: normal; font-style: normal; text-decoration: none; font-family: Helvetica,Arial,sans-serif; font-size: 12pt;"><b>From: </b>"Valentin FERON" <valferon@gmail.com><br><b>To: </b>"Jake Smith" <jsmith@argotec.com><br><b>Sent: </b>Thursday, November 3, 2011 11:32:37 AM<br><b>Subject: </b>Re: [DRBD-user] DRBD sync speed is too fast comparing to my conf<br><br>Hello and thanks for replies,<div><br></div><div>The reason why I have this problem is because I'm replicating a resource between two servers witch are in two different buildings connected by a fiber link.</div><div><br>
</div><div>I have only one link, that's why my data and the drbd replication are on the same link and not separated...<br><br>I'm looking around QoS on the switch but it's not giving good results.</div><div><br>
</div><div id="DWT12714">Any low (no) cost idea ?<br></div></blockquote>Not a real simple one that I know of. Maybe some of the DRBD experts here have an idea. My only thought is more complex - If you have switches that support bandwidth limiting on the switch port(s) then you could add a second NIC to each server (if you don't have already available) and set them up with a different IP's for replication link. Then limit the bandwidth on the switch ports the replication NIC's attach to to something feasible such as 50M or 30M. That way the replication links would never exceed that switch limited bandwidth. If you try something like that make sure to change your syncer rate to something that is a fraction of what you limited the bandwidth to. This is nothing I've tried/tested just sounds good in theory!<br><br>HTH<br><blockquote style="border-left: 2px solid rgb(16, 16, 255); margin-left: 5px; padding-left: 5px; color: rgb(0, 0, 0); font-weight: normal; font-style: normal; text-decoration: none; font-family: Helvetica,Arial,sans-serif; font-size: 12pt;"><div><br><div class="gmail_quote">2011/11/3 Jake Smith <span dir="ltr"><<a href="mailto:jsmith@argotec.com" target="_blank">jsmith@argotec.com</a>></span><br><blockquote class="gmail_quote" style="margin: 0pt 0pt 0pt 0.8ex; border-left: 1px solid rgb(204, 204, 204); padding-left: 1ex;">
<div class="im">----- Original Message -----<br>
> From: "vferon" <<a href="mailto:valferon@gmail.com" target="_blank">valferon@gmail.com</a>><br>
> To: <a href="mailto:drbd-user@lists.linbit.com" target="_blank">drbd-user@lists.linbit.com</a><br>
> Sent: Friday, October 14, 2011 9:13:12 AM<br>
> Subject: [DRBD-user] DRBD sync speed is too fast comparing to my conf<br>
><br>
> Hello,<br>
><br>
> In my config file I specified a sync rate of 30M but in reality, when<br>
> a<br>
> create a file on the master, it replicates it at more thant 70M/s and<br>
> using most of my bandwidth.<br>
<br>
</div>Syncer rate only applies to background sync i.e. when one node is not "uptodate" it will bring it back "uptodate" at that rate. If you are talking about creating a file on one node when both are "uptodate" then that is foreground replication and that goes as fast as the stack/IO/bandwidth will allow. If you tried to change that speed it would negatively impact the application speed. You do have the replication on a seperate link right?<br>
<br>
HTH<br>
<br>
Jake<br>
<div><div></div><div class="h5"><br>
><br>
> Here is my config file :<br>
> /****************************************************************************************************/<br>
> global { usage-count no; }<br>
> resource r0 {<br>
> protocol C;<br>
> startup {<br>
> wfc-timeout 15; # wfc-timeout can be dangerous<br>
> (<a href="http://forum.proxmox.com/threads/3465-Is-it-safe-to-use-wfc-timeout-in-DRBD-configuration" target="_blank">http://forum.proxmox.com/threads/3465-Is-it-safe-to-use-wfc-timeout-in-DRBD-configuration</a>)<br>
> degr-wfc-timeout 60;<br>
> }<br>
> net {<br>
> cram-hmac-alg sha1;<br>
> shared-secret "sdlkjSFcqseazecj3345Sd";<br>
> after-sb-0pri discard-zero-changes;<br>
> after-sb-1pri discard-secondary;<br>
> after-sb-2pri disconnect;<br>
> max-buffers 80000;<br>
> max-epoch-size 20000;<br>
> unplug-watermark 16;<br>
> sndbuf-size 10485760;<br>
><br>
> }<br>
> syncer {<br>
> rate 30M;<br>
> }<br>
> on plmbnas01 {<br>
> device /dev/drbd0;<br>
> disk /dev/sda4;<br>
> address <a href="http://10.0.0.11:7788" target="_blank">10.0.0.11:7788</a>;<br>
> meta-disk internal;<br>
> }<br>
> on plmanas01 {<br>
> device /dev/drbd0;<br>
> disk /dev/sda4;<br>
> address <a href="http://10.0.0.10:7788" target="_blank">10.0.0.10:7788</a>;<br>
> meta-disk internal;<br>
> }<br>
> }<br>
> /****************************************************************************************************/<br>
><br>
><br>
> Did someone has the same problem and how dit he solve it?<br>
><br>
> Thanks<br>
><br>
> Valentin Feron<br>
><br>
</div></div>> _______________________________________________<br>
> drbd-user mailing list<br>
> <a href="mailto:drbd-user@lists.linbit.com" target="_blank">drbd-user@lists.linbit.com</a><br>
> <a href="http://lists.linbit.com/mailman/listinfo/drbd-user" target="_blank">http://lists.linbit.com/mailman/listinfo/drbd-user</a><br>
><br>
><br>
</blockquote></div><br></div>
</blockquote><br></div></body></html>