Hi,<br><br>> I'm not familiar with this tool. Does it mix reads and writes?<br>No it does not mix them.<br>Here is it's testing logic ( seems to be ):<br>Writing a byte at a time...done<br>Writing intelligently...done<br>
Rewriting...done<br>Reading a byte at a time...done<br>Reading intelligently...done<br>start 'em...done...done...done...done...done...<br>Create files in sequential order...done.<br>Stat files in sequential order...done.<br>
Delete files in sequential order...done.<br>Create files in random order...done.<br>Stat files in random order...done.<br>Delete files in random order...done.<br><br>> You may want to consider activating disk flushes. Why does your config disable those?<br>
Because of BBU present on RAID controller.<br><a href="http://www.drbd.org/users-guide/s-throughput-tuning.html">http://www.drbd.org/users-guide/s-throughput-tuning.html</a><br><br>> Looking at your write throughput being much larger than your read<br>
> throuhgput(!) I suspect you're writing to some sort of aggressive cache,<br>> and your reads must wait until the writes are completed to the disk(?)<br>> If that is indeed the case, this might explain why syncing to the<br>
> secondary slows down the read performance.<br>I noticed this. Could be "feature" of H700 RAID controller. Anyway I have changed Write Cache policy to <br>Write-Through, so theoretically I would have more "real" results. But I still get same results. I think I will try to contact DELL support/forum ...<br>
<br>Thank you.<br>--<br>Maxim Ianoglo<br><br><div class="gmail_quote">On Mon, May 2, 2011 at 11:05 AM, Felix Frank <span dir="ltr"><<a href="mailto:ff@mpexnet.de">ff@mpexnet.de</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex;">
Hi,<br>
<div class="im"><br>
On 04/30/2011 11:12 PM, Maxim Ianoglo wrote:<br>
> Hello,<br>
><br>
> Have two DELL PE R410, RAID 10 15K rpm Seagate SAS disks. Both servers<br>
> are connected via 1Gb link.<br>
><br>
> Testing performance with bonnie++.<br>
> On a RAW device I get: ~240MB/s Read and ~390MB/s Write<br>
> On DRBD device with Secondary DRBD host connected: ~80MB/s Read and<br>
> ~340MB/s Write<br>
> On DRBD without Secondary DRBD host connected: ~245MB/s Read and ~340Write<br>
<br>
</div>I'm not familiar with this tool. Does it mix reads and writes?<br>
<br>
Looking at your write throughput being much larger than your read<br>
throuhgput(!) I suspect you're writing to some sort of aggressive cache,<br>
and your reads must wait until the writes are completed to the disk(?)<br>
If that is indeed the case, this might explain why syncing to the<br>
secondary slows down the read performance.<br>
This seems especially likely since your write throughput is<br>
a) not affected by live syncing and<br>
b) WAY above the capacity of a 1GB link!<br>
<br>
Long story short, I disbelieve your mode of testing is yielding results<br>
of any operational value.<br>
<br>
Also note that a volatile write cache may leave your secondary (and<br>
primary) disks in a corrupted state on primary crash (and the secondary<br>
even during connection loss). You may want to consider activating disk<br>
flushes. Why does your config disable those?<br>
<div class="im"><br>
> Why I get so bad Read performance on DRBD device with Secondary host<br>
> connected ? Per documentation all reads should be made l<br>
<br>
</div>Yes they are local.<br>
<br>
Also note:<br>
<div class="im"><br>
> syncer {<br>
> cpu-mask 4095;<br>
> rate 85M;<br>
> verify-alg "crc32c";<br>
> csums-alg "crc32c";<br>
> }<br>
<br>
</div>It is advisable to not use syncer rates that exceed 30% of your<br>
bandwidth by much.<br>
<br>
This rate is only used if an outdated peer is doing a full or partial<br>
sync, it doesn't affect live synchronization (which is always done as<br>
soon as each block is written).<br>
If you use up most of your bandwidth for syncing, the write performance<br>
on your DRBD will suffer.<br>
<br>
HTH,<br>
<font color="#888888">Felix<br>
</font></blockquote></div><br><div style="visibility: hidden; left: -5000px; position: absolute; z-index: 9999; padding: 0px; margin-left: 0px; margin-top: 0px; overflow: hidden; word-wrap: break-word; color: black; font-size: 10px; text-align: left; line-height: 130%;" id="avg_ls_inline_popup">
</div>