<P>
><BR>
> It has been observed that when protocol A and B is used, there is<BR>
>some impact on the I/O performance (especially on sequential writes)<BR>
>compared to protocol C. But theoretically, protocols A and B should be faster than C. Why Tiobench results (http://www.drbd.org/performance .html) says that protocol C is good ?<BR>
><BR>
> Even i carried out performance tests (using Tiobench tool) on 8.0.3 build and i observed that the results are almost same as the one mentioned on the DRBD site.<BR>
><BR>
> Please let me know the reason behind it.<BR>
><BR>
<BR>
LARS REPLY:<BR>
<BR>
><BR>
>because those results are horribly out of date?<BR>
>note, kernel 2.4 and DRBD-0.6.6 were used for these tests.<BR>
>protocol A and B implementation was broken then.<BR>
><BR>
>because there are two aspects to protocol choice in DRBD?<BR>
> * latency<BR>
> * data safety<BR>
><BR>
>because you measured throughput,<BR>
>and the choice of DRBD protocol has no impact on sustainable throughput?<BR>
><BR>
>we realy should take that page offline.<BR>
>or better yet, replace with some more recent benchmarks.<BR>
>I'll try to dig some up.<BR>
><BR>
>with recent 8.0 resp. 8.2,<BR>
>the protocols finally are implemented correct,<BR>
>and indeed show the expected latency improvements:<BR>
>A exposes nearly local latency,<BR>
>B has some latency overhead (order of round-trip-time),<BR>
>C latency overhead is (order of round-trip-time plus some).<BR>
><BR>
>as mentioned, you trade data safety against latency.<BR>
><BR>
>--<BR>
>: Lars Ellenberg http://www.linbit.com :<BR>
>: DRBD/HA support and consulting sales at linbit.com :<BR>
>: LINBIT Information Technologies GmbH Tel +43-1-8178292-0 :<BR>
>: Vivenotgasse 48, A-1120 Vienna/Europe Fax +43-1-8178292-82 :<BR>
>__<BR>
<BR>
<BR>
I have carried out performance test on 8.0.3 version using Tiobench<BR>
tool. But, results says that Avg latency in case of protocol C is<BR>
less compared to protocol A and B (during write operation),<BR>
theoretically which is not expected.<BR>
<BR>
Can someone please tell me the reason for it? (refer results given below for Sequential and Random writes) <BR>
<BR>
<BR>
<BR>
--------------<BR>
>File size= 1024 DRBD partition size=2GB<BR>
>Block size= 4096 RAM=256MB<BR>
>Number of threads =8 RTT=~0.3ms<BR>
><BR>
><BR>
>Random writes:<BR>
>---------------------------------------------------------------<BR>
> Rate Max Avg Max CPU<BR>
> CPU % latency latency Efficiency<BR>
>---------------------------------------------------------------<BR>
>2.6.18-53.el5 2.40 20.60 2.616 1426.08 12<BR>
>(without DRBD)<BR>
>2.6.18-53.el5<BR>
>(protocol A) 1.05 11.15 4.206 1503.61 9<BR>
><BR>
>2.6.18-53.el5<BR>
>(protocol B) 1.11 12.53 5.024 1705.55 9<BR>
><BR>
>2.6.18-53.el5<BR>
>(protocol C) 0.99 10.09 1.834 1853.46 10<BR>
>---------------------------------------------------------------<BR>
><BR>
><BR>
>Sequential writes:<BR>
>---------------------------------------------------------------<BR>
> Rate Max Avg Max CPU<BR>
> CPU % latency latency Efficiency<BR>
>---------------------------------------------------------------<BR>
>2.6.18-53.el5 37.85 307.7 0.600 2511.88 12<BR>
>(without DRBD)<BR>
>2.6.18-53.el5<BR>
>(protocol A) 11.37 77.77 1.938 14701.06 15<BR>
>2.6.18-53.el5<BR>
>(protocol B) 9.52 62.69 2.240 21407.66 15<BR>
>2.6.18-53.el5<BR>
>(protocol C) 12.16 100.6 2.131 15969.09 12<BR>
>---------------------------------------------------------------<BR>
><BR>
<BR>
<BR>
<BR>
Regards,<BR>
Azeez<BR>
</P>
<br><br>
<Table border=0 Width=644 Height=57 cellspacing=0 cellpadding=0 style='font-family:Verdana;font-size:11px;line-height:15px;'><TR><td><a href='http://adworks.rediff.com/cgi-bin/AdWorks/click.cgi/www.rediff.com/signature-home.htm/1050715198@Middle5/2047758_2040452/2047768/1?PARTNER=3&OAS_QUERY=null' target=new ><img src ='http://imadworks.rediff.com/cgi-bin/AdWorks/adimage.cgi/2047758_2040452/creative_2047768.gif' alt='Monster' border=0></a></td></TR></Table>