<div>Hi,</div>
<div> </div>
<div>Most of the issues related to insufficient memory while using client & OSS on same Server, in our case more than 50% of the memory is alway free on OSS as we have 32GB on MDT/OSS nodes.</div>
<div> </div>
<div>We do'nt see the performance issues while the filesize is large & number of files are less; performance issues starts as number of files increase. So it might be the turn around time which is causing the problem as each transaction need to go to MGS/MDT & then again back to OSS.</div>
<div> </div>
<div>Example: </div>
<div>Case1 : If we try to write data set of 1GB with 200Files ; write, delete & read is faster </div>
<div>Case2 : if we try to write data set of 1GB with 14000Files ; write, delete & read is very slow.</div>
<div> </div>
<div> Even we see most of the issues reported related to small files & work around to improve performance by disable debug mode. But in our case disabling debug mode did not help to improve the performance.</div>
<div> </div>
<div>Please let us know if there is any other alternate options to improve performance while using more number of file with small size.</div>
<div> </div>
<div>Thanks,</div>
<div>Anil</div>
<div> </div>
<div> </div>
<div><br><br> </div>
<div class="gmail_quote">On Fri, Dec 5, 2008 at 4:51 PM, Balagopal Pillai <span dir="ltr"><<a href="mailto:pillai@mathstat.dal.ca" target="_blank">pillai@mathstat.dal.ca</a>></span> wrote:<br>
<blockquote class="gmail_quote" style="PADDING-LEFT: 1ex; MARGIN: 0px 0px 0px 0.8ex; BORDER-LEFT: #ccc 1px solid">
<div>"OST - 13 ( also act as nfsserver)"<br><br></div> Then I am assuming that your OSS is also a Lustre client.<br>It might be useful to search through this<br>list to find out the potential pitfalls of mounting Lustre volumes on OSS.<br>
<div>
<div></div>
<div><br><br><br>siva murugan wrote:<br>> We are trying to uptake Lustre in one of our heavy read/write<br>> intensive infrastructure(daily writes - 8million files, 1TB ). Average<br>> size of files written is 1KB ( I know , Lustre can't scale well for<br>
> small size files, but just wanted to analyze the possibility of<br>> uptaking )<br>><br>> Following are some of the tests conducted to see the difference in<br>> large and small file size writting,<br>><br>
> MDT - 1<br>> OST - 13 ( also act as nfsserver)<br>> Clients access lustrefs via NFS ( not patchless clients)<br>><br>> Test 1 :<br>><br>> Number of clients - 10<br>> Dataset size read/written - 971M (per client)<br>
> Number of files in the dataset- 14000<br>> Total data written - 10gb<br>> Time taken - 1390s<br>><br>> Test2 :<br>><br>> Number of clients - 10<br>> Dataset size read/written -1001M (per client)<br>
> Number of files in the dataset - 4<br>> Total data written - 10gb<br>><br>> Time taken - 215s<br>><br>><br>> Test3 :<br>><br>> Number of clients - 10<br>> Dataset size read.written- 53MB (per client)<br>
> Number of files in the dataset- 14000<br>> Total data written - 530MB<br>> Time taken - 1027s<br>> MDT was heavily loaded during Test3 ( Load average > 25 ). Since the<br>> file size in Test 3 is small(1kb) and number of files written is too<br>
> large(14000 x 10 clients ), obvisouly mdt gets loaded in allocating<br>> inodes, total data written in test3 is only 530MB.<br>><br>> Question : Is there any parameter that I can tune in MDT to increase<br>
> the performance when writting large number of small files ?<br>><br>> Please help<br>><br></div></div>> ------------------------------------------------------------------------<br>><br>> _______________________________________________<br>
> Lustre-discuss mailing list<br>> <a href="mailto:Lustre-discuss@lists.lustre.org" target="_blank">Lustre-discuss@lists.lustre.org</a><br>> <a href="http://lists.lustre.org/mailman/listinfo/lustre-discuss" target="_blank">http://lists.lustre.org/mailman/listinfo/lustre-discuss</a><br>
><br>_______________________________________________<br>Lustre-discuss mailing list<br><a href="mailto:Lustre-discuss@lists.lustre.org" target="_blank">Lustre-discuss@lists.lustre.org</a><br><a href="http://lists.lustre.org/mailman/listinfo/lustre-discuss" target="_blank">http://lists.lustre.org/mailman/listinfo/lustre-discuss</a><br>
</blockquote></div><br>