<div class="socmaildefaultfont" dir="ltr" style="font-family:Arial, Helvetica, sans-serif;font-size:10pt" ><div dir="ltr" style="font-family:Arial, Helvetica, sans-serif;font-size:10pt" ><div dir="ltr" >Russell,</div>
<div dir="ltr" > </div>
<div dir="ltr" >The maximum number of replicas, as specified via</div>
<div dir="ltr" > </div>
<div dir="ltr" style="margin-left: 40px;" >mmcrfs -R MaxDataReplicas</div>
<div dir="ltr" > </div>
<div dir="ltr" >causes space to be reserved in the inode and indirect block for the replicated disk address for each data block.</div>
<div dir="ltr" > </div>
<div dir="ltr" >For files which are small enough to fit in the inode, it should not make a difference.</div>
<div dir="ltr" > </div>
<div dir="ltr" >For files with indirection level "direct" then -R 2 or -R3 will result in fewer disk addresses to fit in the inode, and for the larger files force the use of indirect blocks. Then each indirect block will have space for half (-R 2) or a third (-R 3) of the number of disk addresses, as compared with -R 1. Which means that additional indirect blocks will be needed to accommodate a file of a given size.</div>
<div dir="ltr" > </div>
<div dir="ltr" >For even larger files requiring higher indirection levels, the statement above applies to the leaf indirect blocks.</div>
<div dir="ltr" > </div>
<div dir="ltr" style="margin-left: 40px;" >>>> This information was written prior to GPFSv5 when subblocks changed from only 32. Does it still hold true that unused data replica\u2019s use metadata space with v5? </div>
<div dir="ltr" style="margin-left: 40px;" > </div>
<div dir="ltr" >The same remains true for v5 file systems with more than 32 subblocks per block.</div>
<div dir="ltr" > </div>
<div dir="ltr" > Felipe</div>
<div dir="ltr" > </div>
<div dir="ltr" >----<br>Felipe Knop knop@us.ibm.com<br>GPFS Development and Security<br>IBM Systems<br>IBM Building 008<br>2455 South Rd, Poughkeepsie, NY 12601<br>(845) 433-9314 T/L 293-9314<br> </div>
<div dir="ltr" > </div>
<div dir="ltr" > </div>
<blockquote data-history-content-modified="1" data-history-expanded="1" dir="ltr" style="border-left:solid #aaaaaa 2px; margin-left:5px; padding-left:5px; direction:ltr; margin-right:0px" >----- Original message -----<br>From: Russell Nordquist <russell@nordquist.info><br>Sent by: gpfsug-discuss-bounces@spectrumscale.org<br>To: gpfsug-discuss@spectrumscale.org<br>Cc:<br>Subject: [EXTERNAL] [gpfsug-discuss] data replicas and metadata space used<br>Date: Wed, Sep 2, 2020 10:55 AM<br> <br><!--Notes ACF
<meta http-equiv="Content-Type" content="text/html; charset=utf8" >-->
<div>I was reading this slide deck on GPFS metadata sizing and I ran across something</div>
<div><a href="http://files.gpfsug.org/presentations/2016/south-bank/D2_P2_A_spectrum_scale_metadata_dark_V2a.pdf" target="_blank">http://files.gpfsug.org/presentations/2016/south-bank/D2_P2_A_spectrum_scale_metadata_dark_V2a.pdf</a></div>
<div> </div>
<div>On slide 51 it says</div>
<div> </div>
<div><i>"Max replicas for Data, multiplies the MD capacity used<br><br>\u2022 Reserves space in MD for the replicas even if no files replicated!\u201d</i></div>
<div> </div>
<div>This is something I did not realize - setting data replicas to 2 or even 3 consumes metadata space even if you are not using the data replicas. For metadata replicas it says unused replica\u2019s have little impact - great. I like to set data and metadata replica\u2019s to 3 when I make a filesystem even when the initial replicas used are 1 since you never know what will change down the road. However this makes me wonder about that idea for the data replica\u2019s - it\u2019s really expensive metadata spacewise.</div>
<div> </div>
<div>This information was written prior to GPFSv5 when subblocks changed from only 32. Does it still hold true that unused data replica\u2019s use metadata space with v5? </div>
<div> </div>
<div>thanks</div>
<div>Russell</div>
<div> </div>
<div> </div>
<div><font size="2" face="Default Monospace,Courier New,Courier,monospace" >_______________________________________________<br>gpfsug-discuss mailing list<br>gpfsug-discuss at spectrumscale.org<br><a href="http://gpfsug.org/mailman/listinfo/gpfsug-discuss" target="_blank">http://gpfsug.org/mailman/listinfo/gpfsug-discuss</a> </font></div></blockquote>
<div dir="ltr" > </div></div></div><BR>