<div dir="ltr">Patrick,<div><br></div><div>You bring up an interesting point on read vs. write performance. We can't use lfs_migrate control the stripe count used for writes (obviously), so that is left up to the application developer or at least the user to intelligently place shared access files in a directory with wider striping. Restriping a file with lfs_migrate could change *read* performance characteristics, so there is indeed some risk there... but your work implies that is not too bad. If we only restripe files that are "old", then the likelyhood that they will be read again goes way down, and balancing capacity used plays a bigger factor. Bottom line is that I think restriping has more potential for upsides than down. :)</div><div><br></div><div>Thanks,</div><div>Nathan</div><div><br></div><div><div class="gmail_extra"><br><div class="gmail_quote">On Wed, May 18, 2016 at 1:22 PM, Patrick Farrell <span dir="ltr"><<a href="mailto:paf@cray.com" target="_blank">paf@cray.com</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
<div bgcolor="#FFFFFF" text="#000000">
Nathan,<br>
<br>
This *is* excellent fodder for discussion.<br>
<br>
A few thoughts from a developer perspective. When you stripe a file
to multiple OSTs, you're spreading the data out across multiple
targets, which (to my mind) has two purposes:<br>
1) More even space usage across OSTs (mostly relevant for *really*
big files, since in general, singly striped files are distributed
across OSTs anyway)<br>
2) Better bandwidth/parallelism for accesses to the file.<br>
<br>
The first one lends itself well to a file size based heuristic, but
I'm not sure the second one does. That's more about access
patterns. I'm not sure that you see much bandwidth benefit from
striping with a single client, at least as long as an individual OST
is fast relative to a client (increasingly common, I think, with
flash and larger RAID arrays). So then, whatever the file size, if
it's accessed from one client, it should probably be single striped.<br>
<br>
Also, for shared files, client count relative to stripe count has a
huge impact on write performance. Assuming strided I/O patterns,
anything more than 1 client per stripe/OST is actually worse than 1
client. (See my lock ahead presentation at LUG'15 for more on
this.) Read performance doesn't share this weirdness, though.<br>
<br>
All that's to say that for case 2 above, at least for writing, it's
access pattern/access parallelism, not size, which matters. I'm
sure there's some correlation between file size and how parallel the
access pattern is, but it might be very loose, and at least write
performance doesn't scale linearly with stripe size. Instead, the
behavior is complex.<br>
<br>
So in order to pick an ideal striping with case 2 in mind, you
really need to understand the application access pattern. I can't
see another way to do that goal justice. (The Lustre ADIO in the
MPI I/O library does this, partly by controlling the I/O pattern
through I/O aggregation for collective I/Os.)<br>
<br>
So I think your tool can definitely help with case 1, not so sure
about case 2.<br>
<br>
- Patrick<br>
<br>
<div>On 05/18/2016 12:22 PM, Nathan Dauchy -
NOAA Affiliate wrote:<br>
</div>
<blockquote type="cite">
<div dir="ltr">
<div class="gmail_quote">
<div dir="ltr">
<div>
<div>Greetings All,</div>
<div><br>
</div>
<div>I'm looking for your experience and perhaps some
lively discussion regarding "best practices" for
choosing a file stripe count. The Lustre manual has
good tips on "Choosing a Stripe Size", and in practice
the default 1M rarely causes problems on our systems.
Stripe Count on the other hand is far more difficult to
chose a single value that is efficient for a general
purpose and multi-use site-wide file system.</div>
<div><br>
</div>
<div>Since there is the "increased overhead" of striping,
and weather applications do unfortunately write MANY
tiny files, we usually keep the filesystem default
stripe count at 1. Unfortunately, there are several
users who then write very large and shared-access files
with that default. I would like to be able to tell them
to restripe... but without digging into the specific
application and access pattern it is hard to know what
count to recommend. Plus there is the "stripe these but
not those" confusion... it is common for users to have a
few very large data files and many small log or output
image files in the SAME directory.</div>
<div><br>
</div>
<div>What do you all recommend as a reasonable rule of
thumb that works for "most" user's needs, where stripe
count can be determined based only on static data
attributes (such as file size)? I have heard a "stripe
per GB" idea, but some have said that escalates to too
many stripes too fast. ORNL has a knowledge base
article that says use a stripe count of "File size / 100
GB", but does that make sense for smaller, non-DOE
sites? Would stripe count = Log2(size_in_GB)+1 be more
generally reasonable? For a 1 TB file, that actually
works out to be similar to ORNL, only gets there more
gradually:</div>
<div>
<div> <a href="https://www.olcf.ornl.gov/kb_articles/lustre-basics/#Stripe_Count" target="_blank">https://www.olcf.ornl.gov/kb_articles/lustre-basics/#Stripe_Count</a><br>
</div>
</div>
<div><br>
</div>
<div>Ideally, I would like to have a tool to give the
users and say "go restripe your directory with this
command" and it will do the right thing in 90% of
cases. See the rough patch to lfs_migrate (included
below) which should help explain what I'm thinking.
Probably there are more efficient ways of doing things,
but I have tested it lightly and it works as a
proof-of-concept.</div>
<div><br>
</div>
<div>With a good programmatic rule of thumb, we (as a
Lustre community!) can eventually work with application
developers to embed the stripe count selection into
their code and get things at least closer to right up
front. Even if trial and error is involved to find the
optimal setting, at least the rule of thumb can be a
_starting_point_ for the users, and they can tweak it
from there based on application, model, scale, dataset,
etc.</div>
<div><br>
</div>
<div>Thinking farther down the road, with progressive file
layout, what algorithm will be used as the default? If
Lustre gets to the point where it can rebalance OST
capacity behind the scenes, could it also make some
intelligent choice about restriping very large files to
spread out load and better balance capacity? (Would
that mean we need a bit set on the file to flag whether
the stripe info was set specifically by the user or
automatically by Lustre tools or it was just using the
system default?) Can the filesystem track concurrent
access to a file, and perhaps migrate the file and
adjust stripe count based on number of active clients?</div>
<div><br>
</div>
<div>I appreciate any and all suggestions, clarifying
questions, heckles, etc. I know this is a lot of
questions, and I certainly don't expect definitive
answers on all of them, but I hope it is at least food
for thought and discussion! :)</div>
<div><br>
</div>
<div>Thanks,</div>
<div>Nathan</div>
<div><br>
</div>
</div>
<div><br>
</div>
<div>
<div>--- lfs_migrate-2.7.1<span style="white-space:pre-wrap"> </span>2016-05-13
12:46:06.828032000 +0000</div>
<div>+++ lfs_migrate.auto-count<span style="white-space:pre-wrap"> </span>2016-05-17
21:37:19.036589000 +0000</div>
<div>@@ -21,8 +21,10 @@</div>
<div> </div>
<div> usage() {</div>
<div> cat -- <<USAGE 1>&2</div>
<div>-usage: lfs_migrate [-c <stripe_count>] [-h]
[-l] [-n] [-q] [-R] [-s] [-y] [-0]</div>
<div>+usage: lfs_migrate [-A] [-c <stripe_count>]
[-h] [-l] [-n] [-q] [-R] [-s] [-v] [-y] [-0]</div>
<div> [file|dir ...]</div>
<div>+ -A restripe file using an automatically selected
stripe count</div>
<div>+ currently Stripe Count = Log2(size_in_GB)</div>
<div> -c <stripe_count></div>
<div> restripe file using the specified stripe
count</div>
<div> -h show this usage message</div>
<div>@@ -31,11 +33,11 @@</div>
<div> -q run quietly (don't print filenames or status)</div>
<div> -R restripe file using default directory
striping</div>
<div> -s skip file data comparison after migrate</div>
<div>+ -v be verbose and print information about each
file</div>
<div> -y answer 'y' to usage question</div>
<div> -0 input file names on stdin are separated by a
null character</div>
<div> </div>
<div>-The -c <stripe_count> option may not be
specified at the same time as</div>
<div>-the -R option.</div>
<div>+Only one of the '-A', '-c', or '-R' options may be
specified at a time.</div>
<div> </div>
<div> If a directory is an argument, all files in the
directory are migrated.</div>
<div> If no file/directory is given, the file list is read
from standard input.</div>
<div>@@ -48,15 +50,19 @@</div>
<div> </div>
<div> OPT_CHECK=y</div>
<div> OPT_STRIPE_COUNT=""</div>
<div>+OPT_AUTOSTRIPE=""</div>
<div>+OPT_VERBOSE=""</div>
<div> </div>
<div>-while getopts "c:hlnqRsy0" opt $*; do</div>
<div>+while getopts "Ac:hlnqRsvy0" opt $*; do</div>
<div> case $opt in</div>
<div>+<span style="white-space:pre-wrap"> </span>A)
OPT_AUTOSTRIPE=y;;</div>
<div> <span style="white-space:pre-wrap"> </span>c)
OPT_STRIPE_COUNT=$OPTARG;;</div>
<div> <span style="white-space:pre-wrap"> </span>l)
OPT_NLINK=y;;</div>
<div> <span style="white-space:pre-wrap"> </span>n)
OPT_DRYRUN=n; OPT_YES=y;;</div>
<div> <span style="white-space:pre-wrap"> </span>q)
ECHO=:;;</div>
<div> <span style="white-space:pre-wrap"> </span>R)
OPT_RESTRIPE=y;;</div>
<div> <span style="white-space:pre-wrap"> </span>s)
OPT_CHECK="";;</div>
<div>+<span style="white-space:pre-wrap"> </span>v)
OPT_VERBOSE=y;;</div>
<div> <span style="white-space:pre-wrap"> </span>y)
OPT_YES=y;;</div>
<div> <span style="white-space:pre-wrap"> </span>0)
OPT_NULL=y;;</div>
<div> <span style="white-space:pre-wrap"> </span>h|\?)
usage;;</div>
<div>@@ -69,6 +75,16 @@</div>
<div> <span style="white-space:pre-wrap"> </span>echo
"$(basename $0) error: The -c <stripe_count>
option may not" 1>&2</div>
<div> <span style="white-space:pre-wrap"> </span>echo "be
specified at the same time as the -R option."
1>&2</div>
<div> <span style="white-space:pre-wrap"> </span>exit 1</div>
<div>+elif [ "$OPT_STRIPE_COUNT" -a "$OPT_AUTOSTRIPE" ];
then</div>
<div>+<span style="white-space:pre-wrap"> </span>echo ""</div>
<div>+<span style="white-space:pre-wrap"> </span>echo
"$(basename $0) error: The -c <stripe_count>
option may not" 1>&2</div>
<div>+<span style="white-space:pre-wrap"> </span>echo "be
specified at the same time as the -A option."
1>&2</div>
<div>+<span style="white-space:pre-wrap"> </span>exit 1</div>
<div>+elif [ "$OPT_AUTOSTRIPE" -a "$OPT_RESTRIPE" ]; then</div>
<div>+<span style="white-space:pre-wrap"> </span>echo ""</div>
<div>+<span style="white-space:pre-wrap"> </span>echo
"$(basename $0) error: The -A option may not be
specified at" 1>&2</div>
<div>+<span style="white-space:pre-wrap"> </span>echo
"the same time as the -R option." 1>&2</div>
<div>+<span style="white-space:pre-wrap"> </span>exit 1</div>
<div> fi</div>
<div> </div>
<div> if [ -z "$OPT_YES" ]; then</div>
<div>@@ -107,7 +123,7 @@</div>
<div> <span style="white-space:pre-wrap"> </span>$ECHO -n
"$OLDNAME: "</div>
<div> </div>
<div> <span style="white-space:pre-wrap"> </span># avoid
duplicate stat if possible</div>
<div>-<span style="white-space:pre-wrap"> </span>TYPE_LINK=($(LANG=C
stat -c "%h %F" "$OLDNAME" || true))</div>
<div>+<span style="white-space:pre-wrap"> </span>TYPE_LINK=($(LANG=C
stat -c "%h %F %s" "$OLDNAME" || true))</div>
<div> </div>
<div> <span style="white-space:pre-wrap"> </span># skip
non-regular files, since they don't have any objects</div>
<div> <span style="white-space:pre-wrap"> </span># and
there is no point in trying to migrate them.</div>
<div>@@ -127,11 +143,6 @@</div>
<div> <span style="white-space:pre-wrap"> </span>continue</div>
<div> <span style="white-space:pre-wrap"> </span>fi</div>
<div> </div>
<div>-<span style="white-space:pre-wrap"> </span>if [
"$OPT_DRYRUN" ]; then</div>
<div>-<span style="white-space:pre-wrap"> </span>echo -e
"dry run, skipped"</div>
<div>-<span style="white-space:pre-wrap"> </span>continue</div>
<div>-<span style="white-space:pre-wrap"> </span>fi</div>
<div>-</div>
<div> <span style="white-space:pre-wrap"> </span>if [
"$OPT_RESTRIPE" ]; then</div>
<div> <span style="white-space:pre-wrap"> </span>UNLINK=""</div>
<div> <span style="white-space:pre-wrap"> </span>else</div>
<div>@@ -140,16 +151,43 @@</div>
<div> <span style="white-space:pre-wrap"> </span># then
we don't need to do this getstripe/mktemp stuff.</div>
<div> <span style="white-space:pre-wrap"> </span>UNLINK="-u"</div>
<div> </div>
<div>-<span style="white-space:pre-wrap"> </span>[
"$OPT_STRIPE_COUNT" ] && COUNT=$OPT_STRIPE_COUNT
||</div>
<div>-<span style="white-space:pre-wrap"> </span>COUNT=$($LFS
getstripe -c "$OLDNAME" \</div>
<div>-<span style="white-space:pre-wrap"> </span>2>
/dev/null)</div>
<div> <span style="white-space:pre-wrap"> </span>SIZE=$($LFS
getstripe $LFS_SIZE_OPT "$OLDNAME" \</div>
<div> <span style="white-space:pre-wrap"> </span>
2> /dev/null)</div>
<div>+<span style="white-space:pre-wrap"> </span>if [
"$OPT_AUTOSTRIPE" ]; then</div>
<div>+<span style="white-space:pre-wrap"> </span>FILE_SIZE=${TYPE_LINK[3]}<br>
</div>
<div>+<span style="white-space:pre-wrap"> </span># (math
in bash is dumb, so depend on common tools, and there
are options for that...)</div>
<div>+<span style="white-space:pre-wrap"> </span># Stripe
Count = Log2(size_in_GB)</div>
<div>+<span style="white-space:pre-wrap"> </span>#COUNT=$(echo
$FILE_SIZE | awk '{printf
"%.0f\n",log($1/1024/1024/1024)/log(2)}')</div>
<div>+<span style="white-space:pre-wrap"> </span>#COUNT=$(printf
"%.0f\n" $(echo "l($FILE_SIZE/1024/1024/1024) / l(2)" |
bc -l))</div>
<div>+<span style="white-space:pre-wrap"> </span>COUNT=$(echo
"l($FILE_SIZE/1024/1024/1024) / l(2) + 1" | bc -l | cut
-d . -f 1)</div>
<div>+<span style="white-space:pre-wrap"> </span># Stripe
Count = size_in_GB</div>
<div>+<span style="white-space:pre-wrap"> </span>#COUNT=$(echo
"scale=0; $FILE_SIZE/1024/1024/1024" | bc -l | cut -d .
-f 1)</div>
<div>+<span style="white-space:pre-wrap"> </span>[
"$COUNT" -lt 1 ] && COUNT=1</div>
<div>+<span style="white-space:pre-wrap"> </span># (does
it make sense to skip the file if old</div>
<div>+<span style="white-space:pre-wrap"> </span># and
new stripe count are identical?)</div>
<div>+<span style="white-space:pre-wrap"> </span>else</div>
<div>+<span style="white-space:pre-wrap"> </span>[
"$OPT_STRIPE_COUNT" ] && COUNT=$OPT_STRIPE_COUNT
||</div>
<div>+<span style="white-space:pre-wrap"> </span>COUNT=$($LFS
getstripe -c "$OLDNAME" \</div>
<div>+<span style="white-space:pre-wrap"> </span>2>
/dev/null)</div>
<div>+<span style="white-space:pre-wrap"> </span>fi</div>
<div> </div>
<div> <span style="white-space:pre-wrap"> </span>[ -z
"$COUNT" -o -z "$SIZE" ] && UNLINK=""</div>
<div>-<span style="white-space:pre-wrap"> </span>SIZE=${LFS_SIZE_OPT}${SIZE}</div>
<div> <span style="white-space:pre-wrap"> </span>fi</div>
<div> </div>
<div>+<span style="white-space:pre-wrap"> </span>if [
"$OPT_DRYRUN" ]; then</div>
<div>+<span style="white-space:pre-wrap"> </span>if [
"$OPT_VERBOSE" ]; then</div>
<div>+<span style="white-space:pre-wrap"> </span>echo -e
"dry run, would use count=${COUNT} size=${SIZE}"</div>
<div>+<span style="white-space:pre-wrap"> </span>else</div>
<div>+<span style="white-space:pre-wrap"> </span>echo -e
"dry run, skipped"</div>
<div>+<span style="white-space:pre-wrap"> </span>fi</div>
<div>+<span style="white-space:pre-wrap"> </span>continue</div>
<div>+<span style="white-space:pre-wrap"> </span>fi</div>
<div>+<span style="white-space:pre-wrap"> </span>if [
"$OPT_VERBOSE" ]; then</div>
<div>+<span style="white-space:pre-wrap"> </span>echo -n
"(count=${COUNT} size=${SIZE}) "</div>
<div>+<span style="white-space:pre-wrap"> </span>fi</div>
<div>+</div>
<div>+<span style="white-space:pre-wrap"> </span>[
"$SIZE" ] && SIZE=${LFS_SIZE_OPT}${SIZE}</div>
<div>+</div>
<div> <span style="white-space:pre-wrap"> </span># first
try to migrate inside lustre</div>
<div> <span style="white-space:pre-wrap"> </span># if
failed go back to old rsync mode</div>
<div> <span style="white-space:pre-wrap"> </span>if [[
$RSYNC_MODE == false ]]; then</div>
</div>
<div><br>
</div>
</div>
</div>
</div>
<br>
<fieldset></fieldset>
<br>
<pre>_______________________________________________
lustre-discuss mailing list
<a href="mailto:lustre-discuss@lists.lustre.org" target="_blank">lustre-discuss@lists.lustre.org</a>
<a href="http://lists.lustre.org/listinfo.cgi/lustre-discuss-lustre.org" target="_blank">http://lists.lustre.org/listinfo.cgi/lustre-discuss-lustre.org</a>
</pre>
</blockquote>
<br>
</div>
<br>_______________________________________________<br>
lustre-discuss mailing list<br>
<a href="mailto:lustre-discuss@lists.lustre.org">lustre-discuss@lists.lustre.org</a><br>
<a href="http://lists.lustre.org/listinfo.cgi/lustre-discuss-lustre.org" rel="noreferrer" target="_blank">http://lists.lustre.org/listinfo.cgi/lustre-discuss-lustre.org</a><br>
<br></blockquote></div><br></div></div></div>