[lustre-discuss] trimming flash-based external journal device

[Nathan Dauchy - NOAA Affiliate]

Greetings ext4 and flash storage experts!

Motivation:  We have ldiskfs OSTs that are primarily HDDs and use a Flash
device for an external journal device.  Recent IOR benchmarks showed that
write performance dropped (suddenly?) to about 25% of the original
baseline, yet read performance remains fine.  We saw similar
characteristics at one point in the past when the OSTs were mounted without
the external flash journal enabled.  I verified that the journals are
currently enabled, but the past experience still led me to question whether
the journals were performing well.

Question:  Is it possible that a flash journal device on an ext4 filesystem
can reach a point where there are not enough clean blocks to write to, and
they can suffer from very degraded write performance?

I know that "fstrim" can be run for mounted ldiskfs file systems, but when
I try that it doesn't see the OSTs as using flash, because they are
primarily HDD-based.  Is there some other way to tell the system which
blocks can be discarded on the journal flash device?  (I found "blkdiscard"
but that seems heavyweight and dangerous.)

Another related question would be how to benchmark the journal device on
it's own, particularly write performance, without losing data on an
existing file system; similar to the very useful obdfilter-survey tool, but
at a lower level.  But I am primarily looking to understand the nuances of
flash devices and ldiskfs external journals a bit better.

Thanks in advance for any references, tuning suggestions and other info!!!

-Nathan
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.lustre.org/pipermail/lustre-discuss-lustre.org/attachments/20210805/93cd1900/attachment.html>