From 2e91c83ebb6152dbed1a47e4f2c3af142d8c825a Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Johannes=20W=C3=BCnsche?= <johannes@spacesnek.rocks>
Date: Tue, 19 Sep 2023 15:46:57 +0200
Subject: [PATCH] docs: update fio example and add job tips to fio

---
 docs/src/fio-haura/run.md | 5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)

diff --git a/docs/src/fio-haura/run.md b/docs/src/fio-haura/run.md
index 0d5ef03b..75bbf40f 100644
--- a/docs/src/fio-haura/run.md
+++ b/docs/src/fio-haura/run.md
@@ -26,7 +26,7 @@ $ fio \
     --random_distribution=zipf \
     --bs=4k \
     --ioengine=external:src/fio-engine-haura.o \
-    --numjobs=2 \
+    --numjobs=1 \
     --runtime=30 \
     --time_based \
     --group_reporting \
@@ -55,6 +55,9 @@ reached.
 > - Increase the underlying block size while retaining the same `io_size`
 > - Choose a random distribution with a higher skew to specific regions (e.g.
 >   zipf) to avoid frequent evictions of nodes from the internal cache
+> - Reduce the number of jobs; More jobs put more pressure on the cache leading
+>   to more frequent evictions which lead to more writeback operations worsening
+>   fragmentation
 > 
 > As a general rule this leads to two things: reduce the amount of write
 > operations, enlarge the allocation space.