From: Ian Kelling <iank@fsf.org>
Date: Mon, 4 Apr 2022 06:19:52 +0000 (-0400)
Subject: cam deal with filesystem bug
X-Git-Url: https://iankelling.org/git/?p=log-quiet;a=commitdiff_plain;h=559cd04634ae77a27b830c811c38f240993c1775

cam deal with filesystem bug
---

diff --git a/sysd-prom-fail b/sysd-prom-fail
index 20e7c9b..ea84f74 100755
--- a/sysd-prom-fail
+++ b/sysd-prom-fail
@@ -66,8 +66,17 @@ fi
 if $write_count; then
   printf 'node_systemd_unit_result_fail_count{name="%s"} %s\n' "$ser_name" "$count" >$ftmp
   mv $ftmp $f
+  finaltmp=$dir/sysd-result-fail.prom.$$
+  # i had a minor problem that multiple of these scripts executed at the same time, and ended up with
+  # multiple instances of the same file
+  # ls -lai
+  # ...
+  # 3896242 -rw-r--r-- 1 root       root         439 Apr  4 00:02 sysd-result-fail.prom
+  # 3896242 -rw-r--r-- 1 root       root         439 Apr  4 00:02 sysd-result-fail.prom
+  # 3896242 -rw-r--r-- 1 root       root         439 Apr  4 00:02 sysd-result-fail.prom
+  # this seems harmless, https://www.spinics.net/lists/linux-btrfs/msg111245.html
+  # however, lets do a bit to avoid it: sleep a random amount of time from 0-.5 seconds.
+  sleep 0.$(( RANDOM % 500 ))
+  cat $dir/*-result-fail.premerge >$finaltmp
+  mv $finaltmp $dir/sysd-result-fail.prom
 fi
-
-finaltmp=$dir/sysd-result-fail.prom.$$
-cat $dir/*-result-fail.premerge >$finaltmp
-mv $finaltmp $dir/sysd-result-fail.prom