From 93bed1b081bb23624b988627d2db6e71d55e796a Mon Sep 17 00:00:00 2001
From: Christian Hesse <mail@eworm.de>
Date: Mon, 26 Feb 2024 23:18:02 +0100
Subject: netwatch-notify: work around race condition

This used to crash every now and then with:

> script;error script error: no such item (4)

I guess this is caused by querying the dns cache just exactly when a
record expires. The chance is maximized: The script is started by
scheduler every minute, and the record's ttl is a multiple of a minute.

Let's query records that are not about to expire immediately, and
try again.
---
 netwatch-notify.rsc | 20 ++++++++++++--------
 1 file changed, 12 insertions(+), 8 deletions(-)

(limited to 'netwatch-notify.rsc')

diff --git a/netwatch-notify.rsc b/netwatch-notify.rsc
index 186ac0a..c8de7e1 100644
--- a/netwatch-notify.rsc
+++ b/netwatch-notify.rsc
@@ -80,14 +80,18 @@ $ScriptLock $0;
         :do {
           :local Resolve [ :resolve ($HostInfo->"resolve") ];
           :if ($Resolve != $HostVal->"host" && \
-               [ :len [ /ip/dns/cache/find where name=($HostInfo->"resolve") data=[ :tostr ($HostVal->"host") ] ] ] = 0) do={
-             $LogPrintExit2 info $0 ("Name '" . $HostInfo->"resolve" . [ $IfThenElse \
-                 ($HostInfo->"resolve" != $HostInfo->"name") ("' for " . $Type . " '" . \
-                 $HostInfo->"name") "" ] . "' resolves to different address " . $Resolve . \
-                 ", updating.") false;
-            /tool/netwatch/set host=$Resolve $Host;
-            :set ($Metric->"resolve-failcnt") 0;
-            :set ($HostVal->"status") "unknown";
+               [ :len [ /ip/dns/cache/find where name=($HostInfo->"resolve") data=[ :tostr ($HostVal->"host") ] ttl>0s ] ] = 0) do={
+            :delay 1500ms;
+            :resolve ($HostInfo->"resolve");
+            :if ([ :len [ /ip/dns/cache/find where name=($HostInfo->"resolve") data=[ :tostr ($HostVal->"host") ] ] ] = 0) do={
+              $LogPrintExit2 info $0 ("Name '" . $HostInfo->"resolve" . [ $IfThenElse \
+                  ($HostInfo->"resolve" != $HostInfo->"name") ("' for " . $Type . " '" . \
+                  $HostInfo->"name") "" ] . "' resolves to different address " . $Resolve . \
+                  ", updating.") false;
+              /tool/netwatch/set host=$Resolve $Host;
+              :set ($Metric->"resolve-failcnt") 0;
+              :set ($HostVal->"status") "unknown";
+            }
           }
         } on-error={
           :set ($Metric->"resolve-failcnt") ($Metric->"resolve-failcnt" + 1);
-- 
cgit v1.2.3-54-g00ecf