From 54e21b1a6f4f83a957c31d0bb93b8c7a8e256edd Mon Sep 17 00:00:00 2001
From: Marina Gourtovaia <mg8@sanger.ac.uk>
Date: Thu, 25 Jan 2024 14:34:42 +0000
Subject: [PATCH] Added logging to help monitor shadow run folderes.

Added runfolder path to some of log messages so that it is
easier to search the logs.

Added comments to the staging monitor scripts to highlight
that the script is likely to discover and process shadow
runfolders.
---
 bin/staging_area_monitor         | 19 ++++++++++++++++++-
 lib/Monitor/RunFolder.pm         |  2 +-
 lib/Monitor/RunFolder/Staging.pm |  6 +++---
 3 files changed, 22 insertions(+), 5 deletions(-)

diff --git a/bin/staging_area_monitor b/bin/staging_area_monitor
index 1ed21c40..93e98d34 100755
--- a/bin/staging_area_monitor
+++ b/bin/staging_area_monitor
@@ -55,7 +55,24 @@ sub main {
     while (1) {
 
         _log_noindent($time->());
-
+       
+        #######
+        # find_live finds all run folders in 'incoming and 'analysis'.
+        # It can return two run folders for the same run, one in each
+        # of 'incoming and 'analysis'. In this case the  run folder in
+        # incoming is a 'shadow' run folder, which is sometimes created
+        # by the instrument well after the original run.
+        # 
+        # It is also possible for a 'shadow' folder to exist in 'incoming',
+        # while the run folder with the original data is already in 'outgoing'.
+        #
+        # Depending on how much data is present in a 'shadow' run folder,
+        # the staging monitor might attempt to update the trackign database.
+        # Mostly the values will be the same as for teh main run folder.
+        # However, the actual cycle count might be different. Therefore, it is
+        # important never to upload to the database a cycle count that is
+        # lower than the one already stored in the database.
+        #  
         foreach my $run_path ( $monitor->find_live($staging_area) ) {
             _log_noindent("Considering $run_path");
             my $done;
diff --git a/lib/Monitor/RunFolder.pm b/lib/Monitor/RunFolder.pm
index 5d9dcb25..14cfadbe 100644
--- a/lib/Monitor/RunFolder.pm
+++ b/lib/Monitor/RunFolder.pm
@@ -23,7 +23,7 @@ sub update_cycle_count {
   defined $latest_cycle or croak 'Latest cycle count not supplied';
   my $actual_cycle = $self->tracking_run()->actual_cycle_count();
   $actual_cycle ||= 0;
-  if ($latest_cycle > $actual_cycle) {
+  if ($latest_cycle > $actual_cycle) { # A very important condition!
     $self->tracking_run()->update({actual_cycle_count => $latest_cycle});
     return 1;
   }
diff --git a/lib/Monitor/RunFolder/Staging.pm b/lib/Monitor/RunFolder/Staging.pm
index 3b6802b1..2718c09a 100644
--- a/lib/Monitor/RunFolder/Staging.pm
+++ b/lib/Monitor/RunFolder/Staging.pm
@@ -368,7 +368,7 @@ sub update_run_from_path { ##### Method to deal with one run folder
     }
 
     if ($folder->is_in_analysis) {
-        _log('Folder is in /analysis/');
+        _log("Folder $run_path is in /analysis/");
 
         if( $run_status eq 'qc complete') {
             _log('Moving run folder to /outgoing/');
@@ -377,7 +377,7 @@ sub update_run_from_path { ##### Method to deal with one run folder
         return; # Nothing else to do for a folder in /analysis/
     }
 
-    _log('Folder is in /incoming/');
+    _log("Folder $run_path is in /incoming/");
 
     # If we don't remember seeing it before, set the folder name and glob;
     # set staging tag, if appropriate, set/fix instrument side, workflow side.
@@ -411,7 +411,7 @@ sub update_run_from_path { ##### Method to deal with one run folder
         
         my $latest_cycle = $folder->get_latest_cycle();
         if ($folder->update_cycle_count($latest_cycle)) {
-            _log("Cycle count updated to $latest_cycle");
+            _log("Cycle count updated to $latest_cycle for $run_path");
         }
         
         if ( $run_status eq 'run pending' )  {