Backport43: troubleshoot public cloud setup bashlion (#2748)

* Troubleshoot public cloud setup with BYOS instances -AWS- tapasn (#2714)
* added-public-cloud-troubleshoot-document
	new file:   modules/administration/assets/images/actual-files.png
	new file:   modules/administration/assets/images/broken-symlink.png
	new file:   modules/administration/assets/images/error-1.png
	new file:   modules/administration/assets/images/incomplete-files.png
	new file:   modules/administration/assets/images/partitions.png
	new file:   modules/administration/pages/troubleshooting/tshoot-suma-installation-public-cloud.md

Co-authored-by: Karl Eichwalder <[email protected]>
Co-authored-by: Ornela Marić <[email protected]>
Co-authored-by: Bashlion(TapasN) <[email protected]>

.changelog

@@ -4,6 +4,8 @@
 # - Fixed error in Bat section of Upgrade Guide (bsc#1234567)
 # For guidelines: https://en.opensuse.org/openSUSE:Creating_a_changes_file_(RPM)#Changelog_section_.28.25changelog.29
 
+- Added troubleshooting instructions for setting up in public cloud
+ (BYOS) to Administration Guide
 - Added section about migrating Enterprise Linux (EL) clients to SUSE Liberty
   Linux to Client Configuration Guide
 - Added detailed information about the messages produced by subscription matcher
@@ -576,3 +578,4 @@
   client operating systems.
 - Documented upgrading to version 4.2 in Upgrade Guide (bsc#1185711).
 - Fixed URL of API documentation
+- Added troubleshooting instructions for the setting up in public cloud ({byos}) to Administration Guide

modules/administration/pages/troubleshooting/tshoot-suma-installation-public-cloud.md

@@ -0,0 +1,120 @@
+= Setting up SUSE Manager with separate disk for data in Public Cloud (BYOS)
+
+
+**Issue:** 
+
+{productname} setup results with an error while restarting services during setup. 
+
+
+The setup fails on restarting Tomcat.
+
+
+
+<img src="../../assets/images/error-1.png">
+
+
+
+**Reason:**
+
+When {productname} is set up using separate disks for cache, pgsql and spacewalk, the existing data within these partitions are covered under the new mount point.
+
+
+ <img src="../../assets/images/partitions.png">
+
+Because of this, data for certain services like tomcat go missing due to a broken symbolic link.
+
+
+<img src="../../assets/images/broken-symlink.png"> 
+
+As {productname} instance of cloud already has some services set up, when the data is not available during set up it causes the service to not restart as expected.
+Therefore the setup fails.
+
+
+ <img src="../../assets/images/incomplete-files.png"> 
+
+Actual data which should be present:
+
+
+ <img src="../../assets/images/actual-files.png"> 
+
+
+
+**How to Recreate:**
+
+1. {productname} Server provisioned in {aws} uses relevant AMI for {byos}
+
+
+2. Separate storage for spacewalk, {postgresql} and cache. 
+  The size of the partition depends on the repository to be synchronized.
+
+
+3. Set up {productname} using [command]``yast2 susemanager_setup``.
+
+
+
+
+**Resolution:**
+
+To resolve or avoid the issue, copy the data of the underlying partition into the new partition by following the steps below:
+
+
+. Create a backup partition
++
+----
+mkdir /bkp
+----
+
+
+. Unmount the affected partition for example /var/cache. 
++
+----
+umount /var/cache
+----
+
+
+. Check the underlying data in the partition. 
+  This should list the actual data present in the directory.
++
+----
+ll /var/cache
+----
+
+. Copy the data onto the backup partition created in step 1.
++
+----
+cp -prav /var/cache/* /bkp/
+----
+
+
+. Mount the partition again using option 1 or 2:
++
+----
+(option 1)   /dev/nvme0nxx /var/cache
+(option 2)   mount -a (if you have the entries under /etc/fstab)
+----
+
+
+. Copy the backed up data from step 4 to the mounted partition
++
+----
+cp -prav /bkp/* /var/cache/
+----
+
+. Verify the data is present
++
+----
+ll /var/cache/
+----
+
+. Repeat the {productname} setup
++
+----
+yast2 susemanager_setup
+----
+
+
+
+**Conclusion**:
+
+When using {procustname} in public cloud with separate data partitions, make sure that the existing data in those directories are copied onto the new partition.
+