From 3759664e49a31bfe176290ed84c8ccf378ce1a71 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?J=C3=B6rg=20Sachse?= <Joerg.Sachse@slub-dresden.de>
Date: Tue, 31 May 2022 09:44:15 +0200
Subject: [PATCH] doc: add ADR for why this role even exists

---
 doc/adr/init.asciidoc | 32 ++++++++++++++++++++++++++++++++
 1 file changed, 32 insertions(+)
 create mode 100644 doc/adr/init.asciidoc

diff --git a/doc/adr/init.asciidoc b/doc/adr/init.asciidoc
new file mode 100644
index 0000000..7371d8f
--- /dev/null
+++ b/doc/adr/init.asciidoc
@@ -0,0 +1,32 @@
+// architecture decision record
+:lang: de
+:encoding: utf-8
+:date: 2022-05-30
+:author: Jörg Sachse
+
+// Hinweis: Bitte kurz fassen!
+
+Reason for creating this role
+=============================
+
+== DECISION
+
+// Describe the decision in one short sentence.
+Rosetta NFS-Shares will be backed up using a backup proxy server that can mount the shares in read-only mode and copy their contents using IBMSP.
+
+
+== Status
+
+// Status: Proposed -> Accepted -> Superseded
+//                  -> Declined
+Accepted
+
+== Context
+
+// Reasons and constraints that led to the decision
+Severs in a Rosetta cluster share some NFS shares (among others, these are /operational_shared, /operational_storage und /deposit_storage). In production use, redundant backups would be created if these shares would be backed up from each server individually using the IBMSP client.
+
+== Consequenzes
+
+// What follows from the decision?
+Duplicate backups will be prevented without having to use storage-side deduplication, which would create additional server load. Also, the Rosetta servers will not have to perform the backups themselves, reducing their load as well. Backup creation can be monitored easier and the backup proxy server itself can easily be backed up as well and reinstalled in case of a disaster just using the Ansible role.
-- 
GitLab