From c48ef696622a0aef5b753d88ccf798e195d44c9c Mon Sep 17 00:00:00 2001
From: eikek <701128+eikek@users.noreply.github.com>
Date: Thu, 26 Sep 2024 15:14:41 +0200
Subject: [PATCH] Reset lock documents on startup (#211)

In the situation the services crashes or gets killed while a lock is
being held, it would startup into an incorrect state. On startup, the
lock documents are resetted to recover from such a case.

This can be problematic when many services are deployed, because it
would reset all documents right now. This will be addressed in a
future change and for the currently foreseeable future, there will be
only one instance.
---
 .../scala/io/renku/search/provision/Microservice.scala    | 8 ++++++--
 1 file changed, 6 insertions(+), 2 deletions(-)

diff --git a/modules/search-provision/src/main/scala/io/renku/search/provision/Microservice.scala b/modules/search-provision/src/main/scala/io/renku/search/provision/Microservice.scala
index 599a2f88..a1c3a26e 100644
--- a/modules/search-provision/src/main/scala/io/renku/search/provision/Microservice.scala
+++ b/modules/search-provision/src/main/scala/io/renku/search/provision/Microservice.scala
@@ -44,8 +44,12 @@ object Microservice extends IOApp:
           logger
             .warn(s"There were ${migrateResult.migrationsSkipped} skipped migrations!")
         )
-        // this is only safe for a single provisioning service
-        _ <- services.resetLockDocuments
+        // this is really only safe for a single provisioning service,
+        // but for the immediate future that is the situation. So do
+        // this to recover from crashes during a lock is held
+        _ <- services.resetLockDocuments.handleErrorWith { ex =>
+          logger.warn(s"Resetting locks on start failed", ex)
+        }
         registryBuilder = CollectorRegistryBuilder[IO].withJVMMetrics
           .add(RedisMetrics.queueSizeGauge)
           .add(RedisMetrics.unprocessedGauge)