From ccbb6acc51d0ca50026fc7ac79ed52ce80d55ae6 Mon Sep 17 00:00:00 2001
From: Jakub Kadlcik <frostyx@email.cz>
Date: Thu, 21 Nov 2024 12:27:49 +0100
Subject: [PATCH] backend: implement a check for every storage whether a
 repository is available

Otherwise, builds in Pulp projects will fail with this:

    Waiting for copr_base repository
    Waiting for copr_base repository
    Backend process error: Giving up waiting for copr_base repository

So far it worked only because we didn't create Pulp projects directly
but rather migrated existing projects into Pulp and therefore this
path existed on the backend.
---
 .../copr_backend/background_worker_build.py   |  3 +-
 backend/copr_backend/storage.py               | 31 +++++++++++++++++++
 2 files changed, 32 insertions(+), 2 deletions(-)

diff --git a/backend/copr_backend/background_worker_build.py b/backend/copr_backend/background_worker_build.py
index 451d7e1f8..128e899b9 100644
--- a/backend/copr_backend/background_worker_build.py
+++ b/backend/copr_backend/background_worker_build.py
@@ -348,10 +348,9 @@ def _wait_for_repo(self):
             # we don't need copr_base repodata for srpm builds
             return
 
-        repodata = os.path.join(self.job.chroot_dir, "repodata/repomd.xml")
         waiting_since = time.time()
         while time.time() - waiting_since < 60:
-            if os.path.exists(repodata):
+            if self.storage.repository_exists(self.job.chroot):
                 return
 
             # Either (a) the very first copr-repo run in this chroot dir
diff --git a/backend/copr_backend/storage.py b/backend/copr_backend/storage.py
index 1d816b6c1..7c40ef63c 100644
--- a/backend/copr_backend/storage.py
+++ b/backend/copr_backend/storage.py
@@ -5,6 +5,8 @@
 import os
 import json
 import shutil
+from urllib.parse import urlparse
+import requests
 from copr_common.enums import StorageEnum
 from copr_backend.helpers import call_copr_repo, build_chroot_log_name
 from copr_backend.pulp import PulpClient
@@ -85,6 +87,12 @@ def delete_builds(self, dirname, chroot_builddirs, build_ids):
         """
         raise NotImplementedError
 
+    def repository_exists(self, chroot):
+        """
+        Does a repository exist?
+        """
+        raise NotImplementedError
+
 
 class BackendStorage(Storage):
     """
@@ -172,6 +180,11 @@ def delete_builds(self, dirname, chroot_builddirs, build_ids):
                         self.log.debug("can't remove %s", log_path)
         return result
 
+    def repository_exists(self, chroot):
+        repodata = os.path.join(self.opts.destdir, self.owner, self.project,
+                                chroot, "repodata", "repomd.xml")
+        return os.path.exists(repodata)
+
 
 class PulpStorage(Storage):
     """
@@ -342,6 +355,24 @@ def delete_builds(self, dirname, chroot_builddirs, build_ids):
 
         return result
 
+    def repository_exists(self, chroot):
+        name = self._distribution_name(chroot)
+        response = self.client.get_distribution(name)
+        if not response.ok:
+            return False
+        distribution = response.json()["results"][0]
+
+        # For some instances (local container) the distribution base_url
+        # contains only path, for some instances (hosted STG) it returns fully
+        # qualified URL. The problem is that there is a lot of magic in the
+        # hosted deployment in order to provide the data publicly without
+        # a Red Hat login. And we cannot use the returned URL, only its path.
+        path = urlparse(distribution["base_url"]).path.lstrip("/")
+        host = self.client.config["base_url"].rstrip("/")
+        repodata = "{0}/{1}/repodata/repomd.xml".format(host, path)
+        response = requests.head(repodata)
+        return response.ok
+
     def _repository_name(self, chroot, dirname=None):
         return "/".join([
             self.owner,