Merge pull request #1106

Fix stale volumes/devices in bareos-sd after failure to acquire a read device on MAC jobs

CHANGELOG.md

@@ -25,6 +25,7 @@ and since Bareos version 20 this project adheres to [Semantic Versioning](https:
 - webui: fix deprecated notice required param follows optional param [PR #1097]
 - webui: fix uncaught TypeError if node.data is null [PR #1087]
 - core cats: Add DROP VIEWS instruction in drop_bareos_table script [PR #1092]
+- Don't keep volume open after acquiring a read-storage failed in migrate/copy/virtual full [PR #1106]
 
 ### Added
 - Python plugins: add default module_path to search path [PR #1038]

core/src/dird/migrate.cc

@@ -1495,11 +1495,13 @@ static inline bool DoActualMigration(JobControlRecord* jcr)
     WaitForStorageDaemonTermination(jcr);
     WaitForStorageDaemonTermination(mig_jcr);
     jcr->setJobStatus(jcr->impl->SDJobStatus);
-    mig_jcr->db_batch->WriteBatchFileRecords(mig_jcr);
+    if (mig_jcr->batch_started) {
+      mig_jcr->db_batch->WriteBatchFileRecords(mig_jcr);
+    }
   } else {
     WaitForStorageDaemonTermination(jcr);
     jcr->setJobStatus(jcr->impl->SDJobStatus);
-    jcr->db_batch->WriteBatchFileRecords(jcr);
+    if (jcr->batch_started) { jcr->db_batch->WriteBatchFileRecords(jcr); }
   }
 
 bail_out:

core/src/stored/mac.cc

@@ -3,7 +3,7 @@
 
    Copyright (C) 2006-2012 Free Software Foundation Europe e.V.
    Copyright (C) 2011-2012 Planets Communications B.V.
-   Copyright (C) 2013-2021 Bareos GmbH & Co. KG
+   Copyright (C) 2013-2022 Bareos GmbH & Co. KG
 
    This program is Free Software; you can redistribute it and/or
    modify it under the terms of version three of the GNU Affero General Public
@@ -686,16 +686,17 @@ bool DoMacRun(JobControlRecord* jcr)
          job_elapsed / 3600, job_elapsed % 3600 / 60, job_elapsed % 60,
          edit_uint64_with_suffix(jcr->JobBytes / job_elapsed, ec1));
 
-    // Release the device -- and send final Vol info to DIR
-    ReleaseDevice(jcr->impl->dcr);
-
+    // send final Vol info to DIR
     if (!ok || JobCanceled(jcr)) {
       DiscardAttributeSpool(jcr);
     } else {
       CommitAttributeSpool(jcr);
     }
   }
 
+  if (!jcr->impl->remote_replicate && jcr->impl->dcr) {
+    if (!ReleaseDevice(jcr->impl->dcr)) { ok = false; }
+  }
   if (jcr->impl->read_dcr) {
     if (!ReleaseDevice(jcr->impl->read_dcr)) { ok = false; }
   }

systemtests/tests/CMakeLists.txt

@@ -1,6 +1,6 @@
 #   BAREOS® - Backup Archiving REcovery Open Sourced
 #
-#   Copyright (C) 2021-2021 Bareos GmbH & Co. KG
+#   Copyright (C) 2021-2022 Bareos GmbH & Co. KG
 #
 #   This program is Free Software; you can redistribute it and/or
 #   modify it under the terms of version three of the GNU Affero General Public
@@ -31,8 +31,8 @@ add_subdirectory(chflags)
 add_subdirectory(client-initiated)
 add_subdirectory(config-dump)
 add_subdirectory(config-syntax-crash)
+add_subdirectory(copy)
 add_subdirectory(copy-archive-job)
-add_subdirectory(copy-bscan)
 add_subdirectory(copy-remote-bscan)
 add_subdirectory(dbcopy-mysql-postgresql)
 add_subdirectory(deprecation)

systemtests/tests/copy-bscan/CMakeLists.txt → systemtests/tests/copy/CMakeLists.txt

@@ -1,6 +1,6 @@
 #   BAREOS® - Backup Archiving REcovery Open Sourced
 #
-#   Copyright (C) 2021-2021 Bareos GmbH & Co. KG
+#   Copyright (C) 2021-2022 Bareos GmbH & Co. KG
 #
 #   This program is Free Software; you can redistribute it and/or
 #   modify it under the terms of version three of the GNU Affero General Public
@@ -19,3 +19,26 @@
 
 get_filename_component(BASENAME ${CMAKE_CURRENT_BINARY_DIR} NAME)
 create_systemtest(${SYSTEMTEST_PREFIX} ${BASENAME})
+
+set_tests_properties(
+  system:copy:01-backup PROPERTIES FIXTURES_SETUP "system:copy:01-backup"
+                                   FIXTURES_REQUIRED "system:copy-fixture"
+)
+set_tests_properties(
+  system:copy:02-copy
+  PROPERTIES FIXTURES_SETUP "system:copy:02-copy" FIXTURES_REQUIRED
+             "system:copy:01-backup;system:copy-fixture"
+)
+set_tests_properties(
+  system:copy:03-impossible-copy
+  PROPERTIES FIXTURES_SETUP "system:copy:03-impossible-copy" FIXTURES_REQUIRED
+             "system:copy:02-copy;system:copy:01-backup;system:copy-fixture"
+)
+set_tests_properties(
+  system:copy:04-bscan
+  PROPERTIES
+    FIXTURES_SETUP
+    "system:copy:04-bscan"
+    FIXTURES_REQUIRED
+    "system:copy:03-impossible-copy;system:copy:02-copy;system:copy:01-backup;system:copy-fixture"
+)

systemtests/tests/copy/etc/bareos/bareos-dir.d/storage/NoSuchStorage.conf

@@ -0,0 +1,7 @@
+Storage {
+  Name = NoSuchStorage
+  Address = localhost
+  Password = "does not matter"
+  Device = NoSuchDevice
+  Media Type = NoSuchType
+}

systemtests/tests/copy/testrunner-01-backup

@@ -0,0 +1,46 @@
+#!/bin/bash
+set -e
+set -o pipefail
+set -u
+
+TestName=01-backup
+
+#shellcheck source=../environment.in
+. ./environment
+
+#shellcheck source=../scripts/functions
+. "${rscripts}"/functions
+
+log="$tmp/$TestName.log"
+
+check_log() {
+  if ! grep -q "$@" "$log"; then
+    echo "'$*' not found in $log" >&2
+    estat=1
+  fi
+}
+
+setup_data
+
+start_test
+
+cat <<END_OF_DATA >"$tmp/bconcmds"
+@$out /dev/null
+messages
+@$out $log
+label volume=TestVolume001 storage=File pool=Full
+run job=backup-bareos-fd yes
+wait
+update volume=TestVolume001 volstatus=Used
+messages
+quit
+END_OF_DATA
+
+run_bconsole "$tmp/bconcmds"
+
+check_for_zombie_jobs storage=File
+
+check_log -F 'Catalog record for Volume "TestVolume001", Slot 0 successfully created.'
+check_log 'Termination:.*Backup OK'
+
+end_test

systemtests/tests/copy/testrunner-02-copy

@@ -0,0 +1,43 @@
+#!/bin/bash
+set -e
+set -o pipefail
+set -u
+
+TestName=02-copy
+
+#shellcheck source=../environment.in
+. ./environment
+
+#shellcheck source=../scripts/functions
+. "${rscripts}"/functions
+
+log="$tmp/$TestName.log"
+
+check_log() {
+  if ! grep -q "$@" "$log"; then
+    echo "'$*' not found in $log" >&2
+    estat=1
+  fi
+}
+
+start_test
+
+cat <<END_OF_DATA >"$tmp/bconcmds"
+@$out /dev/null
+messages
+@$out $log
+label volume=TestCopyVolume001 storage=File2 pool=FullCopy
+run copy yes
+wait
+messages
+quit
+END_OF_DATA
+
+run_bconsole "$tmp/bconcmds"
+
+check_for_zombie_jobs storage=File
+
+check_log -F 'Catalog record for Volume "TestCopyVolume001", Slot 0 successfully created.'
+check_log 'Termination:.*Copying OK'
+
+end_test

systemtests/tests/copy/testrunner-03-impossible-copy

@@ -0,0 +1,56 @@
+#!/bin/bash
+set -e
+set -o pipefail
+set -u
+
+TestName=03-impossible-copy
+
+#shellcheck source=../environment.in
+. ./environment
+
+#shellcheck source=../scripts/functions
+. "${rscripts}"/functions
+
+log="$tmp/$TestName.log"
+
+check_logged() {
+  if ! grep -q "$@" "$log"; then
+    echo "'$*' not found in $log" >&2
+    estat=1
+  fi
+}
+check_not_logged() {
+  if grep -q "$@" "$log"; then
+    echo "'$*' found in $log" >&2
+    estat=1
+  fi
+}
+
+start_test
+
+cat <<END_OF_DATA >"$tmp/bconcmds"
+@$out /dev/null
+messages
+@$out $log
+sqlquery
+UPDATE Media SET mediatype = 'NoSuchType' WHERE volumename = 'TestVolume001';
+
+
+run copy yes
+wait
+messages
+
+status storage=File2
+quit
+END_OF_DATA
+
+run_bconsole "$tmp/bconcmds"
+
+check_for_zombie_jobs storage=File
+
+check_logged -F 'Changing read device. Want Media Type="NoSuchType"'
+check_logged -F '*** Copying Error ***'
+# check if there is a stale volume in the storage's volume list
+check_not_logged -F 'on device "FileStorage2" (storage)'
+
+end_test