Merge PR #42106 into master

* refs/pull/42106/head:
	mds: create file system with specific ID

Reviewed-by: Patrick Donnelly <pdonnell@redhat.com>

PendingReleaseNotes

@@ -19,6 +19,11 @@
 
 * The ``device_health_metrics`` pool has been renamed ``.mgr``. It is now
   used as a common store for all ``ceph-mgr`` modules.
+
+* fs: A file system can be created with a specific ID ("fscid"). This is useful
+  in certain recovery scenarios, e.g., monitor database lost and rebuilt, and
+  the restored file system is expected to have the same ID as before.
+
 * fs: A file system can be renamed using the `fs rename` command. Any cephx
   credentials authorized for the old file system name will need to be
   reauthorized to the new file system name. Since the operations of the clients

doc/cephfs/administration.rst

@@ -395,3 +395,14 @@ This removes a rank from the failed set.
 
 This command resets the file system state to defaults, except for the name and
 pools. Non-zero ranks are saved in the stopped set.
+
+
+::
+
+    fs new <file system name> <metadata pool name> <data pool name> --fscid <fscid> --force
+
+This command creates a file system with a specific **fscid** (file system cluster ID).
+You may want to do this when an application expects the file system's ID to be
+stable after it has been recovered, e.g., after monitor databases are lost and
+rebuilt. Consequently, file system IDs don't always keep increasing with newer
+file systems.

qa/tasks/cephfs/test_admin.py

@@ -112,6 +112,7 @@ class TestFsNew(TestAdminCommands):
     """
     Test "ceph fs new" subcommand.
     """
+    MDSS_REQUIRED = 3
 
     def test_fsnames_can_only_by_goodchars(self):
         n = 'test_fsnames_can_only_by_goodchars'
@@ -213,6 +214,72 @@ class TestFsNew(TestAdminCommands):
             self.check_pool_application_metadata_key_value(
                 pool_names[i], 'cephfs', keys[i], fs_name)
 
+    def test_fs_new_with_specific_id(self):
+        """
+        That a file system can be created with a specific ID.
+        """
+        fs_name = "test_fs_specific_id"
+        fscid = 100
+        keys = ['metadata', 'data']
+        pool_names = [fs_name+'-'+key for key in keys]
+        for p in pool_names:
+            self.run_cluster_cmd(f'osd pool create {p}')
+        self.run_cluster_cmd(f'fs new {fs_name} {pool_names[0]} {pool_names[1]} --fscid  {fscid} --force')
+        self.fs.status().get_fsmap(fscid)
+        for i in range(2):
+            self.check_pool_application_metadata_key_value(pool_names[i], 'cephfs', keys[i], fs_name)
+
+    def test_fs_new_with_specific_id_idempotency(self):
+        """
+        That command to create file system with specific ID is idempotent.
+        """
+        fs_name = "test_fs_specific_id"
+        fscid = 100
+        keys = ['metadata', 'data']
+        pool_names = [fs_name+'-'+key for key in keys]
+        for p in pool_names:
+            self.run_cluster_cmd(f'osd pool create {p}')
+        self.run_cluster_cmd(f'fs new {fs_name} {pool_names[0]} {pool_names[1]} --fscid  {fscid} --force')
+        self.run_cluster_cmd(f'fs new {fs_name} {pool_names[0]} {pool_names[1]} --fscid  {fscid} --force')
+        self.fs.status().get_fsmap(fscid)
+
+    def test_fs_new_with_specific_id_fails_without_force_flag(self):
+        """
+        That command to create file system with specific ID fails without '--force' flag.
+        """
+        fs_name = "test_fs_specific_id"
+        fscid = 100
+        keys = ['metadata', 'data']
+        pool_names = [fs_name+'-'+key for key in keys]
+        for p in pool_names:
+            self.run_cluster_cmd(f'osd pool create {p}')
+        try:
+            self.run_cluster_cmd(f'fs new {fs_name} {pool_names[0]} {pool_names[1]} --fscid  {fscid}')
+        except CommandFailedError as ce:
+            self.assertEqual(ce.exitstatus, errno.EINVAL,
+                "invalid error code on creating a file system with specifc ID without --force flag")
+        else:
+            self.fail("expected creating file system with specific ID without '--force' flag to fail")
+
+    def test_fs_new_with_specific_id_fails_already_in_use(self):
+        """
+        That creating file system with ID already in use fails.
+        """
+        fs_name = "test_fs_specific_id"
+        # file system ID already in use
+        fscid =  self.fs.status().map['filesystems'][0]['id']
+        keys = ['metadata', 'data']
+        pool_names = [fs_name+'-'+key for key in keys]
+        for p in pool_names:
+            self.run_cluster_cmd(f'osd pool create {p}')
+        try:
+            self.run_cluster_cmd(f'fs new {fs_name} {pool_names[0]} {pool_names[1]} --fscid  {fscid} --force')
+        except CommandFailedError as ce:
+            self.assertEqual(ce.exitstatus, errno.EINVAL,
+                "invalid error code on creating a file system with specifc ID that is already in use")
+        else:
+            self.fail("expected creating file system with ID already in use to fail")
+
 
 class TestRenameCommand(TestAdminCommands):
     """

src/mds/FSMap.cc

@@ -449,7 +449,8 @@ mds_gid_t Filesystem::get_standby_replay(mds_gid_t who) const
 }
 
 Filesystem::ref FSMap::create_filesystem(std::string_view name,
-    int64_t metadata_pool, int64_t data_pool, uint64_t features)
+    int64_t metadata_pool, int64_t data_pool, uint64_t features,
+    fs_cluster_id_t fscid)
 {
   auto fs = Filesystem::create();
   fs->mds_map.epoch = epoch;
@@ -461,10 +462,21 @@ Filesystem::ref FSMap::create_filesystem(std::string_view name,
   fs->mds_map.created = ceph_clock_now();
   fs->mds_map.modified = ceph_clock_now();
   fs->mds_map.enabled = true;
-  fs->fscid = next_filesystem_id++;
-  // ANONYMOUS is only for upgrades from legacy mdsmaps, we should
-  // have initialized next_filesystem_id such that it's never used here.
-  ceph_assert(fs->fscid != FS_CLUSTER_ID_ANONYMOUS);
+  if (fscid == FS_CLUSTER_ID_NONE) {
+    fs->fscid = next_filesystem_id++;
+  } else {
+    fs->fscid = fscid;
+    next_filesystem_id = std::max(fscid,  (fs_cluster_id_t)next_filesystem_id) + 1;
+  }
+
+  // File system's ID can be FS_CLUSTER_ID_ANONYMOUS if we're recovering
+  // a legacy file system by passing FS_CLUSTER_ID_ANONYMOUS as the desired
+  // file system ID
+  if (fscid != FS_CLUSTER_ID_ANONYMOUS) {
+    // ANONYMOUS is only for upgrades from legacy mdsmaps, we should
+    // have initialized next_filesystem_id such that it's never used here.
+    ceph_assert(fs->fscid != FS_CLUSTER_ID_ANONYMOUS);
+  }
   filesystems[fs->fscid] = fs;
 
   // Created first filesystem?  Set it as the one

src/mds/FSMap.h

@@ -400,7 +400,8 @@ public:
    */
   Filesystem::ref create_filesystem(
       std::string_view name, int64_t metadata_pool,
-      int64_t data_pool, uint64_t features);
+      int64_t data_pool, uint64_t features,
+      fs_cluster_id_t fscid);
 
   /**
    * Remove the filesystem (it must exist).  Caller should already

src/mon/FSCommands.cc

@@ -241,6 +241,18 @@ class FsNewHandler : public FileSystemCommandHandler
       }
     }
 
+    int64_t fscid = FS_CLUSTER_ID_NONE;
+    if (cmd_getval(cmdmap, "fscid", fscid)) {
+      if (!force) {
+        ss << "Pass --force to create a file system with a specific ID";
+        return -EINVAL;
+      }
+      if (fsmap.filesystem_exists(fscid)) {
+        ss << "filesystem already exists with id '" << fscid << "'";
+        return -EINVAL;
+      }
+    }
+
     pg_pool_t const *data_pool = mon->osdmon()->osdmap.get_pg_pool(data);
     ceph_assert(data_pool != NULL);  // Checked it existed above
     pg_pool_t const *metadata_pool = mon->osdmon()->osdmap.get_pg_pool(metadata);
@@ -280,7 +292,7 @@ class FsNewHandler : public FileSystemCommandHandler
 
     // All checks passed, go ahead and create.
     auto&& fs = fsmap.create_filesystem(fs_name, metadata, data,
-        mon->get_quorum_con_features());
+        mon->get_quorum_con_features(), fscid);
 
     ss << "new fs with metadata pool " << metadata << " and data pool " << data;
 

src/mon/MonCommands.h

@@ -331,7 +331,8 @@ COMMAND("fs new "
 	" name=metadata,type=CephString "
 	"name=data,type=CephString "
 	"name=force,type=CephBool,req=false "
-	"name=allow_dangerous_metadata_overlay,type=CephBool,req=false",
+	"name=allow_dangerous_metadata_overlay,type=CephBool,req=false "
+	"name=fscid,type=CephInt,range=0,req=false",
 	"make new filesystem using named pools <metadata> and <data>",
 	"fs", "rw")
 COMMAND("fs fail "