initial code repo

This patch creates initial code repo.

For ceph, luminous stable release will be used for base code,
and next changes and optimization for ceph will be added to it.

For opensds, currently any changes can be upstreamed into original
opensds repo (https://github.com/opensds/opensds), and so stor4nfv
will directly clone opensds code to deploy stor4nfv environment.
And the scripts for deployment based on ceph and opensds will be
put into 'ci' directory.

Change-Id: I46a32218884c75dda2936337604ff03c554648e4
Signed-off-by: Qiaowei Ren <qiaowei.ren@intel.com>

1 files changed, 80 insertions, 0 deletions

diff --git a/src/ceph/qa/tasks/mon_recovery.py b/src/ceph/qa/tasks/mon_recovery.py
new file mode 100644
index 0000000..bfa2cdf
--- /dev/null
+++ b/src/ceph/qa/tasks/mon_recovery.py
@@ -0,0 +1,80 @@
+"""
+Monitor recovery
+"""
+import logging
+import ceph_manager
+from teuthology import misc as teuthology
+
+
+log = logging.getLogger(__name__)
+
+def task(ctx, config):
+    """
+    Test monitor recovery.
+    """
+    if config is None:
+        config = {}
+    assert isinstance(config, dict), \
+        'task only accepts a dict for configuration'
+    first_mon = teuthology.get_first_mon(ctx, config)
+    (mon,) = ctx.cluster.only(first_mon).remotes.iterkeys()
+
+    manager = ceph_manager.CephManager(
+        mon,
+        ctx=ctx,
+        logger=log.getChild('ceph_manager'),
+        )
+
+    mons = [f.split('.')[1] for f in teuthology.get_mon_names(ctx)]
+    log.info("mon ids = %s" % mons)
+
+    manager.wait_for_mon_quorum_size(len(mons))
+
+    log.info('verifying all monitors are in the quorum')
+    for m in mons:
+        s = manager.get_mon_status(m)
+        assert s['state'] == 'leader' or s['state'] == 'peon'
+        assert len(s['quorum']) == len(mons)
+
+    log.info('restarting each monitor in turn')
+    for m in mons:
+        # stop a monitor
+        manager.kill_mon(m)
+        manager.wait_for_mon_quorum_size(len(mons) - 1)
+
+        # restart
+        manager.revive_mon(m)
+        manager.wait_for_mon_quorum_size(len(mons))
+
+    # in forward and reverse order,
+    rmons = mons
+    rmons.reverse()
+    for mons in mons, rmons:
+        log.info('stopping all monitors')
+        for m in mons:
+            manager.kill_mon(m)
+
+        log.info('forming a minimal quorum for %s, then adding monitors' % mons)
+        qnum = (len(mons) / 2) + 1
+        num = 0
+        for m in mons:
+            manager.revive_mon(m)
+            num += 1
+            if num >= qnum:
+                manager.wait_for_mon_quorum_size(num)
+
+    # on both leader and non-leader ranks...
+    for rank in [0, 1]:
+        # take one out
+        log.info('removing mon %s' % mons[rank])
+        manager.kill_mon(mons[rank])
+        manager.wait_for_mon_quorum_size(len(mons) - 1)
+
+        log.info('causing some monitor log activity')
+        m = 30
+        for n in range(1, m):
+            manager.raw_cluster_cmd('log', '%d of %d' % (n, m))
+
+        log.info('adding mon %s back in' % mons[rank])
+        manager.revive_mon(mons[rank])
+        manager.wait_for_mon_quorum_size(len(mons))