teuthology 1.0.0__py3-none-any.whl → 1.2.0__py3-none-any.whl

scripts/describe.py

@@ -70,6 +70,7 @@ options only for describing combinations represented by a suite:
                                      operator, where negative value used for
                                      a random seed
                                      [default: -1]
+  --no-nested-subset                 Disable nested subsets
 """
 
 

scripts/dispatcher.py

@@ -0,0 +1,62 @@
+import argparse
+import sys
+
+import teuthology.dispatcher.supervisor
+
+from .supervisor import parse_args as parse_supervisor_args
+
+
+def parse_args(argv):
+    parser = argparse.ArgumentParser(
+        description="Start a dispatcher for the specified tube. Grab jobs from a beanstalk queue and run the teuthology tests they describe as subprocesses. The subprocess invoked is teuthology-supervisor."
+    )
+    parser.add_argument(
+        "-v",
+        "--verbose",
+        action="store_true",
+        help="be more verbose",
+    )
+    parser.add_argument(
+        "-a",
+        "--archive-dir",
+        type=str,
+        help="path to archive results in",
+    )
+    parser.add_argument(
+        "-t",
+        "--tube",
+        type=str,
+        help="which beanstalk tube to read jobs from",
+        required=True,
+    )
+    parser.add_argument(
+        "-l",
+        "--log-dir",
+        type=str,
+        help="path in which to store the dispatcher log",
+        required=True,
+    )
+    parser.add_argument(
+        "--exit-on-empty-queue",
+        action="store_true",
+        help="if the queue is empty, exit",
+    )
+    return parser.parse_args(argv)
+
+
+def main():
+    if "--supervisor" in sys.argv:
+        # This is for transitional compatibility, so the old dispatcher can
+        # invoke the new supervisor. Once old dispatchers are phased out,
+        # this block can be as well.
+        sys.argv.remove("--supervisor")
+        sys.argv[0] = "teuthology-supervisor"
+        sys.exit(teuthology.dispatcher.supervisor.main(
+            parse_supervisor_args(sys.argv[1:])
+        ))
+    else:
+        sys.exit(teuthology.dispatcher.main(parse_args(sys.argv[1:])))
+
+
+if __name__ == "__main__":
+    main()

scripts/exporter.py

@@ -0,0 +1,18 @@
+import docopt
+
+import teuthology.exporter
+
+doc = """
+usage: teuthology-exporter --help
+       teuthology-exporter [--interval INTERVAL]
+
+optional arguments:
+  -h, --help                     show this help message and exit
+  --interval INTERVAL            update metrics this often, in seconds
+                                 [default: 60]
+"""
+
+
+def main():
+    args = docopt.docopt(doc)
+    teuthology.exporter.main(args)

scripts/lock.py

@@ -27,8 +27,8 @@ def parse_args(argv):
             teuthology-lock --lock-many 1 --machine-type vps
             teuthology-lock --lock -t target.yaml
             teuthology-lock --list-targets plana01
-            teuthology-lock --list --brief --owner user@host
             teuthology-lock --brief
+            teuthology-lock --brief --owner user@host
             teuthology-lock --update --status down --desc testing plana01
         '''),
         formatter_class=argparse.RawTextHelpFormatter)

scripts/node_cleanup.py

@@ -0,0 +1,58 @@
+import argparse
+import logging
+import sys
+
+import teuthology
+from teuthology.lock import query, ops
+
+def main():
+    args = parse_args(sys.argv[1:])
+    if args.verbose:
+        teuthology.log.setLevel(logging.DEBUG)
+    log = logging.getLogger(__name__)
+    stale = query.find_stale_locks(args.owner)
+    if not stale:
+        return
+    by_owner = {}
+    for node in stale:
+        if args.owner and node['locked_by'] != args.owner:
+            log.warning(
+                f"Node {node['name']} expected to be locked by {args.owner} "
+                f"but found {node['locked_by']} instead"
+            )
+            continue
+        by_owner.setdefault(node['locked_by'], []).append(node)
+    if args.dry_run:
+        log.info("Would attempt to unlock:")
+        for owner, nodes in by_owner.items():
+            for node in nodes:
+                log.info(f"{node['name']}\t{node['description']}")
+    else:
+        for owner, nodes in by_owner.items():
+            ops.unlock_safe([node["name"] for node in nodes], owner)
+
+def parse_args(argv):
+    parser = argparse.ArgumentParser(
+        description="Find and unlock nodes that are still locked by jobs that are no "
+            "longer active",
+    )
+    parser.add_argument(
+        '-v', '--verbose',
+        action='store_true',
+        default=False,
+        help='Be more verbose',
+    )
+    parser.add_argument(
+        '--dry-run',
+        action='store_true',
+        default=False,
+        help="List nodes that would be unlocked if the flag were omitted",
+    )
+    parser.add_argument(
+        '--owner',
+        help='Optionally, find nodes locked by a specific user',
+    )
+    return parser.parse_args(argv)
+
+if __name__ == "__main__":
+    main()

scripts/openstack.py

@@ -49,7 +49,7 @@ def get_suite_parser():
     parser.add_argument(
         '-c', '--ceph',
         help='The ceph branch to run against',
-        default=os.getenv('TEUTH_CEPH_BRANCH', 'master'),
+        default=os.getenv('TEUTH_CEPH_BRANCH', 'main'),
     )
     parser.add_argument(
         '-k', '--kernel',
@@ -58,9 +58,9 @@ def get_suite_parser():
     )
     parser.add_argument(
         '-f', '--flavor',
-        help=("The kernel flavor to run against: ('basic',"
-              "'gcov', 'notcmalloc')"),
-        default='basic',
+        help=("The ceph packages shaman flavor to run with:"
+        "('default', 'crimson', 'notcmalloc', 'jaeger')"),
+        default='default',
     )
     parser.add_argument(
         '-d', '--distro',
@@ -69,7 +69,7 @@ def get_suite_parser():
     parser.add_argument(
         '--suite-branch',
         help='Use this suite branch instead of the ceph branch',
-        default=os.getenv('TEUTH_SUITE_BRANCH', 'master'),
+        default=os.getenv('TEUTH_SUITE_BRANCH', 'main'),
     )
     parser.add_argument(
         '-e', '--email',
@@ -251,8 +251,8 @@ def get_openstack_parser():
     )
     parser.add_argument(
         '--teuthology-branch',
-        help="use this teuthology branch instead of master",
-        default=os.getenv('TEUTH_BRANCH', 'master'),
+        help="use this teuthology branch instead of main",
+        default=os.getenv('TEUTH_BRANCH', 'main'),
     )
     parser.add_argument(
         '--ceph-workbench-git-url',
@@ -260,8 +260,8 @@ def get_openstack_parser():
     )
     parser.add_argument(
         '--ceph-workbench-branch',
-        help="use this ceph-workbench branch instead of master",
-        default='master',
+        help="use this ceph-workbench branch instead of main",
+        default='main',
     )
     parser.add_argument(
         '--upload',

scripts/results.py

@@ -1,19 +1,20 @@
 """
-usage: teuthology-results [-h] [-v] [--dry-run] [--email EMAIL] [--timeout TIMEOUT] --archive-dir DIR --name NAME [--subset SUBSET] [--seed SEED]
+usage: teuthology-results [-h] [-v] [--dry-run] [--email EMAIL] [--timeout TIMEOUT] --archive-dir DIR --name NAME [--subset SUBSET] [--seed SEED] [--no-nested-subset]
 
 Email teuthology suite results
 
 optional arguments:
-  -h, --help         show this help message and exit
-  -v, --verbose      be more verbose
-  --dry-run          Instead of sending the email, just print it
-  --email EMAIL      address to email test failures to
-  --timeout TIMEOUT  how many seconds to wait for all tests to finish
-                     [default: 0]
-  --archive-dir DIR  path under which results for the suite are stored
-  --name NAME        name of the suite
-  --subset SUBSET    subset passed to teuthology-suite
-  --seed SEED        random seed used in teuthology-suite
+  -h, --help            show this help message and exit
+  -v, --verbose         be more verbose
+  --dry-run             Instead of sending the email, just print it
+  --email EMAIL         address to email test failures to
+  --timeout TIMEOUT     how many seconds to wait for all tests to finish
+                        [default: 0]
+  --archive-dir DIR     path under which results for the suite are stored
+  --name NAME           name of the suite
+  --subset SUBSET       subset passed to teuthology-suite
+  --seed SEED           random seed used in teuthology-suite
+  --no-nested-subset    disable nested subsets used in teuthology-suite
 """
 import docopt
 import teuthology.results

scripts/run.py

@@ -23,6 +23,10 @@ optional arguments:
   --name NAME                    name for this teuthology run
   --suite-path SUITE_PATH        Location of ceph-qa-suite on disk. If not specified,
                                  it will be fetched
+  --interactive-on-error         drop to a python shell on failure, which will
+                                 halt the job; developer can then ssh to targets
+                                 and examine cluster state.
+
 """
 import docopt
 

scripts/schedule.py

@@ -12,6 +12,7 @@ Schedule ceph integration tests
 
 positional arguments:
   <conf_file>                          Config file to read
+                                       "-" indicates read stdin.
 
 optional arguments:
   -h, --help                           Show this help message and exit
@@ -46,6 +47,9 @@ optional arguments:
                                        Only applies to the last job in a suite.
   --subset <subset>                    The subset option passed to teuthology-suite.
                                        Only applies to the last job in a suite.
+  --no-nested-subset                   The no-nested-subset option passed to
+                                       teuthology-suite.
+                                       Only applies to the last job in a suite.
   --dry-run                            Instead of scheduling, just output the
                                        job config.
 

scripts/suite.py

@@ -7,8 +7,8 @@ from teuthology.config import config
 
 doc = """
 usage: teuthology-suite --help
-       teuthology-suite [-v | -vv ] --machine-type <type> --suite <suite> [options] [<config_yaml>...]
-       teuthology-suite [-v | -vv ] --machine-type <type> --rerun <name>  [options] [<config_yaml>...]
+       teuthology-suite [-v | -vv ] --suite <suite> [options] [<config_yaml>...]
+       teuthology-suite [-v | -vv ] --rerun <name>  [options] [<config_yaml>...]
 
 Run a suite of ceph integration tests. A suite is a directory containing
 facets. A facet is a directory containing config snippets. Running a suite
@@ -49,9 +49,9 @@ Standard arguments:
                               use 'none' to bypass kernel task.
                               [default: distro]
   -f <flavor>, --flavor <flavor>
-                              The kernel flavor to run against: ('basic',
-                              'gcov', 'notcmalloc')
-                              [default: basic]
+                              The ceph packages shaman flavor to run with:
+                              ('default', 'crimson', 'notcmalloc', 'jaeger')
+                              [default: default]
   -t <branch>, --teuthology-branch <branch>
                               The teuthology branch to run against.
                               Default value is determined in the next order.
@@ -60,7 +60,7 @@ Standard arguments:
                               the suite repo and contains non-empty string.
                               There is `teuthology_branch` present in one of
                               the user or system `teuthology.yaml` configuration
-                              files respectively, otherwise use `master`.
+                              files respectively, otherwise use `main`.
   -m <type>, --machine-type <type>
                               Machine type [default: {default_machine_type}]
   -d <distro>, --distro <distro>
@@ -86,6 +86,9 @@ Standard arguments:
   --validate-sha1 <bool>
                               Validate that git SHA1s passed to -S exist.
                               [default: true]
+  --kdb <bool>
+                              Enable/disable kdb in kernel
+                              [default: true]
   --sleep-before-teardown <seconds>
                               Number of seconds to sleep before teardown.
                               Use with care, as this applies to all jobs in the
@@ -109,6 +112,13 @@ Scheduler arguments:
                               When tests finish or time out, send an email
                               here. May also be specified in ~/.teuthology.yaml
                               as 'results_email'
+  --expire <datetime>         Do not execute jobs in the run if they have not
+                              completed by this time. Valid formats include
+                              ISO 8601, and relative offsets like '90s', '30m',
+                              '1h', '3d', or '1w'
+  --rocketchat <rocketchat>   Comma separated list of Rocket.Chat channels where
+                              to send a message when tests finished or time out.
+                              To be used with --sleep-before-teardown option.
   -N <num>, --num <num>       Number of times to run/queue the job
                               [default: 1]
   -l <jobs>, --limit <jobs>   Queue at most this many jobs
@@ -118,7 +128,8 @@ Scheduler arguments:
                               contain each facet at least once) and schedule
                               piece <index>.  Scheduling 0/<outof>, 1/<outof>,
                               2/<outof> ... <outof>-1/<outof> will schedule all
-                              jobs in the suite (many more than once).
+                              jobs in the suite (many more than once). If specified,
+                              this value can be found in results.log.
   -p <priority>, --priority <priority>
                               Job priority (lower is sooner)
                               [default: 1000]
@@ -154,20 +165,53 @@ Scheduler arguments:
                               using the same logic as --filter.
                               Of all the flags that were passed when scheduling
                               the original run, the resulting one will only
-                              inherit the suite value. Any others must be
-                              passed as normal while scheduling with this
-                              feature. For random tests involving facet whose
-                              path ends with '$' operator, you might want to
-                              use --seed argument to repeat them.
+                              inherit the --suite value. Any other arguments
+                              must be passed again while scheduling. By default,
+                              'seed' and 'subset' will be taken from results.log,
+                              but can be overide if passed again.
+                              This is important for tests involving random facet
+                              (path ends with '$' operator).
  -R, --rerun-statuses <statuses>
                               A comma-separated list of statuses to be used
                               with --rerun. Supported statuses are: 'dead',
                               'fail', 'pass', 'queued', 'running', 'waiting'
                               [default: fail,dead]
  --seed SEED                  An random number mostly useful when used along
-                              with --rerun argument. This number can be found
-                              in the output of teuthology-suite command. -1
-                              for a random seed [default: -1].
+                              with --rerun argument to rerun the exact
+                              same jobs that can only be picked at random.
+                              This number can be found in the output of
+                              teuthology-suite command or in results.log.
+                              Pass -1 for a random seed [default: -1].
+ --force-priority             Skip the priority check.
+ --job-threshold <threshold>  Do not allow to schedule the run if the number
+                              of jobs exceeds <threshold>. Use 0 to allow
+                              any number [default: {default_job_threshold}].
+ --no-nested-subset           Do not perform nested suite subsets [default: false].
+
++=================+=================================================================+
+| Priority        | Explanation                                                     |
++=================+=================================================================+
+| N < 10          | Use this if the sky is falling and some group of tests          |
+|                 | must be run ASAP.                                               |
++-----------------+-----------------------------------------------------------------+
+| 10 <= N < 50    | Use this if your tests are urgent and blocking other            |
+|                 | important development.                                          |
++-----------------+-----------------------------------------------------------------+
+| 50 <= N < 75    | Use this if you are testing a particular feature/fix            |
+|                 | and running fewer than about 25 jobs. This range is also        |
+|                 | used for urgent release testing.                                |
++-----------------+-----------------------------------------------------------------+
+| 75 <= N < 100   | Tech Leads regularly schedule integration tests with this       |
+|                 | priority to verify pull requests against main.                  |
++-----------------+-----------------------------------------------------------------+
+| 100 <= N < 150  | This priority is used for QE validation of point releases.      |
++-----------------+-----------------------------------------------------------------+
+| 150 <= N < 200  | Use this priority for 100 jobs or fewer that test a particular  |
+|                 | feature or fix. Results are available in about 24 hours.        |
++-----------------+-----------------------------------------------------------------+
+| 200 <= N < 1000 | Use this priority for large test runs. Results are available    |
+|                 | in about a week.                                                |
++-----------------+-----------------------------------------------------------------+
 
 """.format(
     default_machine_type=config.default_machine_type,
@@ -176,7 +220,8 @@ Scheduler arguments:
                             config.get_ceph_git_url()),
     default_suite_repo=defaults('--suite-repo',
                             config.get_ceph_qa_suite_git_url()),
-    default_ceph_branch=defaults('--ceph-branch', 'master'),
+    default_ceph_branch=defaults('--ceph-branch', 'main'),
+    default_job_threshold=config.job_threshold,
 )
 
 

scripts/supervisor.py

@@ -0,0 +1,44 @@
+import argparse
+import sys
+
+import teuthology.dispatcher.supervisor
+
+
+def parse_args(argv):
+    parser = argparse.ArgumentParser(
+        description="Supervise and run a teuthology job; normally only run by the dispatcher",
+    )
+    parser.add_argument(
+        "-v",
+        "--verbose",
+        action="store_true",
+        help="be more verbose",
+    )
+    parser.add_argument(
+        "-a",
+        "--archive-dir",
+        type=str,
+        help="path in which to store the job's logfiles",
+        required=True,
+    )
+    parser.add_argument(
+        "--bin-path",
+        type=str,
+        help="teuthology bin path",
+        required=True,
+    )
+    parser.add_argument(
+        "--job-config",
+        type=str,
+        help="file descriptor of job's config file",
+        required=True,
+    )
+    return parser.parse_args(argv)
+
+
+def main():
+    sys.exit(teuthology.dispatcher.supervisor.main(parse_args(sys.argv[1:])))
+
+
+if __name__ == "__main__":
+    main()

scripts/update_inventory.py

@@ -10,14 +10,16 @@ import logging
 
 doc = """
 usage: teuthology-update-inventory -h
-       teuthology-update-inventory [-v] REMOTE [REMOTE ...]
+       teuthology-update-inventory [-v] [-m type] REMOTE [REMOTE ...]
 
 Update the given nodes' inventory information on the lock server
 
 
-  -h, --help            show this help message and exit
-  -v, --verbose         be more verbose
-  REMOTE                hostnames of machines whose information to update
+  -h, --help                        show this help message and exit
+  -v, --verbose                     be more verbose
+  -m <type>, --machine-type <type>  optionally specify a machine type when 
+                                    submitting nodes for the first time
+  REMOTE                            hostnames of machines whose information to update
 
 """
 
@@ -27,9 +29,13 @@ def main():
     if args['--verbose']:
         teuthology.log.setLevel(logging.DEBUG)
 
+    machine_type = args.get('--machine-type')
     remotes = args.get('REMOTE')
     for rem_name in remotes:
         rem_name = teuthology.misc.canonicalize_hostname(rem_name)
         remote = teuthology.orchestra.remote.Remote(rem_name)
+        remote.connect()
         inventory_info = remote.inventory_info
+        if machine_type:
+          inventory_info['machine_type'] = machine_type
         teuthology.lock.ops.update_inventory(inventory_info)

scripts/wait.py

@@ -0,0 +1,31 @@
+import docopt
+import sys
+
+import logging
+
+import teuthology
+import teuthology.suite
+from teuthology.config import config
+
+doc = """
+usage: teuthology-wait --help
+       teuthology-wait [-v] --run <name>
+
+Wait until run is finished. Returns exit code 0 on success, otherwise 1.
+
+Miscellaneous arguments:
+  -h, --help                  Show this help message and exit
+  -v, --verbose               Be more verbose
+
+Standard arguments:
+  -r, --run <name>          Run name to watch.
+"""
+
+
+def main(argv=sys.argv[1:]):
+    args = docopt.docopt(doc, argv=argv)
+    if args.get('--verbose'):
+        teuthology.log.setLevel(logging.DEBUG)
+    name = args.get('--run')
+    return teuthology.suite.wait(name, config.max_job_time, None)
+

teuthology/__init__.py

@@ -1,5 +1,11 @@
 from __future__ import print_function
-import os
+import os, sys
+try:
+    import importlib.metadata as importlib_metadata
+except ImportError:
+    import importlib_metadata
+
+__version__ = importlib_metadata.version("teuthology")
 
 # Tell gevent not to patch os.waitpid() since it is susceptible to race
 # conditions. See:
@@ -8,17 +14,25 @@ os.environ['GEVENT_NOWAITPID'] = 'true'
 
 # Use manhole to give us a way to debug hung processes
 # https://pypi.python.org/pypi/manhole
-import manhole
-manhole.install(
-    verbose=False,
-    # Listen for SIGUSR1
-    oneshot_on="USR1"
-)
+try:
+    import manhole
+    manhole.install(
+        verbose=False,
+        # Listen for SIGUSR1
+        oneshot_on="USR1"
+    )
+except ImportError:
+    pass
 from gevent import monkey
+patch_threads=True
+for arg in sys.argv:
+    if "teuthology_api" in arg:
+        patch_threads=False
 monkey.patch_all(
     dns=False,
     # Don't patch subprocess to avoid http://tracker.ceph.com/issues/14990
     subprocess=False,
+    thread=patch_threads,
 )
 import sys
 from gevent.hub import Hub
@@ -30,20 +44,6 @@ from teuthology.orchestra import monkey
 monkey.patch_all()
 
 import logging
-import subprocess
-
-__version__ = '1.0.0'
-
-# do our best, but if it fails, continue with above
-
-try:
-    __version__ += '-' + str(subprocess.check_output(
-        'git rev-parse --short HEAD'.split(),
-        cwd=os.path.dirname(os.path.realpath(__file__))
-    ).decode()).strip()
-except Exception as e:
-    # before logging; should be unusual
-    print("Can't get version from git rev-parse %s" % e, file=sys.stderr)
 
 # If we are running inside a virtualenv, ensure we have its 'bin' directory in
 # our PATH. This doesn't happen automatically if scripts are called without
@@ -58,6 +58,9 @@ logging.getLogger('requests.packages.urllib3.connectionpool').setLevel(
 # if requests doesn't bundle it, shut it up anyway
 logging.getLogger('urllib3.connectionpool').setLevel(
     logging.WARN)
+# We also don't need the "Converted retries value" messages
+logging.getLogger('urllib3.util.retry').setLevel(
+    logging.WARN)
 
 logging.basicConfig(
     level=logging.INFO,

teuthology/beanstalk.py

@@ -1,12 +1,13 @@
 import beanstalkc
+import json
 import yaml
 import logging
 import pprint
 import sys
 from collections import OrderedDict
 
-from teuthology.config import config
 from teuthology import report
+from teuthology.config import config
 
 log = logging.getLogger(__name__)
 
@@ -18,7 +19,7 @@ def connect():
         raise RuntimeError(
             'Beanstalk queue information not found in {conf_path}'.format(
                 conf_path=config.teuthology_yaml))
-    return beanstalkc.Connection(host=host, port=port)
+    return beanstalkc.Connection(host=host, port=port, parse_yaml=yaml.safe_load)
 
 
 def watch_tube(connection, tube_name):
@@ -196,7 +197,7 @@ def main(args):
             # it is not needed for pausing tubes
             watch_tube(connection, machine_type)
         if status:
-            print(stats_tube(connection, machine_type))
+            print(json.dumps(stats_tube(connection, machine_type)))
         elif pause_duration:
             pause_tube(connection, machine_type, pause_duration)
         elif delete: