edf-hpc / hpcstats Goto Github PK
View Code? Open in Web Editor NEWHPC clusters data guzzler for usage statistics and metrics
Home Page: http://edf-hpc.github.io/hpcstats/
License: GNU General Public License v2.0
HPC clusters data guzzler for usage statistics and metrics
Home Page: http://edf-hpc.github.io/hpcstats/
License: GNU General Public License v2.0
HPCStats should try mysql requests several times, eventually restarting the connection, to avoid failing this way during long runs:
Traceback (most recent call last):
File "/usr/bin/hpcstats", line 39, in <module>
launcher.run()
File "/usr/lib/python2.6/dist-packages/HPCStats/CLI/HPCStatsLauncher.py", line 154, in run
self.app.run()
File "/usr/lib/python2.6/dist-packages/HPCStats/CLI/HPCStatsImporter.py", line 93, in run
self.import_cluster_data(db, cluster_name)
File "/usr/lib/python2.6/dist-packages/HPCStats/CLI/HPCStatsImporter.py", line 142, in import_cluster_data
self.jobs.load_update_window()
File "/usr/lib/python2.6/dist-packages/HPCStats/Importer/Jobs/JobImporterSlurm.py", line 166, in load_update_window
batch_id = self.load_window(batch_id) + 1
File "/usr/lib/python2.6/dist-packages/HPCStats/Importer/Jobs/JobImporterSlurm.py", line 181, in load_window
return self.get_jobs_after_batchid(batch_id, self.window_size)
File "/usr/lib/python2.6/dist-packages/HPCStats/Importer/Jobs/JobImporterSlurm.py", line 222, in get_jobs_after_batchid
old_schema = self._is_old_schema()
File "/usr/lib/python2.6/dist-packages/HPCStats/Importer/Jobs/JobImporterSlurm.py", line 198, in _is_old_schema
self.cur.execute(req)
File "/usr/lib/pymodules/python2.6/MySQLdb/cursors.py", line 166, in execute
self.errorhandler(self, exc, value)
File "/usr/lib/pymodules/python2.6/MySQLdb/connections.py", line 35, in defaulterrorhandler
raise errorclass, errorvalue
_mysql_exceptions.OperationalError: (2006, 'MySQL server has gone away')
or :
Traceback (most recent call last):
File "/usr/bin/hpcstats", line 39, in <module>
launcher.run()
File "/usr/lib/python2.6/dist-packages/HPCStats/CLI/HPCStatsLauncher.py", line 152, in run
self.app.run()
File "/usr/lib/python2.6/dist-packages/HPCStats/CLI/HPCStatsImporter.py", line 91, in run
self.import_cluster_data(db, cluster_name)
File "/usr/lib/python2.6/dist-packages/HPCStats/CLI/HPCStatsImporter.py", line 140, in import_cluster_data
self.jobs.load_update_window()
File "/usr/lib/python2.6/dist-packages/HPCStats/Importer/Jobs/JobImporterSlurm.py", line 166, in load_update_window
batch_id = self.load_window(batch_id) + 1
File "/usr/lib/python2.6/dist-packages/HPCStats/Importer/Jobs/JobImporterSlurm.py", line 181, in load_window
return self.get_jobs_after_batchid(batch_id, self.window_size)
File "/usr/lib/python2.6/dist-packages/HPCStats/Importer/Jobs/JobImporterSlurm.py", line 222, in get_jobs_after_batchid
old_schema = self._is_old_schema()
File "/usr/lib/python2.6/dist-packages/HPCStats/Importer/Jobs/JobImporterSlurm.py", line 198, in _is_old_schema
self.cur.execute(req)
File "/usr/lib/pymodules/python2.6/MySQLdb/cursors.py", line 166, in execute
self.errorhandler(self, exc, value)
File "/usr/lib/pymodules/python2.6/MySQLdb/connections.py", line 35, in defaulterrorhandler
raise errorclass, errorvalue
_mysql_exceptions.OperationalError: (1213, 'Deadlock found when trying to get lock; try restarting transaction')
Example of backtrace:
Traceback (most recent call last):
File "/usr/share/hpcstats/bin/launch-jobstats", line 92, in <module>
main()
File "/usr/share/hpcstats/bin/launch-jobstats", line 89, in main
ssh_launch(user, frontend, privkey, script)
File "/usr/share/hpcstats/bin/launch-jobstats", line 70, in ssh_launch
ssh.connect(frontend, username=user, key_filename=privkey)
File "/usr/lib/pymodules/python2.6/paramiko/client.py", line 332, in connect
self._auth(username, password, pkey, key_filenames, allow_agent, look_for_keys)
File "/usr/lib/pymodules/python2.6/paramiko/client.py", line 436, in _auth
key = pkey_class.from_private_key_file(key_filename, password)
File "/usr/lib/pymodules/python2.6/paramiko/pkey.py", line 198, in from_private_key_file
key = cls(filename=filename, password=password)
File "/usr/lib/pymodules/python2.6/paramiko/rsakey.py", line 51, in __init__
self._from_private_key_file(filename, password)
File "/usr/lib/pymodules/python2.6/paramiko/rsakey.py", line 163, in _from_private_key_file
data = self._read_private_key_file('RSA', filename, password)
File "/usr/lib/pymodules/python2.6/paramiko/pkey.py", line 279, in _read_private_key_file
f = open(filename, 'r')
IOError: [Errno 2] No such file or directory: '/path/to/file'
New SlurmDBD schema use multivalues TRES fields that replace several other fields, such as cpu_count in event tables. Many SQL requests accross JobImporterSlurm and EventImporterSlurm must be adapted to use these new fields when needed.
Currently, event_type field in DB is 50 chars long. It can be too short for some event type. It should be increased to 200 chars to avoid issues when importing events.
Some code has been developed to support SlurmDBD 15.08. This new code make 9 unit tests fail, and the new code is not covered by the tests. This must be fixed.
The Slurm job importer checks if runs exist even for new jobs, but this is useless since the DB model prevents runs from existing for new jobs. There might be a great optimization here by avoiding the check for new jobs.
Example of backtrace:
File "/usr/share/hpcstats/bin/launch-jobstats", line 92, in <module>
main()
File "/usr/share/hpcstats/bin/launch-jobstats", line 89, in main
ssh_launch(user, frontend, privkey, script)
File "/usr/share/hpcstats/bin/launch-jobstats", line 70, in ssh_launch
ssh.connect(frontend, username=user, key_filename=privkey)
File "/usr/lib/pymodules/python2.6/paramiko/client.py", line 282, in connect
for (family, socktype, proto, canonname, sockaddr) in socket.getaddrinfo(hostname, port, socket.AF_UNSPEC, socket.SOCK_STREAM):
socket.gaierror: [Errno -2] Name or service not known
HPCStats returns an error if "TLS_REQCERT ALLOW" is not set /etc/ldap/ldap.conf
:
ERROR: HPCStatsLauncher: Source Error: unable to connect to LDAP server: {'info': '(unknown error code)', 'desc': "Can't contact LDAP server"}
Currently, there is nothing to prevent multiple concurrent import runs on the same cluster. This systematically result on failed transaction. A mechanism might be added in HPCStats too avoid such concurrent runs.
Currently, the importer app stops as soon as a source error is encountered. If a source error happens on a cluster, on the data from the other cluster (even if they work perfectly) are not imported. This behaviour should be avoided. The data must be imported for all working clusters and cluster encountering source error must simply be ignored.
A declarative, efficient, and flexible JavaScript library for building user interfaces.
๐ Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
An Open Source Machine Learning Framework for Everyone
The Web framework for perfectionists with deadlines.
A PHP framework for web artisans
Bring data to life with SVG, Canvas and HTML. ๐๐๐
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
Some thing interesting about web. New door for the world.
A server is a program made to process requests and deliver data to clients.
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Some thing interesting about visualization, use data art
Some thing interesting about game, make everyone happy.
We are working to build community through open source technology. NB: members must have two-factor auth.
Open source projects and samples from Microsoft.
Google โค๏ธ Open Source for everyone.
Alibaba Open Source for everyone
Data-Driven Documents codes.
China tencent open source team.