cas March 2009

cas@lists.stg.fedorahosted.org

1 participants
7 discussions

cas-admin lib/cas
by Adam Stokes 31 Mar '09

31 Mar '09

cas-admin | 29 ++++++++--------------------- lib/cas/db.py | 14 ++++++++++++-- 2 files changed, 20 insertions(+), 23 deletions(-) New commits: commit 7513467d1fde862f2abe4127d3ee1ff76a0eb6ca Author: Adam Stokes <adam.stokes(a)gmail.com> Date: Tue Mar 31 19:14:58 2009 -0400 - starting to implement db methods into cas-admin diff --git a/cas-admin b/cas-admin index 7fbc538..e25b778 100755 --- a/cas-admin +++ b/cas-admin @@ -37,6 +37,7 @@ KERNELS = config.get("settings","kernels") RPMFILTER = config.get("settings","rpmFilter") DEBUGS = config.get("settings","debugs") DEBUGLEVEL = config.get("settings","debugLevel") +DATABASE = config.get("settings","database") # read maintenance options PURGELIMIT = config.get("maintenance","purgeLimit") @@ -46,20 +47,11 @@ class CasDatabaseHandler(object): def __init__(self, logger): self.casLog = logger self.util = UtilBase() - if os.path.isfile(RPMS): - self.rpmDB = self.util.load(RPMS) - else: - self.rpmDB = {} - - def rpmExist(self, rpm): - """ checks existence of rpm in db - """ - if self.rpmDB.has_key(rpm): - return True - return False + # setup database connection + self.db = CasStorage(DATABASE) + self.db.connect() def run(self): - rpms = [] # Uses emacs regex -- see `man find` cmd = ["find", "-L", KERNELS, "-iregex", RPMFILTER] pipe = Popen(cmd, stdout=PIPE, stderr=PIPE) @@ -67,17 +59,12 @@ class CasDatabaseHandler(object): count = 0 # create list of rpms from `cmd` for line in pipe.stdout: - rpms.append(line.strip()) + self.db.addDebuginfoRPM(line.strip()) self.casLog.status("(found) %-5d kernel(s)" % (count,)) count = count + 1 - # reset count for further informational messaging - totalRpms = len(rpms) - count = 0 - """ Build database out in the form of - RPM - [( DebugKernel, Timestamp), -' ( DebugKernel2, Timestamp2)] - """ - for x in rpms: + # query database for debuginfo rpms + rpms = self.db.getDebuginfoRPM() + for id, rpm in rpms: if not self.rpmExist(x): self.rpmDB[x] = [] count = count + 1 diff --git a/lib/cas/db.py b/lib/cas/db.py index 5e9ef7c..fb0b322 100644 --- a/lib/cas/db.py +++ b/lib/cas/db.py @@ -22,12 +22,22 @@ class CasStorage(object): (date text, submitter text, corepath text, debugpath text, server text) -create table debuginfo (rpm text) +create table debuginfo (debug_id integer primary key autoincrement, rpm text) -create table timestamp (debuginfo integer, debugpath text, timestamp text)''') +create table timestamp (timestamp_id integer primary key autoincrement, +debuginfo integer, debugpath text, timestamp text, debug_id integer)''') self.conn.commit() return + # DEBUGINFO METHODS + def addDebuginfoRPM(self, debuginfo): + self.cursor.execute("INSERT into debuginfo('rpm') values(?)", debuginfo) + return + + def getDebuginfoRPM(self): + self.cursor.execute("SELECT * FROM debuginfo") + return self.cursor + def connect(self): """ execute connection """ try:

1 0

cas cas-admin cas.conf lib/cas
by Adam Stokes 31 Mar '09

31 Mar '09

cas | 5 +---- cas-admin | 2 +- cas.conf | 11 ++--------- lib/cas/db.py | 46 +++++++++++++++++++++++++++++++--------------- 4 files changed, 35 insertions(+), 29 deletions(-) New commits: commit 31c1d943ca61c8ddf0dc0b9d24ddbc7f464e8a0b Author: Adam Stokes <adam.stokes(a)gmail.com> Date: Tue Mar 31 13:13:07 2009 -0400 - more db changes diff --git a/cas b/cas index e255e5e..0dac60c 100755 --- a/cas +++ b/cas @@ -42,10 +42,7 @@ WORKDIRECTORY = config.get("settings","workDirectory") DEBUGLEVEL = config.get("settings","debugLevel") SMTPHOST = config.get("settings", "mailServer") # Read database parameters -CASDB_CONF = {'dbhost': config.get("database","dbhost"), - 'dbuser': config.get("database","dbuser"), - 'dbpass': config.get("database","dbpass"), - 'dbname': config.get("database","dbname")} +DATABASE = config.get("settings", "database") # Check for some advanced configurations # Test to see if we provide a 32bit crash binary diff --git a/cas-admin b/cas-admin index e394c9b..7fbc538 100755 --- a/cas-admin +++ b/cas-admin @@ -75,7 +75,7 @@ class CasDatabaseHandler(object): count = 0 """ Build database out in the form of RPM - [( DebugKernel, Timestamp), - ( DebugKernel2, Timestamp2)] +' ( DebugKernel2, Timestamp2)] """ for x in rpms: if not self.rpmExist(x): diff --git a/cas.conf b/cas.conf index 5b8808a..2b17e34 100644 --- a/cas.conf +++ b/cas.conf @@ -33,15 +33,8 @@ purgeLimit=90 # whenever cas-admin is run autoPurge=False -[database] -# database connection info (mysql only) -# TODO: perhaps add orm support for multiple backends? -# A database needs to be manually created -# $ mysqladmin -uroot -predhat create $dbname -dbname='casdb' -dbhost='localhost' -dbuser='root' -dbpass='redhat' +# database connection info +database='/var/db/cas/cas.db' [advanced] # if running a x86_64 system and wish to analyze 32bit diff --git a/lib/cas/db.py b/lib/cas/db.py index 57313d2..5e9ef7c 100644 --- a/lib/cas/db.py +++ b/lib/cas/db.py @@ -1,38 +1,54 @@ -import MySQLdb +import os +import sys + +if sys.version_info[:2] > (2,4): + try: + import sqlite3 as sqlite + except ImportError: + import sqlite class CasStorageException(Exception): pass class CasStorage(object): - def __init__(self, sql_server, username, password, database): + def __init__(self, database): """ setup database connection and return db cursor for traversing database """ self.db = database - self.sql_server = sql_server - self.username = username - self.password = password + self.conn = None self.cursor = None + def buildTable(self): + self.cursor.execute('''create table cas_jobs +(date text, submitter text, corepath text, debugpath text, +server text) + +create table debuginfo (rpm text) + +create table timestamp (debuginfo integer, debugpath text, timestamp text)''') + self.conn.commit() + return + def connect(self): """ execute connection """ try: - db_connect = MySQLdb.connect(host=self.sql_server, - user=self.username, - passwd=self.password, - db=self.db) - self.cursor = db_connect.cursor() - except OperationalError, e: - raise CasStorageException(e) + if not os.path.isfile(self.db): + # build out sql table + self.buildTable() + self.conn = sqlite.connect(self.db) + self.cursor = self.conn.cursor() + except: + raise CasStorageException('Cannot connect to database') return def getAllJobs(self): """ all jobs """ self.cursor.execute("SELECT * FROM cas_jobs") - return self.cursor.fetchall() + return self.cursor def getJobById(self, id): """ single job """ - self.cursor.execute("SELECT * FROM cas_jobs where id=%d" % (id,)) - return self.cursor.fetchone() + self.cursor.execute("SELECT * FROM cas_jobs where id=?", id) + return self.cursor def getJobRange(self, days): """ provides jobs based on creation date from

1 0

cas cas-admin cas.conf cas.sql lib/cas
by Adam Stokes 19 Mar '09

19 Mar '09

cas | 7 +++---- cas-admin | 23 +++++++++++++++++++++-- cas.conf | 17 ++++++++++++----- cas.sql | 27 ++++++++++----------------- lib/cas/db.py | 15 ++++++++++----- 5 files changed, 56 insertions(+), 33 deletions(-) New commits: commit f6c4ec270ffedd4d21e9802eee2b26caa4acd433 Author: Adam Stokes <adam.stokes(a)gmail.com> Date: Thu Mar 19 21:07:25 2009 -0400 - conf update - sql update to better reflect the relationships involved - focus on cas-admin to incorporate maintenance diff --git a/cas b/cas index 7303a54..e255e5e 100755 --- a/cas +++ b/cas @@ -40,7 +40,6 @@ config = ConfigParser.ConfigParser() config.read("/etc/cas.conf") WORKDIRECTORY = config.get("settings","workDirectory") DEBUGLEVEL = config.get("settings","debugLevel") -SERVERS = config.get("settings", "servers") SMTPHOST = config.get("settings", "mailServer") # Read database parameters CASDB_CONF = {'dbhost': config.get("database","dbhost"), @@ -143,8 +142,7 @@ class CasApplication(object): parser.add_option("-f","--file", dest="filename", help="Filename") parser.add_option("-e","--email", dest="email", - help="Define email for results (must be valid!)", - action="store") + help="Define email for results") parser.add_option("-m","--modules", dest="kernel_modules", help="Extract associated kernel modules", action="store_true") @@ -241,7 +239,8 @@ class CasApplication(object): # TODO: Randomize server selection # TODO: Verify remote server is reachable casProcessMachine = serverList[debugKernelArch][0] - self.casLog.info("Machine %s found, processing crash output" % (casProcessMachine,)) + self.casLog.info("Machine %s found, processing " \ + "crash output" % (casProcessMachine,)) cmd = os.path.join(self.storagePath,"crash") client = fc.Overlord(casProcessMachine) clientDict = client.command.run(cmd) diff --git a/cas-admin b/cas-admin index 8bd9819..e394c9b 100755 --- a/cas-admin +++ b/cas-admin @@ -23,6 +23,7 @@ import urlparse from cas.core import CoreBase from cas.util import UtilBase, Logging from cas.rpmutils import RPMBase +from cas.db import CasStorage, CasStorageException from subprocess import Popen, PIPE from cas.cas_shutil import rmtree @@ -33,11 +34,13 @@ if sys.version_info[:2] < (2,4): config = ConfigParser.ConfigParser() config.read("/etc/cas.conf") KERNELS = config.get("settings","kernels") -RPMS = config.get("settings","rpms") RPMFILTER = config.get("settings","rpmFilter") DEBUGS = config.get("settings","debugs") DEBUGLEVEL = config.get("settings","debugLevel") -SERVERS = config.get("settings","servers") + +# read maintenance options +PURGELIMIT = config.get("maintenance","purgeLimit") +AUTOPURGE = config.get("maintenance","autoPurge") class CasDatabaseHandler(object): def __init__(self, logger): @@ -151,9 +154,15 @@ class CasAdminApplication(object): help="Build CAS DB", action="store_true", default=False) parser.add_option("-s","--server", dest="server_init", help="Build SERVER DB", action="store_true", default=False) + parser.add_option("-p","--purge", dest="purgeData", + help="Purge files", action="store_true", default=False) + parser.add_option("-d","--days", dest="purgeDataDays", + help="Set how many days back to purge data") (self.opts, args) = parser.parse_args() self.buildDB = self.opts.buildDB self.server_init = self.opts.server_init + self.purgeData = self.opts.purgeData + self.purgeDataDays = self.opts.purgeDataDays def run(self): """ Make sure necessary directories and configuration is setup @@ -166,6 +175,16 @@ class CasAdminApplication(object): if not os.path.isdir(DEBUGS): os.makedirs(DEBUGS) + if self.purgeData: + ans = raw_input("You are about to purge data, is this what you " \ + "really want to do? [Y/y/N/n]: ") + if ans=='Y' or ans=='y': + if not self.purgeDataDays: + self.purgeDataDays = PURGELIMIT + self.casLog.info("Beginning Purge going back %s days" % (self.purgeDataDays,)) + else: + sys.exit(0) + if self.buildDB: self.casLog.info("Starting CAS DB instance.") dbHandler = CasDatabaseHandler(self.casLog).run() diff --git a/cas.conf b/cas.conf index 82d2ce8..5b8808a 100644 --- a/cas.conf +++ b/cas.conf @@ -4,9 +4,6 @@ # indefinately recursive kernels=/mnt/kernels -# Database which houses all kernel found in $kernels -rpms=/var/db/cas/rpms.db - # Compose emacs regular expression for determing what # kernel debug rpms you wish to search for rpmFilter=.*kernel-debuginfo-[0-9].*\.rpm @@ -16,21 +13,31 @@ rpmFilter=.*kernel-debuginfo-[0-9].*\.rpm # cpio is run against an rpm and extract the debug kernel # for a small amount of time. This could be something like # /tmp if space permitted -debugs=/cores/debugs +debugs=/tmp # debug level (DEBUG, INFO) debugLevel=DEBUG # define work directory -workDirectory=/cores/processed +workDirectory=/cores # Mail server, e.g. mail.example.com # Provides job results via email mailServer=mail.example.com +[maintenance] +# Time in days to keep jobs and its associated files +purgeLimit=90 + +# Enable if want to automatically check and purge data +# whenever cas-admin is run +autoPurge=False + [database] # database connection info (mysql only) # TODO: perhaps add orm support for multiple backends? +# A database needs to be manually created +# $ mysqladmin -uroot -predhat create $dbname dbname='casdb' dbhost='localhost' dbuser='root' diff --git a/cas.sql b/cas.sql index 5246295..0448425 100644 --- a/cas.sql +++ b/cas.sql @@ -2,52 +2,45 @@ -- cas jobs -- CREATE TABLE jobs ( - id MEDIUMINT NOT NULL AUTO_INCREMENT, + id INTEGER PRIMARY KEY AUTO_INCREMENT, submitter VARCHAR(255), creation TIMESTAMP DEFAULT CURRENT_TIMESTAMP, jobpath TEXT, corefile TEXT, - debug TEXT, - PRIMARY KEY (id) + debug TEXT ) -- -- cas servers -- CREATE TABLE architecture ( - id MEDIUMINT NOT NULL AUTO_INCREMENT, - arch VARCHAR(255), - PRIMARY KEY (id) + id INTEGER PRIMARY KEY AUTO_INCREMENT, + arch VARCHAR(255) ) -- -- one to many relationship with architectures -- CREATE TABLE servers ( - id MEDIUMINT NOT NULL AUTO_INCREMENT, + id INTEGER PRIMARY KEY AUTO_INCREMENT, hostname VARCHAR(255), - arch_id INTEGER, - PRIMARY KEY (id), - FOREIGN KEY (arch_id) + arch_id INTEGER ) -- -- debug rpm filename -- CREATE TABLE debuginforpm ( - id MEDIUMINT NOT NULL AUTO_INCREMENT, - filename VARCHAR(255), - PRIMARY KEY (id) + id INTEGER PRIMARY KEY AUTO_INCREMENT, + filename VARCHAR(255) ) -- -- debug kernels -- CREATE TABLE debugkernels ( - id MEDIUMINT NOT NULL AUTO_INCREMENT, + id INTEGER PRIMARY KEY AUTO_INCREMENT, debugpath TEXT, debugtimestamp TEXT, - debuginforpm_id INTEGER, - PRIMARY KEY (id), - FOREIGN KEY (debuginforpm_id) + debuginforpm_id INTEGER ) diff --git a/lib/cas/db.py b/lib/cas/db.py index 175a3bb..57313d2 100644 --- a/lib/cas/db.py +++ b/lib/cas/db.py @@ -1,5 +1,7 @@ import MySQLdb +class CasStorageException(Exception): pass + class CasStorage(object): def __init__(self, sql_server, username, password, database): """ setup database connection and return db cursor for @@ -12,11 +14,14 @@ class CasStorage(object): def connect(self): """ execute connection """ - db_connect = MySQLdb.connect(host=self.sql_server, - user=self.username, - passwd=self.password, - db=self.db) - self.cursor = db_connect.cursor() + try: + db_connect = MySQLdb.connect(host=self.sql_server, + user=self.username, + passwd=self.password, + db=self.db) + self.cursor = db_connect.cursor() + except OperationalError, e: + raise CasStorageException(e) return def getAllJobs(self):

1 0

2 commits - cas cas-admin cas.conf cas.sql lib/cas
by Adam Stokes 18 Mar '09

18 Mar '09

cas | 16 ++++++++++------ cas-admin | 8 ++++---- cas.conf | 22 ++++++++++------------ cas.sql | 43 ++++++++++++++++++++++++++++++++++++++++++- lib/cas/db.py | 14 +++++++++++++- 5 files changed, 79 insertions(+), 24 deletions(-) New commits: commit 5b14fda2d3e3017f032fe9a1b866c9025ac516c3 Author: Adam Stokes <adam.stokes(a)gmail.com> Date: Wed Mar 18 23:10:10 2009 -0400 - extend sql - start changes to reflect db schema diff --git a/cas b/cas index 319ca7b..7303a54 100755 --- a/cas +++ b/cas @@ -39,10 +39,14 @@ if sys.version_info[:2] < (2,4): config = ConfigParser.ConfigParser() config.read("/etc/cas.conf") WORKDIRECTORY = config.get("settings","workDirectory") -RPMS = config.get("settings","rpms") DEBUGLEVEL = config.get("settings","debugLevel") SERVERS = config.get("settings", "servers") SMTPHOST = config.get("settings", "mailServer") +# Read database parameters +CASDB_CONF = {'dbhost': config.get("database","dbhost"), + 'dbuser': config.get("database","dbuser"), + 'dbpass': config.get("database","dbpass"), + 'dbname': config.get("database","dbname")} # Check for some advanced configurations # Test to see if we provide a 32bit crash binary @@ -51,7 +55,7 @@ SMTPHOST = config.get("settings", "mailServer") CRASH_32=None if config.has_option("advanced", "crash_32"): CRASH_32=config.get("advanced", "crash_32") - + class CoreHandler(object): def __init__(self, filename, dst, logger): self.filename = filename @@ -161,7 +165,7 @@ class CasApplication(object): dateFormatted = datenow.strftime("%m.%d.%y.%I.%M.%S") self.storagePath = os.path.join(WORKDIRECTORY, self.identifier) self.storagePath = os.path.join(self.storagePath, dateFormatted) - + # build logger object to deal with logging per job and keep things # clean and easy to debug self.casLog = Logging(self.storagePath, self.identifier) @@ -185,7 +189,7 @@ class CasApplication(object): debugKernel = os.path.abspath(debugKernel) # setup crash file to finalize the processing of the core file self.util.buildCrashFile(self.storagePath, corefile, debugKernel) - # Pull the architecture from the elf file to match up with a + # Pull the architecture from the elf file to match up with a # server providing this architecture debugKernelArch = self.util.debugKernelArch(debugKernel) # Read current machine arch to see if we can bypass func and proceed @@ -204,7 +208,7 @@ class CasApplication(object): cmdPipe = Popen([cmd], stdout=PIPE, stderr=PIPE) cmdData = cmdPipe.communicate() # pull status code to verify crash even ran to completeness - sts, out, err = (cmdPipe.returncode, cmdData[0].strip(), + sts, out, err = (cmdPipe.returncode, cmdData[0].strip(), cmdData[1].strip()) if sts: self.casLog.debug("crash problem: err: %s, out: %s" % (err, out)) @@ -219,7 +223,7 @@ class CasApplication(object): cmdPipe = Popen([cmd], stdout=PIPE, stderr=PIPE) cmdData = cmdPipe.communicate() # pull status code to verify crash even ran to completeness - sts, out, err = (cmdPipe.returncode, cmdData[0].strip(), + sts, out, err = (cmdPipe.returncode, cmdData[0].strip(), cmdData[1].strip()) if sts: self.casLog.debug("crash problem: err: %s, out: %s" % (err, out)) diff --git a/cas-admin b/cas-admin index 100140d..8bd9819 100755 --- a/cas-admin +++ b/cas-admin @@ -81,7 +81,7 @@ class CasDatabaseHandler(object): # temporary storage path in form of DEBUGS/COUNT dst = os.path.join(DEBUGS, str(count)) rpmTool = RPMBase() - self.casLog.status("(extracting) [%d/%d] %-50s" % (count, totalRpms, + self.casLog.status("(extracting) [%d/%d] %-50s" % (count, totalRpms, os.path.basename(x))) results = rpmTool.extract(x, dst) # Sort through extracted debug for each type @@ -97,12 +97,12 @@ class CasDatabaseHandler(object): # Cleanup extracted debugs rmtree(dst) return - + class CasServerHandler(object): def __init__(self, logger): self.casLog = logger self.util = UtilBase() - + def run(self): try: serverList = {} @@ -146,7 +146,7 @@ class CasAdminApplication(object): self.casLog = Logging("/var/log","cas-admin") def parse_options(self, args): - parser = optparse.OptionParser(usage="casprint [opts] args") + parser = optparse.OptionParser(usage="cas-admin [opts] args") parser.add_option("-b","--build", dest="buildDB", help="Build CAS DB", action="store_true", default=False) parser.add_option("-s","--server", dest="server_init", diff --git a/cas.conf b/cas.conf index 6ce4579..82d2ce8 100644 --- a/cas.conf +++ b/cas.conf @@ -24,24 +24,22 @@ debugLevel=DEBUG # define work directory workDirectory=/cores/processed -# database to house cas servers -# Server list is stored in a dict in the form of -# {'x86_64' : [servera, serverb, serverc], -# 'x86' : [serverm, servern, servero]} -servers=/var/db/cas/servers.db - -# define processing jobs db -# TODO: determine format type, (e.g raw, json, xml) -jobs=/var/db/cas/jobs.db - # Mail server, e.g. mail.example.com # Provides job results via email mailServer=mail.example.com +[database] +# database connection info (mysql only) +# TODO: perhaps add orm support for multiple backends? +dbname='casdb' +dbhost='localhost' +dbuser='root' +dbpass='redhat' + [advanced] # if running a x86_64 system and wish to analyze 32bit # cores, define the location of 32bit crash binary -# EXAMPLE USAGE: +# EXAMPLE USAGE: # rpm --relocate /usr/bin=/usr/local/i386 -ivh crash*i386.rpm - + # crash_32=/usr/local/i386/crash diff --git a/lib/cas/db.py b/lib/cas/db.py index 9c0cac6..175a3bb 100644 --- a/lib/cas/db.py +++ b/lib/cas/db.py @@ -19,6 +19,18 @@ class CasStorage(object): self.cursor = db_connect.cursor() return - def getJobs(self): + def getAllJobs(self): + """ all jobs """ self.cursor.execute("SELECT * FROM cas_jobs") return self.cursor.fetchall() + + def getJobById(self, id): + """ single job """ + self.cursor.execute("SELECT * FROM cas_jobs where id=%d" % (id,)) + return self.cursor.fetchone() + + def getJobRange(self, days): + """ provides jobs based on creation date from + $days back + """ + pass commit 223325498434a16147675083ffc6bdd5b28e5f47 Author: Adam Stokes <adam.stokes(a)gmail.com> Date: Wed Mar 18 23:10:02 2009 -0400 - add additional tables for db migration diff --git a/cas.sql b/cas.sql index b7e8eb5..5246295 100644 --- a/cas.sql +++ b/cas.sql @@ -1,7 +1,7 @@ -- -- cas jobs -- -CREATE TABLE cas_jobs ( +CREATE TABLE jobs ( id MEDIUMINT NOT NULL AUTO_INCREMENT, submitter VARCHAR(255), creation TIMESTAMP DEFAULT CURRENT_TIMESTAMP, @@ -10,3 +10,44 @@ CREATE TABLE cas_jobs ( debug TEXT, PRIMARY KEY (id) ) + +-- +-- cas servers +-- +CREATE TABLE architecture ( + id MEDIUMINT NOT NULL AUTO_INCREMENT, + arch VARCHAR(255), + PRIMARY KEY (id) +) + +-- +-- one to many relationship with architectures +-- +CREATE TABLE servers ( + id MEDIUMINT NOT NULL AUTO_INCREMENT, + hostname VARCHAR(255), + arch_id INTEGER, + PRIMARY KEY (id), + FOREIGN KEY (arch_id) +) + +-- +-- debug rpm filename +-- +CREATE TABLE debuginforpm ( + id MEDIUMINT NOT NULL AUTO_INCREMENT, + filename VARCHAR(255), + PRIMARY KEY (id) +) + +-- +-- debug kernels +-- +CREATE TABLE debugkernels ( + id MEDIUMINT NOT NULL AUTO_INCREMENT, + debugpath TEXT, + debugtimestamp TEXT, + debuginforpm_id INTEGER, + PRIMARY KEY (id), + FOREIGN KEY (debuginforpm_id) +)

1 0

cas.sql
by Adam Stokes 18 Mar '09

18 Mar '09

cas.sql | 1 - 1 file changed, 1 deletion(-) New commits: commit 6af0beb13bf99eb7127dc8751a545e69964258ec Author: Adam Stokes <adam.stokes(a)gmail.com> Date: Wed Mar 18 13:18:07 2009 -0400 - no double timestamp diff --git a/cas.sql b/cas.sql index ef3aa33..b7e8eb5 100644 --- a/cas.sql +++ b/cas.sql @@ -5,7 +5,6 @@ CREATE TABLE cas_jobs ( id MEDIUMINT NOT NULL AUTO_INCREMENT, submitter VARCHAR(255), creation TIMESTAMP DEFAULT CURRENT_TIMESTAMP, - modified TIMESTAMP DEFAULT CURRENT_TIMESTAMP, jobpath TEXT, corefile TEXT, debug TEXT,

1 0

cas.sql
by Adam Stokes 18 Mar '09

18 Mar '09

cas.sql | 16 +++++++++++----- 1 file changed, 11 insertions(+), 5 deletions(-) New commits: commit d6d04e7de0e22d11d70cb39f44e0c300ce5d9716 Author: Adam Stokes <adam.stokes(a)gmail.com> Date: Wed Mar 18 13:16:26 2009 -0400 - update sql diff --git a/cas.sql b/cas.sql index 7017d37..ef3aa33 100644 --- a/cas.sql +++ b/cas.sql @@ -1,7 +1,13 @@ +-- +-- cas jobs +-- CREATE TABLE cas_jobs ( - 'ID' PRIMARY AUTO_INCREMENT INTEGER, - 'PATH' TEXT, - 'COREFILEPATH' TEXT, - 'DEBUGPATH' TEXT, - 'CREATION_TIME' DATETIME CURRENT_TIMESTAMP + id MEDIUMINT NOT NULL AUTO_INCREMENT, + submitter VARCHAR(255), + creation TIMESTAMP DEFAULT CURRENT_TIMESTAMP, + modified TIMESTAMP DEFAULT CURRENT_TIMESTAMP, + jobpath TEXT, + corefile TEXT, + debug TEXT, + PRIMARY KEY (id) )

1 0

cas.spec cas.sql doc/index.rst lib/cas
by Adam Stokes 18 Mar '09

18 Mar '09

cas.spec | 10 - cas.sql | 7 + doc/index.rst | 356 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ lib/cas/db.py | 34 +++-- 4 files changed, 390 insertions(+), 17 deletions(-) New commits: commit 1c9fbbbcd455db47220ab295c7b2ce9092c01b68 Author: Adam Stokes <adam.stokes(a)gmail.com> Date: Wed Mar 18 11:52:46 2009 -0400 - start work on using mysql backend diff --git a/cas.spec b/cas.spec index e463724..5b53a5a 100644 --- a/cas.spec +++ b/cas.spec @@ -3,7 +3,7 @@ Name: cas Summary: Tool to analyze and configure core file environment Version: 0.13 -Release: 120%{?dist} +Release: 121%{?dist} Source0: https://fedorahosted.org/releases/c/a/cas/%{name}-%{version}.tar.gz License: GPLv3+ Group: Development/Libraries @@ -11,12 +11,12 @@ BuildRoot: %{_tmppath}/%{name}-%{version}-%{release}-buildroot BuildArch: noarch Url: http://fedorahosted.org/cas BuildRequires: python-devel -Requires: crash +Requires: crash MySQL-python %description CAS provides a user the ability to configure an environment for core analysis -quickly. All the hassles of matching kernel versions and machine architecture -types to core dumps are automatically detected and processed. +quickly. All the hassles of matching kernel versions and machine architecture +types to core dumps are automatically detected and processed. %prep %setup -q @@ -76,7 +76,7 @@ rm -rf ${RPM_BUILD_ROOT} - splitting off grabcore to be a download/extract only service - core of the work to be done specifically by their intended modules - + * Fri Dec 7 2007 Adam Stokes <astokes at redhat dot com> - 0.9 - release bump - decompression module added diff --git a/cas.sql b/cas.sql new file mode 100644 index 0000000..7017d37 --- /dev/null +++ b/cas.sql @@ -0,0 +1,7 @@ +CREATE TABLE cas_jobs ( + 'ID' PRIMARY AUTO_INCREMENT INTEGER, + 'PATH' TEXT, + 'COREFILEPATH' TEXT, + 'DEBUGPATH' TEXT, + 'CREATION_TIME' DATETIME CURRENT_TIMESTAMP +) diff --git a/doc/index.rst b/doc/index.rst new file mode 100644 index 0000000..7ecb87c --- /dev/null +++ b/doc/index.rst @@ -0,0 +1,356 @@ +Core Analysis System +==================== + +Introduction +------------ + +.. image:: cas_logo.png + +Description +^^^^^^^^^^^ + +CAS provides a user the ability to configure an environment for core analysis +quickly. All the hassles of matching kernel versions and machine architecture +types to core dumps are automatically detected and processed. + +Prerequisites +^^^^^^^^^^^^^ + +CAS needs at least **Python 2.4** to run. For systems that are not running +Fedora 9 or later (this would include RHEL 4/5) the EPEL repo needs to be +installed. Visit `EPEL <https://fedoraproject.org/wiki/EPEL>`_ to enable +this repository. + +Optionally install `Func <http://fedorahosted.org/func>`_ +for automatic processing on other architecture specific systems. This +can easily be installed via `yum`:: + + $ yum install func + +The amount of storage needed can be determined base on the following +information: + +- The number of kernel-debuginfo packages needed +- How many core dumps will be processed. + +Typically it is recommended to have at least 1TB for cores and another 500GB for +the debuginfo packages. + +Since analyzing cores requires the same architecture specific systems the core +was generated on there will need to be systems available of those same types +in order for analyzation to work properly. + +Finally, root access to the CAS server. + +Configuration +^^^^^^^^^^^^^ + +CAS comes with one main configuration file which is located at ``/etc/cas.conf``. +The overall contents of this file is shown below, further down we will break up +each section and describe its meaning:: + + [settings] + kernels=/mnt/kernels + rpms=/var/db/cas/rpms.db + rpmFilter=.*kernel-debuginfo-[0-9].*\.rpm + debugs=/cores/debugs + debugLevel=DEBUG + workDirectory=/cores/processed + servers=/var/db/cas/servers.db + jobs=/var/db/cas/jobs.db + mailServer=mail.example.com + [advanced] + # crash_32=/usr/local/i386/crash + +``kernels``: (**Required**) Describes the location of where kernel-debuginfo packages are to be +stored. This can range anywhere from an nfs mount, samba share, local disk or +any other type of media the cas server can access. + +``rpms``: (**Required**) Database which houses the processing of kernel-debuginfo and stores the +necessary information for CAS to properly match core dumps. + +``rpmFilters``: (**Required**) This is a emacs based regular expression which is essentially +passed to a find command to locate the various kernel-debuginfo packages defined +in ``kernels`` directive. + +``debugs``: (**Required**) A temporary directory in which to store the extracted vmlinux files +from the kernel-debuginfo packages for processing. Another solution would be to +alter this to point an existing directory like ``/tmp``, for instance. + +``debugLevel``: As the name suggest it will set the debug level for CAS output. +Currently the only accepted values are ``DEBUG|INFO``. + +``workDirectory``: (**Required**) Defines where all processed cores will be placed. This mount +point will need to have the most storage assigned to it. Depending on how many +cores are processed in a given timeframe this area will fill up quickly. + +``servers``: If func is installed and configured all associated servers will +be stored here. + +``jobs``: Contains information on current jobs being processed and eventually +what jobs are in the queue. + +``mailServer``: If wanting output of CAS processing email to a certain address +this directive needs to be set. ``Note`` that the mail server should not +require smtp authentication. + +``crash_32``: Primarily used on x86_64 systems to process x86 cores. If x86 +version of crash is installed this directive can be set to the crash binary +and CAS will automatically process x86 cores on a x86_64 machine. ``Note`` this +is only available if the CAS server is a x86_64 machine. + +Setup & Execution +----------------- + +Preparing CAS Server +^^^^^^^^^^^^^^^^^^^^ + +To install the CAS package simply type:: + + $ yum install cas + +Once installed edit ``/etc/cas.conf`` as root using any preferred text editor. +As described above the required directives need to be altered to suit the +environment in question. + +In this example, ``/mnt/kernels`` is an nfs mount which houses the kernel-debuginfo +packages. ``/cores`` is where all processed cores are stored and ``/tmp`` is the +temporary storage for collecting the necessary data from the kernel-debuginfos. +A mail server is setup within the environment to email CAS results and this +optional directive is shown to reflect that. Finally, the CAS server is an x86_64 +machine and the environment will be processing x86 cores, therefore, the directive +for this is uncommented and path to the x86 crash binary is given. ``Note`` there +is information provided within the configuration file for installing the x86 crash +to a different location. + +Altering the configuration to reflect the above assumptions would show the +following:: + + [settings] + kernels=/mnt/kernels + rpms=/var/db/cas/rpms.db + rpmFilter=.*kernel-debuginfo-[0-9].*\.rpm + debugs=/tmp + debugLevel=DEBUG + workDirectory=/cores + servers=/var/db/cas/servers.db + jobs=/var/db/cas/jobs.db + mailServer=mail.cas-server.com + [advanced] + crash_32=/usr/local/i386/crash + +Now that the configuration file is altered and ``/mnt/kernels`` should be populated +with kernel-debuginfo rpm's the next section will describe running CAS. + +Running CAS +^^^^^^^^^^^ + +First, one or two administrative tasks need to be run. The required task is to build +a database for all the data gathered from the kernel-debuginfo packages.:: + + $ cas-admin -b + +If ``Func`` is installed and several systems are deployed for CAS to use run:: + + $ cas-admin -s + +At this point CAS is configured and looking at the output of CAS help there are +a few options to pass:: + + Usage: cas [opts] args + + Options: + -h, --help show this help message and exit + -i IDENTIFIER, --identifier=IDENTIFIER + Unique ID for core + -f FILENAME, --file=FILENAME + Filename + -e EMAIL, --email=EMAIL + Define email for results (must be valid!) + -m, --modules Extract associated kernel modules + +CAS prepares its directory hierarchy based on the ``identifier`` this option is +therefore required. ``filename`` is also required as it tells CAS exactly which +core to process and associate with ``identifier``. If wanting email results from +CAS simply pass it the email parameter. + +An example, of a user wanting to process a corefile named ``vmcore.12345``:: + + $ cas -i 12345 -f vmcore.12345 -e user(a)cas-server.com + +In the above example an assumption is made that ``12345`` is associated to some +form of ticketing system so to keep things organized an identifier was set of +that number. + +The directory hierarchy for the current job should look like ``/cores/12345``. +In addition to the processing of core files there is also a ``process log`` contained +within this directory for each job processed. If multiple jobs for the same identifier +are issued they are placed within a sub directory marked by the current timestamp +and the relevant data associated with it. + +The last option worth mentioning is for core analyst who are needing to work +within the core that requires one of the kernel modules loaded during the crash. +This can be extracted by passing the ``modules`` parameter in the CAS execution +statement. ``Note`` the ``modules`` parameter is not heavily used but can be +useful when analyzing filesystem issues and the like. + +From this point on CAS will download, process, and email the results of its +initial analysis to the specified email address. From there further instructions +are provided in either the email or the ``process log`` on how to access and analyze +the core. + +Analyzing +--------- + +Continuing with the previous example the results of CAS processing should be emailed +and look something similar to:: + + Subject: CAS results for 12345 + Date: Wed, 11 Feb 2009 08:44:37 -0500 + + Location: /cores/processed/12345/02.11.09.08.44.19 + Server: x86_64.cas-server.com + Output data: + PID: 0 TASK: ffffffff803e9b80 CPU: 0 COMMAND: "swapper" + #0 [ffffffff8047a0a0] smp_call_function_interrupt at ffffffff8011d191 + #1 [ffffffff8047a0b0] call_function_interrupt at ffffffff80110bf5 + --- <IRQ stack> --- + #2 [ffffffff80529f08] call_function_interrupt at ffffffff80110bf5 + [exception RIP: default_idle+32] + RIP: ffffffff8010e7a9 RSP: ffffffff80529fb8 RFLAGS: 00000246 + RAX: 0000000000000000 RBX: 0000000000000000 RCX: 0000000000000018 + RDX: ffffffff8010e789 RSI: ffffffff803e9b80 RDI: 0000010008001780 + RBP: 0000000000000000 R8: ffffffff80528000 R9: 0000000000000080 + R10: 0000000000000100 R11: 0000000000000004 R12: 0000000000000000 + R13: 0000000000000000 R14: 0000000000000000 R15: 0000000000000000 + ORIG_RAX: fffffffffffffffa CS: 0010 SS: 0018 + #3 [ffffffff80529fb8] cpu_idle at ffffffff8010e81c + + PID: 0 TASK: 100f57cb030 CPU: 1 COMMAND: "swapper" + #0 [1000107bfa0] smp_call_function_interrupt at ffffffff8011d191 + #1 [1000107bfb0] call_function_interrupt at ffffffff80110bf5 + --- <IRQ stack> --- + #2 [10001073e98] call_function_interrupt at ffffffff80110bf5 + [exception RIP: default_idle+32] + RIP: ffffffff8010e7a9 RSP: 0000010001073f48 RFLAGS: 00000246 + RAX: 0000000000000000 RBX: 0000000000000e86 RCX: 0000000000000018 + RDX: ffffffff8010e789 RSI: 00000100f57cb030 RDI: 00000102000a4780 + RBP: 0000000000000001 R8: 0000010001072000 R9: 0000000000000040 + R10: 0000000000000000 R11: 0000000000000008 R12: 0000000000000000 + R13: 0000000000000000 R14: 0000000000000000 R15: 0000000000000000 + ORIG_RAX: fffffffffffffffa CS: 0010 SS: 0018 + #3 [10001073f48] cpu_idle at ffffffff8010e81c + + PID: 6122 TASK: 101f3658030 CPU: 2 COMMAND: "gfs_quotad" + #0 [101f21efb20] start_disk_dump at ffffffffa03183ff + #1 [101f21efb50] try_crashdump at ffffffff8014cc1d + #2 [101f21efb60] die at ffffffff80111c90 + #3 [101f21efb80] do_invalid_op at ffffffff80112058 + #4 [101f21efc40] error_exit at ffffffff80110e1d + [exception RIP: do_dlm_lock+366] + + ... snip ... + +From this email a ``location`` is provided ``/cores/processed/12345/02.11.09.08.44.19`` +and the server in which further analyzation can be continued ``x86_64.cas-server.com``. + +Normally from a support perspective this email should contain enough information +for a kernel engineer to begin debugging the problem. Assuming more is needed +the information provided previously will prove beneficial for anyone wishing +to access this data. + +Logging into the stated server and changing into the directory defined several +files are presented:: + + $ pwd + /cores/processed/12345/02.11.09.08.44.19 + $ ls + 12345.log crash crash.in crash.out usr vmcore.12345 + +``12345.log``: contains any informational messages presented during the processing +of the core. Everything from informational to debug statements are provided here. + +``crash``: a script autogenerated to provide an automated way of gathering intial +data from the coredump. ``Note`` if wanting to use this crash wrapper in a more +manual approach some alterations to the script need to occur. + +crash wrapper in its original form:: + + #!/bin/sh + /usr/bin/crash \ + /cores/processed/12345/02.11.09.08.44.19/vmcore.12345 \ + /cores/processed/12345/02.11.09.08.44.19/usr/*/*/*/*/2.6.9*largesmp/vmlinux \ + -s < /cores/processed/12345/02.11.09.08.44.19/crash.in + +In order to remove automation the redirect from crash.in needs to be removed:: + + #!/bin/sh + /usr/bin/crash /cores/processed/12345/02.11.09.08.44.19/vmcore.12345 \ + /cores/processed/12345/02.11.09.08.44.19/usr/*/*/*/*/2.6.9*largesmp/vmlinux + +**Alternative to using the crash wrapper** + +It is possible to specify the vmlinux and corefile with crash on the command line:: + + $ crash /cores/processed/12345/02.11.09.08.44.19/usr/*/*/*/*/2.6.9*largesmp/vmlinux \ + /cores/processed/12345/02.11.09.08.44.19/vmcore.12345 + +``crash.in``: a list of commands to be read into crash during the automated +analysis:: + + bt -a >>/cores/processed/12345/02.11.09.08.44.19/crash.out + sys >>/cores/processed/12345/02.11.09.08.44.19/crash.out + log >>/cores/processed/12345/02.11.09.08.44.19/crash.out + mod >>/cores/processed/12345/02.11.09.08.44.19/crash.out + exit + +``crash.out``: output of initial crash analysis and the same data which +is sent in an email if defined. + +``usr``: directory structure from the extraction of the vmlinux file +from the associated kernel-debuginfo rpm for use within crash:: + + /cores/processed/12345/02.11.09.08.44.19/ + usr/lib/debug/lib/modules/2.6.9-78.18.ELlargesmp/vmlinux + +``vmcore.12345``: corefile from which was either defined or extracted from +a compressed archive during CAS initialization. + +Troubleshooting +--------------- + +Some of the major problems that arise when using CAS usually boils down to some +improper usage of the compression and archiving tools. + +When compressing a core which may need to be sent over the network to a CAS server +one of the proper ways to do so is:: + + $ tar cvjf vmcore.12345.tar.bz2 vmcore.12345 + +Other various ways of compressing archive are as follows:: + + $ tar cvzf vmcore.tar.gz vmcore + $ gzip vmcore + $ bzip2 vmcore + +``Note``: please do not double compress or CAS will fail. + +Another issue, which isn't primarily a fault of CAS, are +incomplete or corrupted cores. If either of these occur +there is a chance that CAS will not be able to process +the data needed to associate a debug kernel or do any +sort of automated analysis. Unfortunately, there is not +much that can be done to resolve these sort of issues +other than verifying that the process which happens when +a system coredump and when that dump reaches the +system specified for retrieval is solid and are seeing +no errors. + +Resources +========= + +* `CAS Wiki <http://fedorahosted.org/cas>`_ +* `CAS FAQ <https://fedorahosted.org/cas/wiki/CasFAQ>`_ +* `Mailing list <https://fedorahosted.org/mailman/listinfo/cas>`_ +* `Upstream releases <https://fedorahosted.org/releases/c/a/cas/>`_ +* Checkout latest from Git, ``git clone git://git.fedorahosted.org/cas.git`` diff --git a/lib/cas/db.py b/lib/cas/db.py index 6f7e942..9c0cac6 100644 --- a/lib/cas/db.py +++ b/lib/cas/db.py @@ -1,14 +1,24 @@ -""" storage container """ +import MySQLdb -class Core(object): - def __init__(self): - self.timestamp = None - self.debugKernel = None +class CasStorage(object): + def __init__(self, sql_server, username, password, database): + """ setup database connection and return db cursor for + traversing database """ + self.db = database + self.sql_server = sql_server + self.username = username + self.password = password + self.cursor = None -class CasServer(object): - """ Handles the definition of cas servers for each architecture - needing to be used for analyzing cores - """ - def __init__(self): - self.hostname = None - self.architecture = None + def connect(self): + """ execute connection """ + db_connect = MySQLdb.connect(host=self.sql_server, + user=self.username, + passwd=self.password, + db=self.db) + self.cursor = db_connect.cursor() + return + + def getJobs(self): + self.cursor.execute("SELECT * FROM cas_jobs") + return self.cursor.fetchall()

1 0

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

cas March 2009