def testGetDate(self):
for ooid in self.yyyyoids:
assert self.baseDate == oo.dateFromOoid(
ooid), 'Expected %s got %s' % (self.baseDate,
oo.dateFromOoid(ooid))
assert 4 == oo.depthFromOoid(
ooid), 'Expected %d, got %d' % (4, oo.depthFromOoid(ooid))
assert None == oo.dateFromOoid(self.badooid0)
assert None == oo.dateFromOoid(self.badooid1)
def toDateFromName(self,uuid): """Given uuid, get the relative path to the top of the date directory from the name location""" depth = socorro_ooid.depthFromOoid(uuid) if not depth: depth = 4 # prior, when hardcoded depth=4, uuid[-8:] was yyyymmdd, year was always (20xx) ups = ['..' for x in range(depth+1)] ups.append(self.dateName) return os.sep.join(ups)
def testCreateNewOoid(self):
ooid = oo.createNewOoid()
ndate = oo.dateFromOoid(ooid)
ndepth = oo.depthFromOoid(ooid)
assert self.nowstamp == ndate, 'Expect date of %s, got %s' %(self.nowstamp,ndate)
assert oo.defaultDepth == ndepth, 'Expect default depth (%d) got %d' % (oo.defaultDepth,ndepth)
ooid = oo.createNewOoid(timestamp=self.xmas05)
ndate = oo.dateFromOoid(ooid)
ndepth = oo.depthFromOoid(ooid)
assert self.xmas05 == ndate, 'Expect date of %s, got %s' %(self.xmas05,ndate)
assert oo.defaultDepth == ndepth, 'Expect default depth (%d) got %d' % (oo.defaultDepth,ndepth)
for d in range(1,5):
ooid0 = oo.createNewOoid(depth=d)
ooid1 = oo.createNewOoid(timestamp=self.xmas05,depth=d)
ndate0 = oo.dateFromOoid(ooid0)
ndepth0 = oo.depthFromOoid(ooid0)
ndate1 = oo.dateFromOoid(ooid1)
ndepth1 = oo.depthFromOoid(ooid1)
assert self.nowstamp == ndate0, 'Expect date of %s, got %s' %(self.nowstamp,ndate0)
assert self.xmas05 == ndate1, 'Expect date of %s, got %s' %(self.xmas05,ndate1)
assert ndepth0 == ndepth1, 'Expect depth0(%d) == depth1(%d)' %(ndepth0,ndepth1)
assert d == ndepth0, 'Expect depth %d, got %d' % (d,ndepth0)
assert None == oo.depthFromOoid(self.badooid0)
assert None == oo.depthFromOoid(self.badooid1)
def testCreateNewOoid(self):
ooid = oo.createNewOoid()
ndate = oo.dateFromOoid(ooid)
ndepth = oo.depthFromOoid(ooid)
assert self.nowstamp == ndate, 'Expect date of %s, got %s' %(self.nowstamp,ndate)
assert oo.defaultDepth == ndepth, 'Expect default depth (%d) got %d' % (oo.defaultDepth,ndepth)
ooid = oo.createNewOoid(timestamp=self.xmas05)
ndate = oo.dateFromOoid(ooid)
ndepth = oo.depthFromOoid(ooid)
assert self.xmas05 == ndate, 'Expect date of %s, got %s' %(self.xmas05,ndate)
assert oo.defaultDepth == ndepth, 'Expect default depth (%d) got %d' % (oo.defaultDepth,ndepth)
for d in range(1,5):
ooid0 = oo.createNewOoid(depth=d)
ooid1 = oo.createNewOoid(timestamp=self.xmas05,depth=d)
ndate0 = oo.dateFromOoid(ooid0)
ndepth0 = oo.depthFromOoid(ooid0)
ndate1 = oo.dateFromOoid(ooid1)
ndepth1 = oo.depthFromOoid(ooid1)
assert self.nowstamp == ndate0, 'Expect date of %s, got %s' %(self.nowstamp,ndate0)
assert self.xmas05 == ndate1, 'Expect date of %s, got %s' %(self.xmas05,ndate1)
assert ndepth0 == ndepth1, 'Expect depth0(%d) == depth1(%d)' %(ndepth0,ndepth1)
assert d == ndepth0, 'Expect depth %d, got %d' % (d,ndepth0)
assert None == oo.depthFromOoid(self.badooid0)
assert None == oo.depthFromOoid(self.badooid1)
def testCreateNewOoid(self):
new_ooid = ooid.create_new_ooid()
ndate = ooid.dateFromOoid(new_ooid)
ndepth = ooid.depthFromOoid(new_ooid)
assert self.nowstamp == ndate
assert ooid.defaultDepth == ndepth
new_ooid = ooid.create_new_ooid(timestamp=self.xmas05)
ndate = ooid.dateFromOoid(new_ooid)
ndepth = ooid.depthFromOoid(new_ooid)
assert self.xmas05 == ndate
assert ooid.defaultDepth == ndepth
for d in range(1, 5):
ooid0 = ooid.create_new_ooid(depth=d)
ooid1 = ooid.create_new_ooid(timestamp=self.xmas05, depth=d)
ndate0 = ooid.dateFromOoid(ooid0)
ndepth0 = ooid.depthFromOoid(ooid0)
ndate1 = ooid.dateFromOoid(ooid1)
ndepth1 = ooid.depthFromOoid(ooid1)
assert self.nowstamp == ndate0
assert self.xmas05 == ndate1
assert ndepth0 == ndepth1
assert d == ndepth0
assert ooid.depthFromOoid(self.badooid0) is None
assert ooid.depthFromOoid(self.badooid1) is None
def testCreateNewOoid(self):
new_ooid = ooid.create_new_ooid()
ndate = ooid.dateFromOoid(new_ooid)
ndepth = ooid.depthFromOoid(new_ooid)
assert self.nowstamp == ndate
assert ooid.defaultDepth == ndepth
new_ooid = ooid.create_new_ooid(timestamp=self.xmas05)
ndate = ooid.dateFromOoid(new_ooid)
ndepth = ooid.depthFromOoid(new_ooid)
assert self.xmas05 == ndate
assert ooid.defaultDepth == ndepth
for d in range(1, 5):
ooid0 = ooid.create_new_ooid(depth=d)
ooid1 = ooid.create_new_ooid(timestamp=self.xmas05, depth=d)
ndate0 = ooid.dateFromOoid(ooid0)
ndepth0 = ooid.depthFromOoid(ooid0)
ndate1 = ooid.dateFromOoid(ooid1)
ndepth1 = ooid.depthFromOoid(ooid1)
assert self.nowstamp == ndate0
assert self.xmas05 == ndate1
assert ndepth0 == ndepth1
assert d == ndepth0
assert ooid.depthFromOoid(self.badooid0) is None
assert ooid.depthFromOoid(self.badooid1) is None
def new_entry(self,
crash_id,
raw_crash,
dumps_dict,
webhead_host_name='webhead01',
timestamp=None):
if not isinstance(dumps_dict, collections.Mapping):
dumps_dict = {self.dump_field: dumps_dict}
name_dir, date_dir = super(JsonDumpStorage, self).newEntry(
crash_id,
timestamp,
webhead_host_name
)
raw_crash_pathname = os.path.join(
name_dir,
crash_id + self.jsonSuffix
)
with open(raw_crash_pathname, "w") as rcf:
json.dump(raw_crash, rcf)
for dump_name, dump in dumps_dict.iteritems():
full_dump_name = self.dump_file_name(crash_id, dump_name)
dump_pathname = os.path.join(
name_dir,
full_dump_name
)
with open(dump_pathname, "w") as dp:
dp.write(dump)
self.osModule.chmod(dump_pathname, self.dumpPermissions)
name_depth = socorro_ooid.depthFromOoid(crash_id)
if not name_depth:
name_depth = 4
rparts = [os.path.pardir, ] * (1 + name_depth)
rparts.append(self.dateName)
date_depth = 2 # .../hh/mm_slot...
if webhead_host_name and self.subSlotCount:
date_depth = 3 # .../webHeadName_slot
date_parts = date_dir.split(os.path.sep)[-date_depth:]
rparts.extend(date_parts)
self.osModule.symlink(
os.path.sep.join(rparts),
os.path.join(name_dir, crash_id)
)
def _get_radix(crash_id):
return [crash_id[i * 2 : (i + 1) * 2] for i in range(depthFromOoid(crash_id))]
def testGetDate(self):
for ooid in self.yyyyoids:
assert self.baseDate == oo.dateFromOoid(ooid), 'Expected %s got %s' %(self.baseDate, oo.dateFromOoid(ooid))
assert 4 == oo.depthFromOoid(ooid), 'Expected %d, got %d' %(4, oo.depthFromOoid(ooid))
assert None == oo.dateFromOoid(self.badooid0)
assert None == oo.dateFromOoid(self.badooid1)
try:
self.logger.debug('%s - about to copy dump %s to %s', threading.currentThread().getName(), dumppath,dumpNewPath)
shutil.copy2(dumppath,dumpNewPath)
self.osModule.chmod(dumpNewPath,self.dumpPermissions)
if self.dumpGID:
self.osModule.chown(dumpNewPath,-1,self.dumpGID)
self.osModule.chown(jsonNewPath,-1,self.dumpGID)
except OSError, e:
try:
self.osModule.unlink(jsonNewPath)
finally:
raise e
if createLinks:
self.logger.debug('%s - building links', threading.currentThread().getName())
dateDir,dparts = self.makeDateDir(timestamp,webheadHostName)
nameDepth = socorro_ooid.depthFromOoid(ooid)
if not nameDepth: nameDepth = 4
nameToDateParts = [os.pardir,]*(1+nameDepth)
nameToDateParts.extend(dparts[2:])
self.osModule.symlink(os.sep.join(nameToDateParts),os.path.join(nameDir,ooid))
try:
dateToNameParts = [os.pardir,]*(len(dparts)-2)
dateToNameParts.extend(nparts[2:])
self.osModule.symlink(os.sep.join(dateToNameParts),os.path.join(dateDir,ooid))
except OSError, e:
self.osModule.unlink(os.path.join(nameDir,ooid))
raise e
if removeOld:
self.logger.debug('%s - removing old %s, %s', threading.currentThread().getName(), jsonpath, dumppath)
try:
self.osModule.unlink(jsonpath)
def __namePath(self,uuid,startswith): """Because the name structure is almost simple, so is the method that creates one""" depth = socorro_ooid.depthFromOoid(uuid) if not depth: depth = 4 # prior, when hardcoded depth=4, uuid[-8:] was yyyymmdd, year was always (20xx) # split the first 2*depth characters into duples, join them, and prepend startswith return os.sep.join([startswith,os.sep.join([ uuid[2*x:2*x+2] for x in range(depth)])])
def relativeNameParts(self, ooid):
depth = socorro_ooid.depthFromOoid(ooid)
if not depth:
depth = 4
return [ooid[2 * x: 2 * x + 2] for x in range(depth)]
def testGetDate(self):
for this_ooid in self.yyyyoids:
assert self.baseDate == ooid.dateFromOoid(this_ooid)
assert 4 == ooid.depthFromOoid(this_ooid)
assert ooid.dateFromOoid(self.badooid0) is None
assert ooid.dateFromOoid(self.badooid1) is None
def testGetDate(self):
for this_ooid in self.yyyyoids:
assert self.baseDate == ooid.dateFromOoid(this_ooid)
assert 4 == ooid.depthFromOoid(this_ooid)
assert ooid.dateFromOoid(self.badooid0) is None
assert ooid.dateFromOoid(self.badooid1) is None
class JsonDumpStorage(socorro_dumpStorage.DumpStorage):
"""
This class implements a file system storage scheme for the JSON and dump
files of the Socorro project. It create a tree with two branches: the name
branch and the date branch.
- The name branch consists of paths based on the first 8 characters of the
crash_id file name. It holds the two data files and a relative symbolic
link to the date branch directory associated with the particular
crash_id. see socorro.lib.ooid.py for details of date and depth
encoding within the crash_id
For the crash_id: 22adfb61-f75b-11dc-b6be-001322081225
- the json file is stored as
%(root)s/%(daypart)s/name/22/ad/22adfb61-f75b-11dc-b6be-001322081225
.json
- the dump file is stored as
%(root)s/name/22/ad/22adfb61-f75b-11dc-b6be-001322081225.dump
- the symbolic link is stored as
%(root)s/name/22/ad/22adfb61-f75b-11dc-b6be-001322081225
and (see below) references
%(toDateFromName)s/date/2008/12/25/12/05/webhead01_0
- The date branch consists of paths based on the year, month, day, hour,
minute-segment, webhead host name and a small sequence number.
For each crash_id, it holds a relative symbolic link referring to the
actual storage (name) directory holding the data for that crash_id.
For the crash_id above, submitted at 2008-12-25T12:05 from webhead01
- the symbolic link is stored as
%(root)s/date/2008/09/30/12/05/webhead01_0/22adfb61-f75b-11dc-b6be-
001322081225 and references %(toNameFromDate)s/name/22/ad/
Note: The symbolic links are relative, so they begin with some rounds of
'../'. This is to avoid issues that might arise from variously mounted nfs
volumes. If the layout changes, self.toNameFromDate and toDateFromName
must be changed to match, as well as a number of the private methods.
Note: If so configured, the bottom nodes in the date path will be
%(webheadName)s_n for n in range(N) for some reasonable (5, perhaps) N.
Files are placed into these buckets in rotation.
"""
#--------------------------------------------------------------------------
def __init__(self, root=".", osModule=os, **kwargs):
"""
Take note of our root directory and other necessities.
Yes, it is perfectly legal to call super(...).__init__() after doing
some other code. As long as you expect the behavior you get, anyway...
"""
kwargs.setdefault('minutesPerSlot', 1)
kwargs.setdefault('subSlotCount', 1) # that is: use xxx_0 every time
# by default
super(JsonDumpStorage, self).__init__(root=root,
osModule=osModule,
**kwargs)
tmp = kwargs.get('cleanIndexDirectories', 'false')
self.cleanIndexDirectories = 'true' == tmp.lower()
self.jsonSuffix = kwargs.get('jsonSuffix', '.json')
if not self.jsonSuffix.startswith('.'):
self.jsonSuffix = ".%s" % (self.jsonSuffix)
self.dumpSuffix = kwargs.get('dumpSuffix', '.dump')
if not self.dumpSuffix.startswith('.'):
self.dumpSuffix = ".%s" % (self.dumpSuffix)
self.logger = kwargs.get('logger', socorro_util.FakeLogger())
#--------------------------------------------------------------------------
def new_entry(self,
crash_id,
raw_crash,
dumps_dict,
webhead_host_name='webhead01',
timestamp=None):
if not isinstance(dumps_dict, collections.Mapping):
dumps_dict = {self.dump_field: dumps_dict}
name_dir, date_dir = super(JsonDumpStorage, self).newEntry(
crash_id,
datetimeFromISOdateString(raw_crash['submitted_timestamp']),
webhead_host_name)
raw_crash_pathname = os.path.join(name_dir, crash_id + self.jsonSuffix)
with open(raw_crash_pathname, "w") as rcf:
json.dump(raw_crash, rcf)
for dump_name, dump in dumps_dict.iteritems():
full_dump_name = self.dump_file_name(crash_id, dump_name)
dump_pathname = os.path.join(name_dir, full_dump_name)
with open(dump_pathname, "w") as dp:
dp.write(dump)
self.osModule.chmod(dump_pathname, self.dumpPermissions)
name_depth = socorro_ooid.depthFromOoid(crash_id)
if not name_depth:
name_depth = 4
rparts = [
os.path.pardir,
] * (1 + name_depth)
rparts.append(self.dateName)
date_depth = 2 # .../hh/mm_slot...
if webhead_host_name and self.subSlotCount:
date_depth = 3 # .../webHeadName_slot
date_parts = date_dir.split(os.path.sep)[-date_depth:]
rparts.extend(date_parts)
self.osModule.symlink(os.path.sep.join(rparts),
os.path.join(name_dir, crash_id))
#--------------------------------------------------------------------------
def newEntry(self, crash_id, webheadHostName='webhead01', timestamp=None):
"""
Sets up the name and date storage directory branches for the given
crash_id. Creates any directories that it needs along the path to the
appropriate storage location.
Creates two relative symbolic links: the date branch link pointing to
the name directory holding the files;
the name branch link pointing to the date branch directory holding that
link.
Returns a 2-tuple containing files open for writing: (jsonfile,
dumpfile)
If self.dumpGID, then the file tree from root to and including the data
files are chown'd
If self.dumpPermissions, then chmod is called on the data files
"""
# note: after this call, dateDir already holds link to nameDir
nameDir, dateDir = super(JsonDumpStorage,
self).newEntry(crash_id, timestamp,
webheadHostName)
df, jf = None, None
jname = os.path.join(nameDir, crash_id + self.jsonSuffix)
try:
jf = open(jname, 'w')
except IOError, x:
if 2 == x.errno:
nameDir = self.makeNameDir(crash_id, timestamp) # deliberately
# leave this dir behind if next line throws
jf = open(jname, 'w')
else:
raise x
try:
# Do all this in a try/finally block to unroll in case of error
self.osModule.chmod(jname, self.dumpPermissions)
dname = os.path.join(nameDir, crash_id + self.dumpSuffix)
df = open(dname, 'w')
self.osModule.chmod(dname, self.dumpPermissions)
nameDepth = socorro_ooid.depthFromOoid(crash_id)
if not nameDepth:
nameDepth = 4
rparts = [
os.path.pardir,
] * (1 + nameDepth)
rparts.append(self.dateName)
dateDepth = 2 # .../hh/mm_slot...
if webheadHostName and self.subSlotCount:
dateDepth = 3 # .../webHeadName_slot
dateParts = dateDir.split(os.path.sep)[-dateDepth:]
rparts.extend(dateParts)
self.osModule.symlink(os.path.sep.join(rparts),
os.path.join(nameDir, crash_id))
if self.dumpGID:
def chown1(path):
self.osModule.chown(path, -1, self.dumpGID)
socorro_fs.visitPath(
self.root, os.path.join(nameDir,
crash_id + self.jsonSuffix),
chown1, self.osModule)
self.osModule.chown(
os.path.join(nameDir, crash_id + self.dumpSuffix), -1,
self.dumpGID)
# socorro_fs.visitPath(self.root,
# os.path.join(dateDir,crash_id),
# chown1
# )
finally:
if not jf or not df:
if jf:
jf.close()
if df:
df.close()
try:
self.osModule.unlink(os.path.join(dateDir, crash_id))
except Exception:
pass # ok if not there
try:
self.osModule.unlink(os.path.join(nameDir, crash_id))
except Exception:
pass # ok if not there
df, jf = None, None
return (jf, df)
def _get_radix(crash_id):
return [
crash_id[i * 2:(i + 1) * 2] for i in range(depthFromOoid(crash_id))
]
self.logger.debug('about to copy dump %s to %s', dumppath,
dumpNewPath)
shutil.copy2(dumppath, dumpNewPath)
self.osModule.chmod(dumpNewPath, self.dumpPermissions)
if self.dumpGID:
self.osModule.chown(dumpNewPath, -1, self.dumpGID)
self.osModule.chown(jsonNewPath, -1, self.dumpGID)
except OSError, e:
try:
self.osModule.unlink(jsonNewPath)
finally:
raise e
if createLinks:
self.logger.debug('building links')
dateDir, dparts = self.makeDateDir(timestamp, webheadHostName)
nameDepth = socorro_ooid.depthFromOoid(ooid)
if not nameDepth:
nameDepth = 4
nameToDateParts = [
os.pardir,
] * (1 + nameDepth)
nameToDateParts.extend(dparts[2:])
self.osModule.symlink(os.sep.join(nameToDateParts),
os.path.join(nameDir, ooid))
try:
dateToNameParts = [
os.pardir,
] * (len(dparts) - 2)
dateToNameParts.extend(nparts[2:])
self.osModule.symlink(os.sep.join(dateToNameParts),
os.path.join(dateDir, ooid))
