def store(self):
"""Sync all queued repositories. Process repositories in batches due to disk space and memory usage."""
self.logger.log("Checking %d repositories." % len(self.repositories))
# Fetch current list of repositories from DB
self.db_repositories = self.repo_store.list_repositories()
# Download all repomd files first
failed = self._download_repomds()
self.logger.log("%d repomd.xml files failed to download." % len(failed))
self._read_repomds(failed)
# Filter all repositories without repomd attribute set (failed download, downloaded repomd is not newer)
batches = BatchList()
to_skip = []
for repository in self.repositories:
if repository.repomd:
batches.add_item(repository)
else:
to_skip.append(repository)
self.clean_repodata(to_skip)
self.logger.log("%d repositories skipped." % len(to_skip))
self.logger.log("Syncing %d repositories." % sum(len(l) for l in batches))
# Download and process repositories in batches (unpacked metadata files can consume lot of disk space)
for batch in batches:
self._download_metadata(batch)
self._unpack_metadata(batch)
for repository in batch:
repository.load_metadata()
self.repo_store.store(repository)
repository.unload_metadata()
self.clean_repodata(batch)
def store(self):
"""Sync all queued CVE lists. Runs in batches due to disk space and memory usage."""
self.logger.info("Checking %d CVE lists.", len(self.repos))
# Download all repomd files first
failed = self._download_meta()
if failed:
FAILED_NIST.inc()
self.logger.warning("%d meta files failed to download.",
len(failed))
self._read_meta(failed)
# filter out failed / unchanged lists
batches = BatchList()
to_skip = []
for repo in self.repos:
if repo.meta:
batches.add_item(repo)
else:
to_skip.append(repo)
self.clean_repo(to_skip)
self.logger.info("%d CVE lists skipped.", len(to_skip))
self.logger.info("Syncing %d CVE lists.", sum(len(l) for l in batches))
# Download and process repositories in batches (unpacked metadata files can consume lot of disk space)
for batch in batches:
try:
self._download_json(batch)
self._unpack_json(batch)
for repo in sorted(batch, key=lambda repo: repo.label):
repo.load_json()
self.cverepo_store.store(repo)
repo.unload_json()
finally:
self.clean_repo(batch)
def store(self):
"""Sync all queued repositories. Process repositories in batches due to disk space and memory usage."""
self.logger.info("Checking %d repositories.", len(self.repositories))
self._write_certificate_cache()
# Download all repomd files first
failed = self._download_repomds()
if failed:
self.logger.warning("%d repomd.xml files failed to download.",
len(failed))
failed_repos = [
repo for repo in self.repositories
if self._repo_download_failed(repo, failed)
]
self.clean_repodata(failed_repos)
self._read_repomds()
# Filter all repositories without repomd attribute set (failed download, downloaded repomd is not newer)
batches = BatchList()
to_skip = []
for repository in self.repositories:
if repository.repomd:
batches.add_item(repository)
else:
to_skip.append(repository)
self.clean_repodata(to_skip)
self.logger.info("%d repositories skipped.", len(to_skip))
self.logger.info("Syncing %d repositories.",
sum(len(l) for l in batches))
# Download and process repositories in batches (unpacked metadata files can consume lot of disk space)
for batch in batches:
failed = self._download_metadata(batch)
if failed:
self.logger.warning("%d metadata files failed to download.",
len(failed))
failed_repos = [
repo for repo in batch
if self._repo_download_failed(repo, failed)
]
self.clean_repodata(failed_repos)
batch = [repo for repo in batch if repo not in failed_repos]
self._unpack_metadata(batch)
for repository in batch:
repository.load_metadata()
self.repo_store.store(repository)
repository.unload_metadata()
self.clean_repodata(batch)
self.repo_store.cleanup_unused_data()
self._clean_certificate_cache()
class TestBatchList:
"""TestBatchList class. Test creating list of lists"""
@pytest.fixture()
def batchlist(self):
"""Setup for batchlist testing."""
self.blist = BatchList()
def test_empty_batch(self, batchlist):
"""Test empty batchlist."""
assert not self.blist.batches
# Assuming default is 50, 102 = 3 batches, 50/50/2 ; 150 = 50/50/50; 157 == 4, 50/50/50/7
# move thru the batches, making sure each other than the last is at most BATCH_MAX_SIZE long
# and each batch has cumulative file_size less than BATCH_MAX_FILESIZE
@pytest.mark.parametrize("list_size", [102, 150, 157])
@pytest.mark.parametrize("item_filesize", ITEM_FILESIZES)
def test_batch_creation(self, batchlist, list_size, item_filesize):
"""Test creation of batch list."""
for i in range(list_size):
self.blist.add_item(i, item_filesize)
# batch size is variable, if items are too large, the batch might contain less than BATCH_MAX_SIZE items
batch_size = min(int(BATCH_MAX_SIZE),
int(BATCH_MAX_FILESIZE) // item_filesize)
total_batches = math.ceil(list_size / batch_size)
last_batch_size = list_size % batch_size
assert len(self.blist.batches) == total_batches
for curr_batch in range(total_batches):
if curr_batch == (total_batches - 1) and last_batch_size > 0:
expected_num_in_batch = last_batch_size
else:
expected_num_in_batch = batch_size
assert len(self.blist.batches[curr_batch]) == expected_num_in_batch
def test_invalid_batch_size(self, batchlist):
"""
Test creation of an invalid batch list.
Should fail because single item is larger than max batch size
"""
with pytest.raises(AssertionError):
self.test_batch_creation(batchlist, 102,
int(BATCH_MAX_FILESIZE) + 1)
def store(self):
"""Sync all queued repositories. Process repositories in batches due to disk space and memory usage."""
self.logger.info("Checking %d repositories.", len(self.repositories))
self._write_certificate_cache()
# Download all repomd files first
failed = self._download_repomds()
if failed:
FAILED_REPOMD.inc(len(failed))
failed_repos = [repo for repo in self.repositories if self._repo_download_failed(repo, failed)]
self.logger.warning("%d repomd.xml files failed to download.", len(failed))
self.clean_repodata(failed_repos)
self._read_repomds()
# Filter all repositories without repomd attribute set (downloaded repomd is not newer)
batches = BatchList()
up_to_date = []
def md_size(repomd, data_type):
try:
mdata = repomd.get_metadata(data_type)
# open-size is not present for uncompressed files
return int(mdata.get('size', 0)) + int(mdata.get('open-size', '0'))
except RepoMDTypeNotFound:
return 0
for repository in self.repositories:
if repository.repomd:
repo_size = md_size(repository.repomd, 'primary_db')
# If we use primary_db, we don't even download primary data xml
if repo_size == 0:
repo_size += md_size(repository.repomd, 'primary')
repo_size += md_size(repository.repomd, 'updateinfo')
repo_size += md_size(repository.repomd, 'modules')
batches.add_item(repository, repo_size)
else:
up_to_date.append(repository)
self.clean_repodata(up_to_date)
self.logger.info("%d repositories are up to date.", len(up_to_date))
total_repositories = batches.get_total_items()
completed_repositories = 0
self.logger.info("%d repositories need to be synced.", total_repositories)
# Download and process repositories in batches (unpacked metadata files can consume lot of disk space)
try:
for batch in batches:
self.logger.info("Syncing a batch of %d repositories", len(batch))
try:
failed = self._download_metadata(batch)
if failed:
self.logger.warning("%d metadata files failed to download.", len(failed))
failed_repos = [repo for repo in batch if self._repo_download_failed(repo, failed)]
self.clean_repodata(failed_repos)
batch = [repo for repo in batch if repo not in failed_repos]
self._unpack_metadata(batch)
for repository in batch:
repository.load_metadata()
completed_repositories += 1
self.logger.info("Syncing repository: %s [%s/%s]", ", ".join(
filter(None, (repository.content_set, repository.basearch, repository.releasever))),
completed_repositories, total_repositories)
self.repo_store.store(repository)
repository.unload_metadata()
finally:
self.clean_repodata(batch)
finally:
self.repo_store.cleanup_unused_data()
self._clean_certificate_cache()
def batchlist(self):
"""Setup for batchlist testing."""
self.blist = BatchList()
class TestBatchList:
""" Batch list tests group. """
batch_list = None
DUMMY_ITEMS = [{"item": 1}, {"item": 2}, {"item": 3}]
@pytest.fixture(autouse=True)
def setup_batch_list(self):
""" Fixture for creating batch list. """
self.batch_list = BatchList()
def test_empty(self):
""" Test for empty batch list. """
assert self.batch_list.get_total_items() == 0
def test_insert_few(self):
""" Test for inserton of one item. """
for item in self.DUMMY_ITEMS:
self.batch_list.add_item(item)
assert self.batch_list.get_total_items() == len(self.DUMMY_ITEMS)
def test_insert_full_batch(self):
""" Test for insertion of multiple items, until new batch needs to be created. """
max_size = int(BATCH_MAX_SIZE)
for count in range(0, max_size + 1):
self.batch_list.add_item({"item": count})
assert self.batch_list.get_total_items() == max_size + 1
assert len(self.batch_list.batches) == 2
assert len(self.batch_list.batches[0]) == max_size
assert len(self.batch_list.batches[1]) == 1
def test_clear(self):
""" Test for clearing the list. """
for item in self.DUMMY_ITEMS:
self.batch_list.add_item(item)
self.batch_list.clear()
assert self.batch_list.batches == []
def setup_batch_list(self):
""" Fixture for creating batch list. """
self.batch_list = BatchList()
def store(self):
"""Sync all queued repositories. Process repositories in batches due to disk space and memory usage."""
self.logger.info("Checking %d repositories.", len(self.repositories))
self._write_certificate_cache()
# Download all repomd files first
failed = self._download_repomds()
if failed:
FAILED_REPOMD.inc(len(failed))
failed_repos = [
repo for repo in self.repositories
if self._repo_download_failed(repo, failed)
]
self.logger.warning("%d repomd.xml files failed to download.",
len(failed))
self.clean_repodata(failed_repos)
self._read_repomds()
# Filter all repositories without repomd attribute set (downloaded repomd is not newer)
batches = BatchList()
up_to_date = []
for repository in self.repositories:
if repository.repomd:
batches.add_item(repository)
else:
up_to_date.append(repository)
self.clean_repodata(up_to_date)
self.logger.info("%d repositories are up to date.", len(up_to_date))
total_repositories = batches.get_total_items()
completed_repositories = 0
self.logger.info("%d repositories need to be synced.",
total_repositories)
# Download and process repositories in batches (unpacked metadata files can consume lot of disk space)
try:
for batch in batches:
try:
failed = self._download_metadata(batch)
if failed:
self.logger.warning(
"%d metadata files failed to download.",
len(failed))
failed_repos = [
repo for repo in batch
if self._repo_download_failed(repo, failed)
]
self.clean_repodata(failed_repos)
batch = [
repo for repo in batch if repo not in failed_repos
]
self._unpack_metadata(batch)
for repository in batch:
repository.load_metadata()
completed_repositories += 1
self.logger.info(
"Syncing repository: %s [%s/%s]", ", ".join(
filter(None, (repository.content_set,
repository.basearch,
repository.releasever))),
completed_repositories, total_repositories)
self.repo_store.store(repository)
repository.unload_metadata()
finally:
self.clean_repodata(batch)
finally:
self.repo_store.cleanup_unused_data()
self._clean_certificate_cache()
