gocsaf/docs/csaf_aggregator.md

csaf_aggregator

Usage

  csaf_aggregator [OPTIONS]

Application Options:
  -c, --config=CFG-FILE    File name of the configuration file (default:
                           aggregator.toml)
      --version            Display version of the binary
  -i, --interim            Perform an interim scan

Help Options:
  -h, --help               Show this help message

Usage example for a single run, to test if the config is good:

./csaf_aggregator -c docs/examples/aggregator.toml

Once the config is good, you can run the aggregator periodically in two modes: full and interim.

Here is a complete example using cron on Ubuntu: After placing the config file in /etc/csaf_aggregator.toml and making sure its permissions only allow the user www-data to read it:

chown www-data /etc/csaf_aggregator.toml
chmod go-rwx /etc/csaf_aggregator.toml

mkdir /var/log/csaf_aggregator
mkdir ~www-data/bin
cp bin-linux-amd64/csaf_aggregator ~www-data/bin/
chown www-data.www-data -R ~www-data/bin /var/log/csaf_aggregator

# list current crontab
crontab -u www-data -l
# edit crontab (add lines like example below)
crontab -u www-data -e

Here is a crontab that runs the full mode once a day and updating interim advisories every 60 minutes:

SHELL=/bin/bash
# run full mode in the night at 04:00
0 4 * * * $HOME/bin/csaf_aggregator --config /etc/csaf_aggregator.toml >> /var/log/csaf_aggregator/full.log 2>&1
# run in interim mode once per hour at 30 minutes, e.g. 00:30, 01:30, ...
30 0-23 * * * $HOME/bin/csaf_aggregator --config /etc/csaf_aggregator.toml --interim >> /var/log/csaf_aggregator/interim.log 2>&1

serve via web server

Serve the paths where the aggregator writes its html/ output by means of a webserver. In the config example below the place in the filesystem is configured by the path given for web.

The user running the aggregator has to be able to write there and the web server must be able to read the files.

If you are using nginx, the setup instructions for the provider give a template. For the aggregator the difference is that you can leave out the cgi-bin part, potentially commend out the TLS client parts and adjust the root path accordingly.

config options

The following options can be used in the config file in TOML format:

workers               // number of parallel workers to start (default 10)
folder                // target folder on disc for writing the downloaded documents
web                   // directory to be served by the webserver
domain                // base url where the contents will be reachable from outside
rate                  // overall downloading limit per worker
insecure              // do not check validity of TLS certificates
write_indices         // write index.txt and changes.csv
openpgp_private_key   // OpenPGP private key (must have no passphrase set, if
                      // you want to be able to run unattended, e.g. via cron.)
openpgp_public_key    // OpenPGP public key
passphrase            // passphrase of the OpenPGP key
lock_file             // path to lockfile, to stop other instances if one is not done
interim_years         // limiting the years for which interim documents are searched
verbose               // print more diagnostic output, e.g. https request
allow_single_provider // debugging option
remote_validator      // use remote validation checker
aggregator            // table with basic infos for the aggregator object
providers             // array of tables, each entry to be mirrored or listed

At least 2 providers have to be configured. The other options are optional.

Rates are specified as floats in HTTPS operations per second. 0 means no limit. The rates can be specified per provider. Any provider which has no specified rate will use the configured default. If no default was configured, the internal default with no limits is used.

providers is an array of tables, each allowing

name
domain
rate
insecure
write_indices
category
update_interval

Any provider needs a valid name and domain. The other options are optional.

If you want an entry to be listed instead of mirrored in a aggregator.category == "aggregator" instance, set category to lister in the entry. Otherwise it is recommended to not set category for entries.

If a provider's domain starts with https:// it is considered a publisher. These publishers are added to the csaf_publishers list, written to the resulting aggregator.json. Each publisher must announce an update_interval there. This can be configured for each entry, by the config option with the same name. If not given it is taken from the configured default Otherwise, the internal default "on best effort" is used.

If a provider's create_service_document option is set to true, a service.json will be written listing its ROLIE feed documents. If it is not set or set to false, then no service.json will be written.

To offer an easy way of assorting CSAF documents by criteria like document category, languages or values of the branch category within the product tree, ROLIE category values can be configured. This can either be done using an array of strings taken literally or, by prepending "expr:". The latter is evaluated as JSONPath and the result will be added into the categories document. For a more detailed explanation and examples, refer to the provider config.

Example config file

workers = 2
folder = "/var/csaf_aggregator"
lock_file = "/var/csaf_aggregator/run.lock"
web = "/var/csaf_aggregator/html"
domain = "https://localhost:9443"
rate = 10.0
insecure = true
#openpgp_private_key =
#openpgp_public_key =
#interim_years =
#passphrase =
#write_indices = false

# specification requires at least two providers (default),
# to override for testing, enable:
# allow_single_provider = true

[aggregator]
  # Set if this instance shall be a mirror (aka `aggregator`) or a `lister`.
  # This determines the default value for the entries in [[provider]].
  category = "aggregator"
  name = "Example Development CSAF Aggregator"
  contact_details = "some @ somewhere"
  issuing_authority = "This service is provided as it is. It is gratis for everybody."
  namespace = "https://testnamespace.example.org"

[[providers]]
  name = "local-dev-provider"
  domain = "localhost"
  categories = ["Example Company Product A", "expr:document.lang"]
  create_service_document = true
#  rate = 1.5
#  insecure = true

[[providers]]
  name = "local-dev-provider2"
  domain = "https://localhost:8443/.well-known/csaf/provider-metadata.json"
#  rate = 1.2
#  insecure = true
  write_indices = true

[[providers]]
  name = "local-dev-provider3"
  domain = "localhost"
#  rate = 1.8
#  insecure = true
  write_indices = true
  # If aggregator.category == "aggregator", set for an entry that should
  # be listed in addition:
  category = "lister"

Publish others' advisories

In case you want to provide CSAF advisories from others that only qualify as CSAF publishers, see how to use the csaf_aggregator as "CSAF proxy provider".