Search

Changelog

Note: please regenerate your vha.vcl after each upgrade to avoid mismatches between Varnish and vha-agent

version 6.0.1r3 (2018-11-23)

  • vha.vcl isn’t generated anymore but instead bundled in the package. Please visit the installation documentation for more information.

version 2.1.5 (2018-08-06) (current version for Varnish Plus 4.1)

  • Add -k argument to allow insecure connection to HTTPS nodes
  • Force the connection header in replication request (close for HTTP, keep-alive for HTTPS), fixing the DNS caching issue.

Varnish High Availability 2.1.4 (2018-01-29)

  • Fix crash when the VHA only sees partial VSL transactions
  • Fix argument parsing error for the -t parameter (VSL timeout)

Varnish High Availability 2.1.3 (2018-01-15)

  • Fix ESI sub-requests containing an Accept-Encoding header losing headers when replicated

Varnish High Availability 2.1.1 (2017-12-06)

  • Only copy request body if the request is not a replication
  • Cache ESI request independently (vs relying on the top request previously)
  • Fix hit-for-pass issues by re-instating vha_skip_grace in vha.vcl
  • Force going through vcl_backend_error in case of error
  • Fix RHEL6 service job not daemonizing
  • Fix permission issues in RHEL packages for configuration files
  • Fix unit file permission for systemd system

Varnish High Availability 2.1.0 (2017-11-20)

  • Don’t omit empty headers
  • Auto-call vha_skip_grace and vha_clean_headers
  • Simplify vha.vcl by checking the request only once
  • Add node name and token in the stat file
  • Support body replication (through vmod-vha)
  • Support header addition (through vmod-vha)
  • Support replication sanity check (VSL overflow, through vmod-vha)
  • Support replication prevention through vmod-vha

Varnish High Availability 2.0.6 (2016-05-10)

  • Empty release

Varnish High Availability 2.0.5 (2016-05-3)

  • Fix generated vcl

Varnish High Availability 2.0.4 (2016-04-28)

  • In new Varnish Enterprise versions, vmod-goto can only be called from a backend context, so vha-generate-vcl had to be changed. Your vha.vcl MUST be re-generated.

Varnish High Availability 2.0.3 (2016-12-14)

  • Fix vha-generate-vcl manual

Varnish High Availability 2.0.2 (2016-12-08)

  • Fix python 2.6 incompatibility

Varnish High Availability 2.0.1 (2016-12-07)

  • Fix headers not being replicated
  • Fix packaging permissions for RHEL
  • Modify time in vha-skip-grace: force miss if obj.ttl <= 1s

Varnish High Availability 2.0.0 (2016-11-18)

  • Add cluster support
  • Allow replication prevention
  • Ability to use stale VSL file

Varnish High Availability 1.2.3 (2016-04-23)

  • Fix packaging issues
  • Fix memory leak in case of faulty VSL records

Varnish High Availability 1.2.2 (2016-01-26)

  • Fix maximum number of job per node

Varnish High Availability 1.2.1 (2015-12-11)

  • Don’t 403 on unknown x-vha headers

Varnish High Availability 1.2.0 (2015-12-11)

  • ESI support
  • New -S (Strict) switch
  • Accept empty configuration file
  • -m is now optional (use gethostbyname as fallback)
  • Create tmp stat file next to the stat file (avoid warnings)
  • Fix allowed character sets in node names
  • Better thread scheduling
  • Backlog is now per node

Varnish High Availability 1.1.5 (2015-12-02)

  • Don’t crash if more than one neighbor and no host header

Varnish High Availability 1.1.4 (2015-10-09)

  • Fix memory leak
  • Fix bogus ssl statement in generated configuration

Varnish High Availability 1.1.3 (2015-10-03)

  • Update the installation doc

Varnish High Availability 1.1.2 (2015-10-02)

Initial release