Upgrading to v8.1.x
Two issues have been identified in the latest v8.1.1 release, which affect certain configurations which includes parent proxying:
This is still being investigated, but anyone experiencing this should reach out to the community and see if we can identify these.
- traffic_cop has been removed. This removes the health checks that traffic_cop was making against traffic_server. Most installations were already performing health checks against traffic_server.
- Removed metrics.config
- Removed plugins
- congestion.config - The separate congestion control feature has been removed. There are more specific congestion management features controlled in records.config specifically proxy.config.net.default_inactivity_timeout, proxy.config.http.server_max_connections, and proxy.config.http.origin_max_connections.
- TSHttpArgIndexReserve - renamed to TSHttpTxnArgIndexReserve
- TSHttpArgIndexNameLookup - renamed to TSHttpTxnArgIndexNameLookup
- TSHttpTxnSetHttpRetStatus - renamed to TSHttpTxnStatusSet
- TSHttpSsnSSLConnectionGet - removed and split up into two calls using TSHttpSsnClientVConnGet and TSVConnSSLConnectionGet - the equivalent would be TSVConnSSLConnectionGet(TSHttpSsnClientVConnGet(x))
- The cache format in this release is compatible with previous 6.x, and 7.x releases
New configuration defaults (changes since 7.x)
proxy.config.http.negative_revalidating_enabled - default value has been changed to 1
New configuration options
proxy.config.http.post_copy_size - added to buffer POST bodies when the follow redirect feature is enabled (proxy.config.http.number_of_redirections > 0). It the POST body exceeds this value a warning message will be printed in diags.log and the POST will fail to the client.
- Removed metrics.config and all related stats