New release

[flionet89]

Good afternoon, could you publish a release candidate to test the resolution of previously open bugs? And then an updated release?

[x4m]

We have some open items regarding memory leaks. But things look like these leaks were inroduced in previous releases.

@reshke do we have something critical preventing release? Did we push all the stuff with fixes?
I know that @NikitaUnisikhin is working on new HBA with support for group authentication, but this feature does not need to be a part of this release.

[marcbachmann]

I'm still getting memory issues with the state from master.
The process exits with the error double free or corruption (out).

No other error gets logged.

Here's the config I'm using

storage "primary" {
  type "remote"
  port 5432
}

database "auth-query" {
  user "pooler" {
    authentication "none"
    pool_routing "internal"
    storage "primary"
    storage_db "postgres"
    pool "transaction"

    pool_size 20
    pool_timeout 10000
    pool_ttl 120
    server_lifetime 3600
  }
}

database default {
  user default {
    storage "primary"
    authentication "md5"
    auth_query "SELECT * from pooler.lookup_user($1)"
    auth_query_db "auth-query"
    auth_query_user "pooler"

    pool "transaction"
    pool_size 0
    pool_timeout 10000
    pool_ttl 120
    pool_cancel yes
    pool_rollback yes
    server_lifetime 3600
    client_fwd_error yes
    application_name_add_host yes
    log_debug no
  }
}

unix_socket_dir "/var/run/postgresql"
unix_socket_mode "0644"
nodelay yes
workers "auto"

log_format "%p %t %l [%i %s] (%c) %m\n"
log_debug no
log_config yes
log_session yes
log_query no
log_stats no

listen {
  host "*"
  port 6543
}

[vrenaville]

@marcbachmann
I think @x4m says that the problem was already there in 1.3, so it can be postpone to another release.

Many changes has been added since 1.3, like HBA syntax support etc ..., It will be great to have a new one :)

[marcbachmann]

I'm not sure it's the same memory issue. We've been running branch v1.3 since its release and only saw about 2 crashes. With the state of master we saw one after 6 hours (Exit Code 139: Segmentation Fault (SIGSEGV)).
This should be communicated as it might affect setups.