[Help-gnu-radius] readius & postgres

[Oleg Gawriloff]

Hello!

When keepopen=no in sqlserver it seems that radiusd waits for long time for
something. E.g. in radius.log I often see following:
Nov 10 13:44:09 hawk2 radiusd: Acct.info: ðÒÏÃÅÓÓ ÚÁ×ÅÒÛÅÎ: ÐÒÅ×ÙÛÅÎÏ ×ÒÅÍÑ
ÏÖÉÄÁÎÉÑ ××ÏÄÁ ÏÔ ÏÓÎÏ×ÎÏÇÏ ÐÒÏÃÅÓÓÁ
Nov 10 13:44:09 hawk2 radiusd: Main.notice: child 59560 ×ÙÛÅÌ ÓÏ ÓÔÁÔÕÓÏÍ 0
Nov 10 13:44:10 hawk2 radiusd: Acct.info: ðÒÏÃÅÓÓ ÚÁ×ÅÒÛÅÎ: ÐÒÅ×ÙÛÅÎÏ ×ÒÅÍÑ
ÏÖÉÄÁÎÉÑ ××ÏÄÁ ÏÔ ÏÓÎÏ×ÎÏÇÏ ÐÒÏÃÅÓÓÁ
Nov 10 13:44:10 hawk2 radiusd: Main.notice: child 59568 ×ÙÛÅÌ ÓÏ ÓÔÁÔÕÓÏÍ 0
In postgres:
56441  ??  I      0:02,92 postmaster: billing RADIUS 172.16.2.135 idle
(postgres)
(i.e. sql query was completed, but radius does not drop db-connection).
sql-lost empty. During periods of heavy radius load (e.g. 50-100 queries per
second) radiusd spawns about 180 processes and completly hangs. How we can
fix this?

With best regards, Oleg Gawriloff.