After takin a swig o' Arrakan spice grog, tgl(at)sss(dot)pgh(dot)pa(dot)us (Tom Lane) belched out:
> Bill Chandler <billybobc1210(at)yahoo(dot)com> writes:
>> LOG: invalid entry in pg_hba.conf file at line 60,
>> token "255.255.255.255"
>> I have found that if I change this value to anything
>> other than "255.255.255.255" (e.g. "255.255.255.254")
>> the message goes away.
> Bizarre. AFAICS this must indicate that getaddrinfo() is rejecting the
> value, which perhaps could happen if some bit of it were miscoded to
> use "-1" as a failure code.
>> Interestingly the same 7.4.1 binary running on a different machine did
>> not exhibit the problem. Could it be a Solaris patch level thing?
> Seems likely. Can you update and see if that fixes it?
Hmm... I think I have seen the same thing.
It went away when I substituted in some "grouped" entries, applying
policies to subnets rather than to individual hosts. That happened to
be a more appropriate model to use, as it cuts WAY down on the need
for pg_hba.conf entries, and on the need to modify configuration in
(We used to have, on test boxes, named IP addresses for people that
now get their IPs assigned by DHCP... Without subnet definitions,
that's a nightmare!)
We've got a box getting patched Saturday; I'll watch for that...
Do Roman paramedics refer to IV's as "4's"?
In response to
pgsql-admin by date
|Next:||From: Tom Lane||Date: 2004-05-19 23:44:21|
|Subject: Re: Servercrash |
|Previous:||From: Christopher Browne||Date: 2004-05-19 22:21:00|
|Subject: Re: postgresql not using enough memory?|