From owner-freebsd-database  Tue Apr 27  2: 7:11 1999
Delivered-To: freebsd-database@freebsd.org
Received: from router.int.odessa.ua (dp-109.EuroCom.Od.UA [212.15.128.109])
	by hub.freebsd.org (Postfix) with ESMTP id 74AB114ED9
	for <freebsd-database@FreeBSD.ORG>; Tue, 27 Apr 1999 02:06:45 -0700 (PDT)
	(envelope-from vns@maxa01.tm.odessa.ua)
Received: from int01 (instructor.int.odessa.ua [192.168.1.2])
	by router.int.odessa.ua (8.8.8/8.8.8) with SMTP id LAA01245;
	Tue, 27 Apr 1999 11:56:41 +0300 (EEST)
	(envelope-from vns@router.int.odessa.ua)
Message-ID: <372589D9.7C14@maxa01.tm.odessa.ua>
Date: Tue, 27 Apr 1999 11:56:41 +0200
From: "Vladimir N.Silyaev" <vns@maxa01.tm.odessa.ua>
Reply-To: vns@maxa01.tm.odessa.ua
Organization: JSC INT
X-Mailer: Mozilla 3.0Gold (Win95; I; 16bit)
MIME-Version: 1.0
To: Doug Poland <dpoland@execpc.com>
Cc: SYBASE-L@UCSBVM.ucsb.edu, freebsd-database@FreeBSD.ORG
Subject: Re: Sybase saga continues...
References: <000001be8db4$dbe1ca70$b16f6478@egypt>
Content-Type: text/plain; charset=koi8-r
Content-Transfer-Encoding: 7bit
Sender: owner-freebsd-database@FreeBSD.ORG
Precedence: bulk

Doug Poland wrote:

> The "strange" thing is, I do a "ps" and the dataserver is running and has
> 16MB of memory allocated and using about 7MB.
> 
> I tried to talk to the dataserver using isql -SSYBASE -Usa -P but got a
> connection timeout.
> 
> On the successful side, I did configure the backup server.  I went through
> the whole configuration and it booted!  Unfortunately, a message similar to
> 
>         "Warning" this process needs 1024 file descriptors but only has 512"
> 
> appeared.
> 
> So I'm wondering, where do I go now?  Am I still seeing a shared memory
> problem?
No, You are already solved shared memory problem.

> Does the files descriptors offer a clue?

For example via sysctl interface:
sysctl -w kern.maxfiles=1536 && sysctl -w kern.maxfilesperproc=1024 && \
echo -n 'Change FileDescriptor '


The rest of our problems there is a SIGIO problem,  the next instruction 
intend for solve this problem. This is my second post of this text, 
previous was delivered without subj.

Please follow the next instruction.
1. Create file sybaseio.sh
-----------------------
#!/bin/sh
pid=`ps -ax|awk  '/dataserver/&&!/awk/{print $1;exit;}'`
echo dataserver pid is $pid
while true; do
 kill -IO $pid || break
done
----------------------
2. At the stage of the installing Sybase SQL, when sybinit say:

Running task: update the SQL Server runserver file.
Task succeeded: update the SQL Server runserver file.
Running task: boot the SQL Server.
waiting for server 'SYBASE' to boot...

launch the sybaseio.sh script, the sybinit must print:

Task succeeded: boot the SQL Server.
Running task: create the sybsystemprocs database.

3. When sybinit finished run isql -U sa,  
enter password [empty], and after isql prompt:
sp_help
go

4. Shutdown dataserver:  kill <dataserver pid>.

5. Create sybase.pl script
---------
#!/usr/bin/perl5
use POSIX;

setsid();
exec $ARGV[0]
---------

6. Open in the editor file ${SYBASE}/install/RUN_SYBASE and
add the 'exec' word at the begin line where dataserver started,
for example:
...
#  interfaces:                    /opt/sybase
#
exec /opt/sybase/bin/dataserver \
...

7. Launch  sybase.pl install/RUN_SYBASE and repeat step 3.
   

This is a symptom of the problem, which are hacked with sybaseio.sh
script.

With best regards, V.Silyaev.


To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-database" in the body of the message