| Oracle FAQ | Your Portal to the Oracle Knowledge Grid | |
 |
 | |||
| Oracle FAQ | Your Portal to the Oracle Knowledge Grid | |
|
| |||
Home -> Community -> Mailing Lists -> Oracle-L -> Re: Does the client "hang" or "disconnect" when the server's
Hi!
Actually it seems that as long as the network (TCP) connection is up at OS level, Oracle is just gonna wait for reply for it's packets sent. So, probably no 3113 error before OS drops the TCP connection (if sqlplus or OCI doesn't have some kind of timeout implemented). Sorry about confusion...
I tested sending new request with sqlplus, by logging on as system and suspending, not killing my server process (this way OS doesn't clear the TCP connection):
SQL> select server from v$session where username = 'SYSTEM';
SERVER
SPID
SQL> !kill -23 3961
SQL> select * from dual;
^C ^C ^C
kill -23 means killing with SIGSTOP in Solaris and after that my next request waited indefinetly (at least for 10 minutes after which I continued the process with kill -25).
Tanel.
----- Original Message -----
From: "Hemant K Chitale" <hkchital_at_singnet.com.sg>
To: "Multiple recipients of list ORACLE-L" <ORACLE-L_at_fatcity.com>
Sent: Friday, July 25, 2003 6:24 PM
Subject: Re: Does the client "hang" or "disconnect" when the server's
>
> Unfortunately, I am not running OPS or RAC.
>
> This is a "failover-cluster" -- if the primary node goes down, the
> database instance
> is started on the secondary node.
>
> I think Tanel is right that sqlnet.expire_time wouldn't apply.
>
> However, Tanel, I didn't understand the paragraph
> "Also, sqlplus is different from normal telnet, because it tries to send a
> request only when it is on it's prompt, thus issuing a command in sqlplus
> session, when it is waiting for response from previous command, doesn't
send
> any network packets."
> Would I be right in understand it to mean that "when waiting for a
response
> from
> the database server, sqlplus doesn't send any network packets. however,
if
> sqlplus is sitting idle and a user enters a command sqlplus attempts to
> send the
> request and only then knows that the database or node is down" ?
> So, if my session were sending a fresh request, it would know that the
server
> was "unavailable" and would return an error [3113 ?] but if it were
waiting
> for
> a response [last issued an INSERT statement and is waiting for the ACK
packet]
> it would just "sit there waiting" ?
>
> Hemant
>
> At 10:04 AM 24-07-03 -0800, you wrote:
>
>
>
>
> >I agree sqlnet.expire_time does only effect the server, but normally when
the
> >server kills the connection the client side will drop. But in this case
it
> >won't because there is no network for the client to get the dead
connection
> >timeout from the server. It sounds like Hemant is running OPS which
AFAIK
> >(version 7.3) he is getting the expected behavior, to get the client to
> >seemlessly switch over to the other instance he needs RAC (again AFAIK).
> >
> >
> >
> >
> >
> >
> > "Tanel
> > Poder"
> >
> > <tanel.poder.003@ To: Multiple
> > recipients of list ORACLE-L <ORACLE-L_at_fatcity.com>
> > mail.ee> cc:
> >
> > Sent by: Subject: Re: Does the
> > client "hang" or "disconnect" when the server's
> > ml-errors_at_fatcity
> >
> > .com
> >
> >
> >
> >
> >
> > 07/24/2003
> > 12:19
> >
> > PM
> >
> > Please respond
> > to
> >
> > ORACLE-L
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >Hi!
> >
> >This sqlnet.expire_time affects server, not sqlplus itself AFAIK. The
issue
> >is likely, that sqlplus has sent a sqlnet request to server, but hasn't
got
> >any response since you pulled the network cable. Thus it'll wait until
TCP
> >connection gets timeout.
> >Also, sqlplus is different from normal telnet, because it tries to send a
> >request only when it is on it's prompt, thus issuing a command in sqlplus
> >session, when it is waiting for response from previous command, doesn't
send
> >any network packets.
> >
> >Tanel.
> >
> >----- Original Message -----
> >To: "Multiple recipients of list ORACLE-L" <ORACLE-L_at_fatcity.com>
> >Sent: Thursday, July 24, 2003 6:54 PM
> >
> >
> > >
> > >
> > >
> > >
> > > Hemant,
> > >
> > > You are right it does get a timeout from the TCP stack, but the
> >default
> > > for the connection timeout is 30 minutes. There is a sqlnet parameter
> >where you
> > > can modify the timeout to make it shorter (sqlnet.expire_time?) Check
> >metalink
> > > note 151972.1 for more info.
> > >
> > > HTH
> > > Mike
> > >
> > >
> > >
> > >
> > > Hemant K Chitale
> > > <hkchital_at_singnet To: Multiple
> >recipients of list ORACLE-L <ORACLE-L_at_fatcity.com>
> > > .com.sg> cc:
> > > Sent by: Subject: Does the
client
> >"hang" or "disconnect" when the server's
> > > ml-errors_at_fatcity
> > > .com
> >
> > >
> > >
> > > 07/24/2003 11:24
> > > AM
> > > Please respond to
> > > ORACLE-L
> > >
> > >
> > >
> > >
> > >
> > >
> > >
> > >
> > > I'm sorry, but I really can't remember how an Oracle SQLPlus client is
> >supposed
> > > to behave when the database server's network interface is "pulled
out".
> > > We were testing a new Sun Cluster and tests for Storage FC Path
Failure,
> > > Instance Failure,
> > > Node Failure, HeartBeat failure and single Network Interface failure
all
> > > went through.
> > >
> > > However, when the engineer pulled out *both* network interfaces of the
> >running
> > > database server node, the SQLPlus client that had been running INSERT
> > > statements
> > > just "hung". It did not exit or error out till we killed the job. At
the
> > > next restart, of course,
> > > it connected to the database which had already failed-over to the
second
> >node.
> > >
> > > The question is : Why does the SQLPlus client "hang" ? Telnet
sessions
> > > also might
> > > "hang" [couldnt' test this out as the Servers are Secured and we
didn't
> > > have SSH on the
> > > client], till the user hits the ENTER key, but shouldn't SQLPlus get
> >server
> > > disconnect
> > > from the TCP stack ?
> > >
> > >
> > > Hemant K Chitale
> > > Oracle 9i Database Administrator Certified Professional
> > > My personal web site is : http://hkchital.tripod.com
> > >
> > >
> > > --
> > > Please see the official ORACLE-L FAQ: http://www.orafaq.net
> > > --
> > > Author: Hemant K Chitale
> > > INET: hkchital_at_singnet.com.sg
> > >
> > > Fat City Network Services -- 858-538-5051 http://www.fatcity.com
> > > San Diego, California -- Mailing list and web hosting services
> > > ---------------------------------------------------------------------
> > > To REMOVE yourself from this mailing list, send an E-Mail message
> > > to: ListGuru_at_fatcity.com (note EXACT spelling of 'ListGuru') and in
> > > the message BODY, include a line containing: UNSUB ORACLE-L
> > > (or the name of mailing list you want to be removed from). You may
> > > also send the HELP command for other information (like subscribing).
> > >
> > >
> > >
> > > --
> > > Please see the official ORACLE-L FAQ: http://www.orafaq.net
> > > --
> > > Author: Michael Boligan
> > > INET: michael.boligan.b_at_bayer.com
> > >
> > > Fat City Network Services -- 858-538-5051 http://www.fatcity.com
> > > San Diego, California -- Mailing list and web hosting services
> > > ---------------------------------------------------------------------
> > > To REMOVE yourself from this mailing list, send an E-Mail message
> > > to: ListGuru_at_fatcity.com (note EXACT spelling of 'ListGuru') and in
> > > the message BODY, include a line containing: UNSUB ORACLE-L
> > > (or the name of mailing list you want to be removed from). You may
> > > also send the HELP command for other information (like subscribing).
> > >
> >
> >
> >--
> >Please see the official ORACLE-L FAQ: http://www.orafaq.net
> >--
> >Author: Tanel Poder
> > INET: tanel.poder.003_at_mail.ee
> >
> >Fat City Network Services -- 858-538-5051 http://www.fatcity.com
> >San Diego, California -- Mailing list and web hosting services
> >---------------------------------------------------------------------
> >To REMOVE yourself from this mailing list, send an E-Mail message
> >to: ListGuru_at_fatcity.com (note EXACT spelling of 'ListGuru') and in
> >the message BODY, include a line containing: UNSUB ORACLE-L
> >(or the name of mailing list you want to be removed from). You may
> >also send the HELP command for other information (like subscribing).
> >
> >
> >
> >--
> >Please see the official ORACLE-L FAQ: http://www.orafaq.net
> >--
> >Author: Michael Boligan
> > INET: michael.boligan.b_at_bayer.com
> >
> >Fat City Network Services -- 858-538-5051 http://www.fatcity.com
> >San Diego, California -- Mailing list and web hosting services
> >---------------------------------------------------------------------
> >To REMOVE yourself from this mailing list, send an E-Mail message
> >to: ListGuru_at_fatcity.com (note EXACT spelling of 'ListGuru') and in
> >the message BODY, include a line containing: UNSUB ORACLE-L
> >(or the name of mailing list you want to be removed from). You may
> >also send the HELP command for other information (like subscribing).
>
> Hemant K Chitale
> Oracle 9i Database Administrator Certified Professional
> My personal web site is : http://hkchital.tripod.com
>
>
> --
> Please see the official ORACLE-L FAQ: http://www.orafaq.net
> --
> Author: Hemant K Chitale
> INET: hkchital_at_singnet.com.sg
>
> Fat City Network Services -- 858-538-5051 http://www.fatcity.com
> San Diego, California -- Mailing list and web hosting services
> ---------------------------------------------------------------------
> To REMOVE yourself from this mailing list, send an E-Mail message
Received on Fri Jul 25 2003 - 09:54:09 CDT
 |
 |