Re: Active Session History count(*) vs. sum(time_waited)
From: David Fitzjarrell <oratune_at_yahoo.com>
Date: Fri, 17 Jan 2014 13:47:28 -0800 (PST)
Message-ID: <1389995248.58215.YahooMailNeo_at_web121603.mail.ne1.yahoo.com>
No, it's waiting for more data from the client such as the next transaction from a user session. It isn't a wait to worry about, really. David Fitzjarrell Primary author, "Oracle Exadata Survival Guide" On Friday, January 17, 2014 2:42 PM, Dba DBA <oracledbaquestions_at_gmail.com> wrote: thank you. What does 'more data from client' mean? Most of the time the application is hung up we get an idle SQLNET wait. This is a query that returns less than 100 bytes, so it should fit in 1 network pass. Is oracle expecting an 'ack' saying i got it from the client? On Thu, Jan 16, 2014 at 4:11 PM, Austin Hackett <hacketta_57_at_me.com> wrote: Hi > >ASH has a "fix-up" mechanism. For long events the last sample gets it's TIME_WAITED value populated, and all the other samples stay at 0. > >If you are seeing sum(time_waited) = 0 then my assumption would be that you have an on-going long wait on SQL*Net more data from client. > >I've previously seen "SQL*Net more data from client" when a client process has died mid-stream. If DCD (dead connection detection) is disabled, the individual wait can go on for days. I think the SEQ# column would allow you to validate this - if the SEQ# is always the same, it should be the same wait. > >Hope that helps > >Austin > > > >-- >http://www.freelists.org/webpage/oracle-l > > >
Date: Fri, 17 Jan 2014 13:47:28 -0800 (PST)
Message-ID: <1389995248.58215.YahooMailNeo_at_web121603.mail.ne1.yahoo.com>
No, it's waiting for more data from the client such as the next transaction from a user session. It isn't a wait to worry about, really. David Fitzjarrell Primary author, "Oracle Exadata Survival Guide" On Friday, January 17, 2014 2:42 PM, Dba DBA <oracledbaquestions_at_gmail.com> wrote: thank you. What does 'more data from client' mean? Most of the time the application is hung up we get an idle SQLNET wait. This is a query that returns less than 100 bytes, so it should fit in 1 network pass. Is oracle expecting an 'ack' saying i got it from the client? On Thu, Jan 16, 2014 at 4:11 PM, Austin Hackett <hacketta_57_at_me.com> wrote: Hi > >ASH has a "fix-up" mechanism. For long events the last sample gets it's TIME_WAITED value populated, and all the other samples stay at 0. > >If you are seeing sum(time_waited) = 0 then my assumption would be that you have an on-going long wait on SQL*Net more data from client. > >I've previously seen "SQL*Net more data from client" when a client process has died mid-stream. If DCD (dead connection detection) is disabled, the individual wait can go on for days. I think the SEQ# column would allow you to validate this - if the SEQ# is always the same, it should be the same wait. > >Hope that helps > >Austin > > > >-- >http://www.freelists.org/webpage/oracle-l > > >
-- http://www.freelists.org/webpage/oracle-lReceived on Fri Jan 17 2014 - 22:47:28 CET