Re: Lot of "listener-thread: err = -25580: oserr =
Hi Art, The SO don't have dropped connections... I increase the tcp_conn_request_max to 256 anyway .... Remember , 99% of conection are with IPCSTR. On rarely occasion have more than 3 connection with SOCTCP. Theoretically, this solution with TCP do not have any with conection over IPCSTR. right ? Once start this messages, don't stop more... The last time i restart the instance, around of 80 clients (IPCSTR) are try connect to database and the messages start instantly . Art S. Kagel (Oninit) wrote: > >You day "the solution doesn't work here" but did netstat -s report >dropped connection attempts? What does > >ndd -get /dev/tcp tcp_conn_request_max > >show is the current setting for tcp_conn_request_max? If it's now 256 >did you try to increase that to 512? > >How many concurrent user connections do you typically have at the time >these messages appear? > >Art S. Kagel >Oninit > Abra sua conta no Yahoo! Mail, o único sem limite de espaço para armazenamento! http://br.mail.yahoo.com/
Cesar Inacio Martins wrote: > Hi Art, > > The SO don't have dropped connections... I increase the tcp_conn_request_max > to 256 anyway .... > > Remember , 99% of conection are with IPCSTR. > On rarely occasion have more than 3 connection with SOCTCP. > > Theoretically, this solution with TCP do not have any with conection over > IPCSTR. right ? > > Once start this messages, don't stop more... > The last time i restart the instance, around of 80 clients (IPCSTR) are try > connect to database and the messages start instantly . > Probably proof, as I suspect but don't know for sure, IDS is using UNIX domain sockets to implement ipcstr connections and that goes through the TCP stack and libraries and so IS INDEED affected by the tcp_conn_request_max. Try setting that to 1024 or something like that and see what happens. Art S. Kagel Oninit > Art S. Kagel (Oninit) wrote: > >> You day "the solution doesn't work here" but did netstat -s report >> dropped connection attempts? What does >> >> ndd -get /dev/tcp tcp_conn_request_max >> >> show is the current setting for tcp_conn_request_max? If it's now 256 >> did you try to increase that to 512? >> >> How many concurrent user connections do you typically have at the time >> these messages appear? >> >> Art S. Kagel >> Oninit >> >> > > Abra sua conta no Yahoo! Mail, o único sem limite de espaço para > armazenamento! > http://br.mail.yahoo.com/ > > > ******************************************************************************* > Forum Note: Use "Reply" to post a response in the discussion forum. > > > ------------------------------------------------------------------------ > > See you at the IIUG Informix 2008 Conference > The Power Conference for Informix Professionals > April 27 - 30, 2008 Marriott Overland Park (Kansas City), Kansas > http://www.iiug.org/conf > Registration Now Open!!
Hi Art, Yesterday I set the tcp_conn_request_max to 1024 ... The errors continue to appear on the message log... This messages started from anywhere , the instance has more than 60 day online when started. I already restart the instance, the machine... always the messages back... ----- Mensagem original ---- De: Art S. Kagel (Oninit) <art@oninit.com> Para: ids@iiug.org Enviadas: Quarta-feira, 9 de Abril de 2008 19:22:22 Assunto: Re: Lot of "listener-thread: err = -25580: ose.... [11810] Cesar Inacio Martins wrote: > Hi Art, > > The SO don't have dropped connections... I increase the tcp_conn_request_max > to 256 anyway .... > > Remember , 99% of conection are with IPCSTR. > On rarely occasion have more than 3 connection with SOCTCP. > > Theoretically, this solution with TCP do not have any with conection over > IPCSTR. right ? > > Once start this messages, don't stop more... > The last time i restart the instance, around of 80 clients (IPCSTR) are try > connect to database and the messages start instantly . > Probably proof, as I suspect but don't know for sure, IDS is using UNIX domain sockets to implement ipcstr connections and that goes through the TCP stack and libraries and so IS INDEED affected by the tcp_conn_request_max. Try setting that to 1024 or something like that and see what happens. Art S. Kagel Oninit > Art S. Kagel (Oninit) wrote: > >> You day "the solution doesn't work here" but did netstat -s report >> dropped connection attempts? What does >> >> ndd -get /dev/tcp tcp_conn_request_max >> >> show is the current setting for tcp_conn_request_max? If it's now 256 >> did you try to increase that to 512? >> >> How many concurrent user connections do you typically have at the time >> these messages appear? >> >> Art S. Kagel >> Oninit >> >> > > Abra sua conta no Yahoo! Mail, o único sem limite de espaço para > armazenamento! > http://br.mail.yahoo.com/ > > > ******************************************************************************* > Forum Note: Use "Reply" to post a response in the discussion forum. > > > ------------------------------------------------------------------------ > > See you at the IIUG Informix 2008 Conference > The Power Conference for Informix Professionals > April 27 - 30, 2008 Marriott Overland Park (Kansas City), Kansas > http://www.iiug.org/conf > Registration Now Open!! ******************************************************************************* Forum Note: Use "Reply" to post a response in the discussion forum. Abra sua conta no Yahoo! Mail, o único sem limite de espaço para armazenamento! http://br.mail.yahoo.com/
Cesar Inacio Martins wrote: > Hi Art, > > Yesterday I set the tcp_conn_request_max to 1024 ... > The errors continue to appear on the message log... > > This messages started from anywhere , the instance has more than 60 day online > when started. > I already restart the instance, the machine... always the messages back... > Cesar, Well, then, the only thing left is to open a support case with IBM. At least you can tell them that you beat this particular workaround to death. Art S. Kagel Oninit > ----- Mensagem original ---- > De: Art S. Kagel (Oninit) <art@oninit.com> > Para: ids@iiug.org > Enviadas: Quarta-feira, 9 de Abril de 2008 19:22:22 > Assunto: Re: Lot of "listener-thread: err = -25580: ose.... [11810] > > Cesar Inacio Martins wrote: > >> Hi Art, >> >> The SO don't have dropped connections... I increase the tcp_conn_request_max >> to 256 anyway .... >> >> Remember , 99% of conection are with IPCSTR. >> On rarely occasion have more than 3 connection with SOCTCP. >> >> Theoretically, this solution with TCP do not have any with conection over >> IPCSTR. right ? >> >> Once start this messages, don't stop more... >> The last time i restart the instance, around of 80 clients (IPCSTR) are try >> connect to database and the messages start instantly . >> >> > > Probably proof, as I suspect but don't know for sure, IDS is using UNIX > domain sockets to implement ipcstr connections and that goes through the > TCP stack and libraries and so IS INDEED affected by the > tcp_conn_request_max. Try setting that to 1024 or something like that > and see what happens. > > Art S. Kagel > Oninit > > >> Art S. Kagel (Oninit) wrote: >> >> >>> You day "the solution doesn't work here" but did netstat -s report >>> dropped connection attempts? What does >>> >>> ndd -get /dev/tcp tcp_conn_request_max >>> >>> show is the current setting for tcp_conn_request_max? If it's now 256 >>> did you try to increase that to 512? >>> >>> How many concurrent user connections do you typically have at the time >>> these messages appear? >>> >>> Art S. Kagel >>> Oninit >>> >>> >>> >> Abra sua conta no Yahoo! Mail, o único sem limite de espaço para >> armazenamento! >> http://br.mail.yahoo.com/ >>