Thread: Incomplete startup packet help needed
Immediately upon starting the server I get an "incomplete startup packet" log message. Just prior there is an "autovacuumlauncher started" message. Any thoughts? David J.
On 01/23/12 4:24 PM, David Johnston wrote: > Immediately upon starting the server I get an "incomplete startup packet" log message. Just prior there is an "autovacuumlauncher started" message. > > Any thoughts? based on the given information: something is probably wrong. -- john r pierce N 37, W 122 santa cruz ca mid-left coast
On Monday, January 23, 2012 4:24:50 pm David Johnston wrote: > Immediately upon starting the server I get an "incomplete startup packet" > log message. Just prior there is an "autovacuum launcher started" > message. Do you have other programs connecting to server on startup? > > Any thoughts? > > David J. -- Adrian Klaver adrian.klaver@gmail.com
On 1/23/2012 5:24 PM, David Johnston wrote: > Immediately upon starting the server I get an "incomplete startup packet" log message. Just prior there is an "autovacuumlauncher started" message. > We've found that this message is printed in the log if a client makes a TCP connection to the PG server, but sends no traffic. For example it happens with our monitoring system, which checks that it can open a TCP connection to the PG port, then closes it.
On Jan 23, 2012, at 19:38, Adrian Klaver <adrian.klaver@gmail.com> wrote: > On Monday, January 23, 2012 4:24:50 pm David Johnston wrote: >> Immediately upon starting the server I get an "incomplete startup packet" >> log message. Just prior there is an "autovacuum launcher started" >> message. > > Do you have other programs connecting to server on startup? > >> >> Any thoughts? >> >> David J. > > -- > Adrian Klaver > adrian.klaver@gmail.com > I have 3 external machines that hit this server plus Apache Tomcat on the same machine. I shutdown Tomcat during the restart. Possible a zombie (or hidden/forgotten) process is running though the log indicates local. Had been using pg_admin3for testing and somehow ended up with a possible zombie process there so maybe... Version 9.0.x - can get the release a little later, running pg_dump at the moment. Still researching/learning but was using WAL shipping and the ship command was failing. In the target directory the mostrecent WAL files is only ~2MB in size (can't check the pg_xlog directory at the moment). My guess is size quota on targetbut cannot check at the moment. Turned off WAL archive and restarted the server. Restarted a couple of times untilI remembered to shutdown all the other applications. Pg_dump seems to be running OK and was able to execute queries. Was still having difficulty getting my Tomcat software toconnect but am unsure if I'm just missing something or what. No obvious messages in either log about that but cannot getback to it until the dump completes. Appreciate any help that can be given. Unfortunately I haven't focused enough attention on Linux (Ubuntu 10.4 LTE) and PostgreSQLadministration as I'd like (education and/or professional services). Apparently that is going to change soonerthan I had thought. David J.
GOOD NEWS On Jan 23, 2012, at 20:00, David Johnston <polobo@yahoo.com> wrote: > On Jan 23, 2012, at 19:38, Adrian Klaver <adrian.klaver@gmail.com> wrote: > >> On Monday, January 23, 2012 4:24:50 pm David Johnston wrote: >>> Immediately upon starting the server I get an "incomplete startup packet" >>> log message. Just prior there is an "autovacuum launcher started" >>> message. >> >> Do you have other programs connecting to server on startup? >> >>> >>> Any thoughts? >>> >>> David J. >> >> -- >> Adrian Klaver >> adrian.klaver@gmail.com >> > > I have 3 external machines that hit this server plus Apache Tomcat on the same machine. I shutdown Tomcat during the restart. Possible a zombie (or hidden/forgotten) process is running though the log indicates local. Had been using pg_admin3for testing and somehow ended up with a possible zombie process there so maybe... > > Version 9.0.x - can get the release a little later, running pg_dump at the moment. > > Still researching/learning but was using WAL shipping and the ship command was failing. In the target directory the mostrecent WAL files is only ~2MB in size (can't check the pg_xlog directory at the moment). My guess is size quota on targetbut cannot check at the moment. Turned off WAL archive and restarted the server. Restarted a couple of times untilI remembered to shutdown all the other applications. > > Pg_dump seems to be running OK and was able to execute queries. Was still having difficulty getting my Tomcat softwareto connect but am unsure if I'm just missing something or what. No obvious messages in either log about that butcannot get back to it until the dump completes. > > Appreciate any help that can be given. Unfortunately I haven't focused enough attention on Linux (Ubuntu 10.4 LTE) andPostgreSQL administration as I'd like (education and/or professional services). Apparently that is going to change soonerthan I had thought. > > David J. > Looking at my process viewer I saw a Java process running (with Catalina...) even though I supposedly shutdown Tomcat. SinceTomcat wasn't working anyway I decided to just kill the process then restart PostgreSQL...no more spurious error message. Restarted Tomcat and I can now hit my Servlet. Just need to get the base backup and WAL shipping enabled again. Assuming since I disabled it, and just ran a pg_dump whileeverything was disconnected, that I can basically throw out the shipped WAL files? Still kinda at a loss for root cause and deciding whether there still may be an underlying problem. Thoughts are welcomebut I'm likely to seek professional help on this one regardless. Thanks for listening. David J.
On Monday, January 23, 2012 5:00:17 pm David Johnston wrote: > On Jan 23, 2012, at 19:38, Adrian Klaver <adrian.klaver@gmail.com> wrote: > > On Monday, January 23, 2012 4:24:50 pm David Johnston wrote: > >> Immediately upon starting the server I get an "incomplete startup > >> packet" log message. Just prior there is an "autovacuum launcher > >> started" message. > > > > Do you have other programs connecting to server on startup? > > > >> Any thoughts? > >> > >> David J. > > I have 3 external machines that hit this server plus Apache Tomcat on the > same machine. I shutdown Tomcat during the restart. Possible a zombie > (or hidden/forgotten) process is running though the log indicates local. > Had been using pg_admin3 for testing and somehow ended up with a possible > zombie process there so maybe... > > Version 9.0.x - can get the release a little later, running pg_dump at the > moment. > > Still researching/learning but was using WAL shipping and the ship command > was failing. In the target directory the most recent WAL files is only > ~2MB in size (can't check the pg_xlog directory at the moment). My guess > is size quota on target but cannot check at the moment. Turned off WAL > archive and restarted the server. Restarted a couple of times until I > remembered to shutdown all the other applications. > > Pg_dump seems to be running OK and was able to execute queries. Was still > having difficulty getting my Tomcat software to connect but am unsure if > I'm just missing something or what. No obvious messages in either log > about that but cannot get back to it until the dump completes. > > Appreciate any help that can be given. Unfortunately I haven't focused > enough attention on Linux (Ubuntu 10.4 LTE) and PostgreSQL administration > as I'd like (education and/or professional services). Apparently that is > going to change sooner than I had thought. Alright alot going on. Searching for the error message tends to indicate what others have said, there is something basically pinging the server with 'empty' packets. Given everything you going on right now I would probably wait until the server has attained a more quiescent state before tracking down the culprit. At guess it not something that is server threatening, but worthy of attention. Once the dump is complete and if it is possible, shut down the server. Go through and make sure the programs hitting the server are shut down, for now at least that probably should include the WAL archiving. Start the server up, tail -f the log and then start up the client programs one at a time and see if any of them tickle the error. > > David J. -- Adrian Klaver adrian.klaver@gmail.com
* David Johnston: > Immediately upon starting the server I get an "incomplete startup > packet" log message. Just prior there is an "autovacuum launcher > started" message. Like this? 2012-01-23 10:42:55.245 UTC 11545 LOG: database system is ready to accept connections 2012-01-23 10:42:55.245 UTC 11549 LOG: autovacuum launcher started 2012-01-23 10:42:55.268 UTC 11551 [unknown] [unknown] LOG: incomplete startup packet I think it's harmless, it's been there for years. It might be related to the init script that starts the database server. -- Florian Weimer <fweimer@bfk.de> BFK edv-consulting GmbH http://www.bfk.de/ Kriegsstraße 100 tel: +49-721-96201-1 D-76133 Karlsruhe fax: +49-721-96201-99
We have it too. I think it might be samba related, or just some SNMP software running. I had indentifed the root of the situation some years ago, do not remember it now. It is 100% harmless. On Τρι 24 Ιαν 2012 10:26:25 Florian Weimer wrote: > * David Johnston: > > Immediately upon starting the server I get an "incomplete startup > > packet" log message. Just prior there is an "autovacuum launcher > > started" message. > > Like this? > > 2012-01-23 10:42:55.245 UTC 11545 LOG: database system is ready to > accept connections 2012-01-23 10:42:55.245 UTC 11549 LOG: autovacuum > launcher started 2012-01-23 10:42:55.268 UTC 11551 [unknown] [unknown] > LOG: incomplete startup packet > > I think it's harmless, it's been there for years. It might be related > to the init script that starts the database server. -- Achilleas Mantzios IT DEPT
You can force this message executing: $ echo | nc SERVER_IP 5432 2014-11-04 21:44:37 CET LOG: incomplete startup packet ----- http://www.administracion-linux.com -- View this message in context: http://postgresql.1045698.n5.nabble.com/Incomplete-startup-packet-help-needed-tp5199030p5825683.html Sent from the PostgreSQL - general mailing list archive at Nabble.com.
Hi Juan Antonio:
On Tue, Nov 4, 2014 at 9:54 PM, memtec <juanantonio.izquierdo@gmail.com> wrote:
You can force this message executing:
$ echo | nc SERVER_IP 5432
2014-11-04 21:44:37 CET LOG: incomplete startup packet
What do you need help on? PG uses a binary protocol, you sent it bad data ( I think this will open connection, send a \n byte, shutdown it ) and it complained in the first check ( I suppose it is due to the connection expecting an startup packet with a given format, and it couldn't get it because it's longer than 1 byte, everything is prety well documented, RTFM ).
Regards.
Francisco Olarte.
On Thu, 6 Nov 2014 13:40:14 +0100
Francisco Olarte <folarte@peoplecall.com> wrote:
> Hi Juan Antonio:
>
> On Tue, Nov 4, 2014 at 9:54 PM, memtec <juanantonio.izquierdo@gmail.com>
> wrote:
>
> > You can force this message executing:
> > $ echo | nc SERVER_IP 5432
> > 2014-11-04 21:44:37 CET LOG: incomplete startup packet
> >
>
> What do you need help on? PG uses a binary protocol, you sent it bad data
> ( I think this will open connection, send a \n byte, shutdown it ) and it
> complained in the first check ( I suppose it is due to the connection
> expecting an startup packet with a given format, and it couldn't get it
> because it's longer than 1 byte, everything is prety well documented, RTFM
> ).
FYI: It's very common to get these messages in your log file if you have something like nagios or other monitoring
toolschecking to see if postgres is up. This is because many of them just test to see if something is listening on the
port,without opening a proper postgres connection (i.e. telnet localhost 5432). The just connect to the port then
disconnectwithout sending any data of any kind.
--
Jon Erdman (aka StuckMojo)
PostgreSQL Zealot