Обсуждение: [postgres no lnger launches
Hi, how do I troubleshoot postgres? It simply won't start any more. I cannot login, not locally nor remotely. I tried a coupleof things without success. bash-3.2$ ps -Ac | egrep postgres bash-3.2$ /Library/PostgreSQL/9.1/bin/pg_ctl start -D /Library/PostgreSQL/9.1/data -l postgres.log pg_ctl: another server might be running; trying to start server anyway server starting bash-3.2$ /Library/PostgreSQL/9.1/bin/pg_ctl start -D /Library/PostgreSQL/9.1/data -l postgres.log pg_ctl: another server might be running; trying to start server anyway server starting bash-3.2$ /Library/PostgreSQL/9.1/bin/pg_ctl start -D /Library/PostgreSQL/9.1/data -l postgres.log pg_ctl: another server might be running; trying to start server anyway server starting bash-3.2$ ps -Ac | egrep postgres So sure enough that process claimed to might be starting does not end up being a process. Where do I look for what's actuallygoing on? This is the log output: 04.06.12 00:26:29,792 com.edb.launchd.postgresql-9.1: If the request size is already small, it's possible that it isless than your kernel's SHMMIN parameter, in which case raising the request size or reconfiguring SHMMIN is called for. 04.06.12 00:26:29,792 com.edb.launchd.postgresql-9.1: The PostgreSQL documentation contains more information about sharedmemory configuration. 04.06.12 00:26:29,793 com.apple.launchd: (com.edb.launchd.postgresql-9.1[1260]) Exited with code: 1 04.06.12 00:26:29,793 com.apple.launchd: (com.edb.launchd.postgresql-9.1) Throttling respawn: Will start in 10 seconds 04.06.12 00:26:30,000 kernel: nstat_lookup_entry failed: 2 It respawns every ten seconds and fails. Now this installation has been running for a few weeks now without a hiccup. I triedto enable remote login, I might have screwed up the pg_hba.conf file. So I copied the same file from an installationof another server I have on which things are running ok. I restarted but it continues to throw up those respawnmessages and shared mem segmets stuff. I changed the sysctl.conf file, but this neither made a difference. This unit has loads of memory....actually the messagesseem to be rather bogus. Also, it worked fine, I didn't change the sysctl file before, it seems silly to assume thatit just changes itself out of the blue and stops working. I installed nothing, there is no source I can assume to havedone so. What can I do to make this server working again? Thanks Alex
Btw, this is OSX 10.7. Am 04.06.2012 um 00:39 schrieb Alexander Reichstadt: > Hi, > > how do I troubleshoot postgres? It simply won't start any more. I cannot login, not locally nor remotely. I tried a coupleof things without success. > > bash-3.2$ ps -Ac | egrep postgres > bash-3.2$ /Library/PostgreSQL/9.1/bin/pg_ctl start -D /Library/PostgreSQL/9.1/data -l postgres.log > pg_ctl: another server might be running; trying to start server anyway > server starting > bash-3.2$ /Library/PostgreSQL/9.1/bin/pg_ctl start -D /Library/PostgreSQL/9.1/data -l postgres.log > pg_ctl: another server might be running; trying to start server anyway > server starting > bash-3.2$ /Library/PostgreSQL/9.1/bin/pg_ctl start -D /Library/PostgreSQL/9.1/data -l postgres.log > pg_ctl: another server might be running; trying to start server anyway > server starting > bash-3.2$ ps -Ac | egrep postgres > > So sure enough that process claimed to might be starting does not end up being a process. Where do I look for what's actuallygoing on? > > This is the log output: > 04.06.12 00:26:29,792 com.edb.launchd.postgresql-9.1: If the request size is already small, it's possible that it isless than your kernel's SHMMIN parameter, in which case raising the request size or reconfiguring SHMMIN is called for. > 04.06.12 00:26:29,792 com.edb.launchd.postgresql-9.1: The PostgreSQL documentation contains more information aboutshared memory configuration. > 04.06.12 00:26:29,793 com.apple.launchd: (com.edb.launchd.postgresql-9.1[1260]) Exited with code: 1 > 04.06.12 00:26:29,793 com.apple.launchd: (com.edb.launchd.postgresql-9.1) Throttling respawn: Will start in 10 seconds > 04.06.12 00:26:30,000 kernel: nstat_lookup_entry failed: 2 > > > It respawns every ten seconds and fails. Now this installation has been running for a few weeks now without a hiccup. Itried to enable remote login, I might have screwed up the pg_hba.conf file. So I copied the same file from an installationof another server I have on which things are running ok. I restarted but it continues to throw up those respawnmessages and shared mem segmets stuff. > > I changed the sysctl.conf file, but this neither made a difference. This unit has loads of memory....actually the messagesseem to be rather bogus. Also, it worked fine, I didn't change the sysctl file before, it seems silly to assume thatit just changes itself out of the blue and stops working. I installed nothing, there is no source I can assume to havedone so. What can I do to make this server working again? > > Thanks > Alex > > -- > Sent via pgsql-general mailing list (pgsql-general@postgresql.org) > To make changes to your subscription: > http://www.postgresql.org/mailpref/pgsql-general
Alexander Reichstadt <lxr@mac.com> writes:
>> So sure enough that process claimed to might be starting does not end up being a process. Where do I look for what's
actuallygoing on?
In the postmaster log ...
>> This is the log output:
>> 04.06.12 00:26:29,792 com.edb.launchd.postgresql-9.1: If the request size is already small, it's possible that
itis less than your kernel's SHMMIN parameter, in which case raising the request size or reconfiguring SHMMIN is called
for.
You seem to have trimmed off the useful part of this error message, but
I'm going to take a wild guess that what is happening is (a) you have
another postmaster running someplace, and (b) it is chewing up all the
allowed space for shared memory. If you have not mucked with OSX's
default shared-memory limits then there is only enough room to run one
postmaster at a time. The trouble with this theory is that the other
postmaster probably ought to be visible in ps output; I'm not sure why
it might not be. Anyway, try running "ipcs" to see if any SysV shared
memory or semaphores are active. If there are any, it's a pretty good
tip that there's a Postgres running somewhere.
regards, tom lane
I found the issue, I changed the connection amount. I changed it back now. Alex Am 04.06.2012 um 00:39 schrieb Alexander Reichstadt: > Hi, > > how do I troubleshoot postgres? It simply won't start any more. I cannot login, not locally nor remotely. I tried a coupleof things without success. > > bash-3.2$ ps -Ac | egrep postgres > bash-3.2$ /Library/PostgreSQL/9.1/bin/pg_ctl start -D /Library/PostgreSQL/9.1/data -l postgres.log > pg_ctl: another server might be running; trying to start server anyway > server starting > bash-3.2$ /Library/PostgreSQL/9.1/bin/pg_ctl start -D /Library/PostgreSQL/9.1/data -l postgres.log > pg_ctl: another server might be running; trying to start server anyway > server starting > bash-3.2$ /Library/PostgreSQL/9.1/bin/pg_ctl start -D /Library/PostgreSQL/9.1/data -l postgres.log > pg_ctl: another server might be running; trying to start server anyway > server starting > bash-3.2$ ps -Ac | egrep postgres > > So sure enough that process claimed to might be starting does not end up being a process. Where do I look for what's actuallygoing on? > > This is the log output: > 04.06.12 00:26:29,792 com.edb.launchd.postgresql-9.1: If the request size is already small, it's possible that it isless than your kernel's SHMMIN parameter, in which case raising the request size or reconfiguring SHMMIN is called for. > 04.06.12 00:26:29,792 com.edb.launchd.postgresql-9.1: The PostgreSQL documentation contains more information aboutshared memory configuration. > 04.06.12 00:26:29,793 com.apple.launchd: (com.edb.launchd.postgresql-9.1[1260]) Exited with code: 1 > 04.06.12 00:26:29,793 com.apple.launchd: (com.edb.launchd.postgresql-9.1) Throttling respawn: Will start in 10 seconds > 04.06.12 00:26:30,000 kernel: nstat_lookup_entry failed: 2 > > > It respawns every ten seconds and fails. Now this installation has been running for a few weeks now without a hiccup. Itried to enable remote login, I might have screwed up the pg_hba.conf file. So I copied the same file from an installationof another server I have on which things are running ok. I restarted but it continues to throw up those respawnmessages and shared mem segmets stuff. > > I changed the sysctl.conf file, but this neither made a difference. This unit has loads of memory....actually the messagesseem to be rather bogus. Also, it worked fine, I didn't change the sysctl file before, it seems silly to assume thatit just changes itself out of the blue and stops working. I installed nothing, there is no source I can assume to havedone so. What can I do to make this server working again? > > Thanks > Alex > > -- > Sent via pgsql-general mailing list (pgsql-general@postgresql.org) > To make changes to your subscription: > http://www.postgresql.org/mailpref/pgsql-general
Thanks, Tom, I messed it up myself, see my other post. Am 04.06.2012 um 01:16 schrieb Tom Lane: > Alexander Reichstadt <lxr@mac.com> writes: >>> So sure enough that process claimed to might be starting does not end up being a process. Where do I look for what'sactually going on? > > In the postmaster log ... > >>> This is the log output: >>> 04.06.12 00:26:29,792 com.edb.launchd.postgresql-9.1: If the request size is already small, it's possible that itis less than your kernel's SHMMIN parameter, in which case raising the request size or reconfiguring SHMMIN is called for. > > You seem to have trimmed off the useful part of this error message, but > I'm going to take a wild guess that what is happening is (a) you have > another postmaster running someplace, and (b) it is chewing up all the > allowed space for shared memory. If you have not mucked with OSX's > default shared-memory limits then there is only enough room to run one > postmaster at a time. The trouble with this theory is that the other > postmaster probably ought to be visible in ps output; I'm not sure why > it might not be. Anyway, try running "ipcs" to see if any SysV shared > memory or semaphores are active. If there are any, it's a pretty good > tip that there's a Postgres running somewhere. > > regards, tom lane > > -- > Sent via pgsql-general mailing list (pgsql-general@postgresql.org) > To make changes to your subscription: > http://www.postgresql.org/mailpref/pgsql-general