Re: Having trouble configuring a Master with multiple standby Servers in PostgreSQL 9.3.3 - Mailing list pgsql-bugs

<span style=3D"font-family:Verdana; color:#000000; font-size:10=
pt;">Sorry folks, we fixed the problem, turned out that in the recover=
y.conf file I had primary_conninfo " ... sslmode=3Drequire" This caused the=
 error "could not connect to the primary server: sslmode value "require" in=
valid when SSL support is not compiled" So we just removed that option and =
bounced the slave and everything is working now.t=
hanks=0A<blockquote id=3D"replyBlockquote" webmail=3D"1" style=3D=
"border-left: 2px solid blue; margin-left: 8px; padding-left: 8px; font-siz=
e:10pt; color:black; font-family:verdana;">=0A=
=0A-------- Original Message --------=0ASubject: Re: [BUGS] Having trou=
ble configuring a Master with multiple=0Astandby Servers in PostgreSQL =
9.3.3=0AFrom: <fburgess@=
radiantblue.com>=0ADate: Fri, April 18, 2014 8:24 am=0ATo: "=
Michael Paquier" <michael.p=
aquier@gmail.com>=0ACc: <a href=3D"mailto:pgsql-bugs@postgresql.=
org">pgsql-bugs@postgresql.org=0A=0A<span style=3D"font-family:=
Verdana; color:#000000; font-size:10pt;">I started the recovery yester=
day and it ran overnight and is still running. Is the problem that the mast=
er is still producing new archivelogs that the slave is trying to recover, =
so that I am currently in a perpetual recovery mode?I can see th=
at the most recent archivelog being processed on the master is also being r=
ecovered on the slave. Do I need to suspend copying the archivelogs to the =
/mnt/server/slave1_archivedir/ directory, or should I wait?<=
/div>thanksFreddie<=
div> <blockquote id=3D"replyBlockquote" webmail=3D"1" style=3D"bo=
rder-left: 2px solid blue; margin-left: 8px; padding-left: 8px; font-size:1=
0pt; color:black; font-family:verdana;">  ------=
-- Original Message -------- Subject: Re: [BUGS] Having trouble configu=
ring a Master with multiple standby Servers in PostgreSQL 9.3.3 Fro=
m: Michael Paquier <<a target=3D"_blank" href=3D"mailto:michael.paquier@=
gmail.com">michael.paquier@gmail.com> Date: Thu, April 17, 2014 =
5:28 pm To: <a target=3D"_blank" href=3D"mailto:fburgess@radiantblue.co=
m">fburgess@radiantblue.com Cc: <a target=3D"_blank" href=3D"mailto=
:pgsql-bugs@postgresql.org">pgsql-bugs@postgresql.org  On Fri, =
Apr 18, 2014 at 1:19 AM, <<a target=3D"_blank" href=3D"mailto:fburgess@r=
adiantblue.com">fburgess@radiantblue.com> wrote: > Hi Michael=
, thanks for your reply. > > I discussed this my colleague, a=
nd we decided to change the archive_command > to execute a shell scr=
ipt. That's wiser as it allows more flexibility.  > This wil=
l copy the archivelogs from the master to both slaves. Will that > a=
void the issue with removing needed WAL files? > slave 1 > ar=
chive_cleanup_command =3D 'pg_archivecleanup /mnt/server/slave1_archivedir/=
 > %r' > slaves #2 > archive_cleanup_command =3D 'pg_a=
rchivecleanup /mnt/server/slave2_archivedir/ > %r' > Does thi=
s look correct? Looks fine. You are copying each WAL file to a differen=
t archive folder, and pg_archivecleanup will clean only the path it use=
s for each folder, so there is no risk to have a WAL file removed by on=
e slave and needed by the other.  > I did a pg_clt reload to=
 change the archivelog destination from > /mnt/server/master_archive=
dir to be redistributed to slave1 and slave2. Do I > need to redo th=
is backup step? Not if the slaves have already fetched necessary WAL fi=
les from the single master archive before you changed the command. =
 > psql -c "select pg_start_backup('initial_backup');" > rsyn=
c -cvar --inplace --exclude=3D*pg_xlog* > /u01/fiber/postgreSQL_data=
/postgres@1.2.3.5:/u01/fiber/postgreSQL_data/ > psql -c " select pg_=
stop_backup ();" > > or can I just copy all of the missing ar=
chivelog files from the > /mnt/server/master_archivedir to the slave=
s, and then restart the slaves in > recovery mode? Taking a new =
base backup will be fine. But you actually do not need to do so if your=
 slaves have already caught up enough. Your slaves are using streaming =
replication and are on the same server as the master AFAIU so they shou=
ld be fine, but there is always a possibility that they need some WAL f=
rom archives if one of them for example was not able to connect to the =
master for a long time and master already dropped the necessary WAL fil=
es from its pg_xlog. --  Michael   --  Sent via pgs=
ql-bugs mailing list (<a target=3D"_blank" href=3D"mailto:pgsql-bugs@postgr=
esql.org">pgsql-bugs@postgresql.org) To make changes to your subscr=
iption: <a target=3D"_blank" href=3D"http://www.postgresql.org/mailpref=
/pgsql-bugs">http://www.postgresql.org/mailpref/pgsql-bugs  <=
/blockquote> =0A=0A