| 1 |
/home/production/cvs/JSOC/doc/dcs_warmstandby.txt |
| 2 |
|
| 3 |
NOTICE: This isn't blind step-by-step procedures. You must understand |
| 4 |
what is being done so that you can interpret the directions for what |
| 5 |
exactly you want to do! |
| 6 |
|
| 7 |
$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$ |
| 8 |
|
| 9 |
!!!BEGIN: To set up warm-standby for dcs0 on dcs2:: |
| 10 |
Assumes no DB for dcs0 running on dcs2. If so, |
| 11 |
pg_ctl -D /var/lib/pgsql/dcs0_data/ stop -m fast |
| 12 |
Make sure no /tmp/pgsql.trigger.5430 on dcs2 |
| 13 |
|
| 14 |
Login to dcs0 as postgres |
| 15 |
> ls -1 / | grep WALarchive | wc -l |
| 16 |
1 |
| 17 |
|
| 18 |
NOTE: the original dcs0 will return 1. Orig dcs2 converted to dcs0 returns 0. |
| 19 |
|
| 20 |
If (0) then |
| 21 |
cp configfiles/postgresql.conf.dcs0.0 dcs0_data/postgresql.conf |
| 22 |
rm -f /data/WALarchive/5430/* |
| 23 |
|
| 24 |
if (1) then |
| 25 |
cp configfiles/postgresql.conf.dcs0.1 dcs0_data/postgresql.conf |
| 26 |
rm -f /WALarchive/5430/* |
| 27 |
|
| 28 |
Restart the dcs0 database to create WAL archives: |
| 29 |
> pg_ctl -D /var/lib/pgsql/dcs0_data -m fast stop |
| 30 |
> pg_ctl -D /var/lib/pgsql/dcs0_data/ start |
| 31 |
|
| 32 |
Begin the backup process on dcs0: |
| 33 |
> psql -p 5430 -c "select pg_start_backup('2010-03-15')" |
| 34 |
|
| 35 |
Create a tar archive of the system data directory: |
| 36 |
> tar -czvf dcs0_backup.tar.gz dcs0_data/ |
| 37 |
|
| 38 |
Stop the backup process on dcs0: |
| 39 |
> psql -p 5430 -c "select pg_stop_backup()" |
| 40 |
|
| 41 |
Copy the archive file from dcs0 to dcs2: |
| 42 |
> scp dcs0_backup.tar.gz postgres@dcs2:/var/lib/pgsql/ |
| 43 |
|
| 44 |
Now login to dcs2 as postgres: |
| 45 |
> rm -rf dcs0_data |
| 46 |
> rm -rf dcs1_data (NOT if just doing warm-standby for dcs0??) |
| 47 |
|
| 48 |
Uncompress the archive file: |
| 49 |
> tar -xzvf dcs0_backup.tar.gz |
| 50 |
|
| 51 |
> ls -1 / | grep WALarchive | wc -l |
| 52 |
0 |
| 53 |
|
| 54 |
if(0) then |
| 55 |
cp configfiles/recovery.conf.dcs0.0 dcs0_data/recovery.conf |
| 56 |
cp configfiles/postgresql.conf.dcs0.ws dcs0_data/postgresql.conf |
| 57 |
if(1) then |
| 58 |
cp configfiles/recovery.conf.dcs0.1 dcs0_data/recovery.conf |
| 59 |
cp configfiles/postgresql.conf.dcs0.ws dcs0_data/postgresql.conf |
| 60 |
|
| 61 |
Remove the postmaster.pid file: |
| 62 |
> rm dcs0_data/postmaster.pid |
| 63 |
|
| 64 |
> chmod 700 /var/lib/pgsql/dcs0_data/ |
| 65 |
|
| 66 |
Start the dcs0 warm standby server, execute the command: |
| 67 |
> pg_ctl -D /var/lib/pgsql/dcs0_data/ start |
| 68 |
|
| 69 |
The server should go into recovery mode and stay there. |
| 70 |
After everything looks good, remove the tar file from dcs2: |
| 71 |
> rm dcs0_backup.tar.gz |
| 72 |
!!!END: To set up warm-standby for dcs0 on dcs2:: |
| 73 |
========================================================================== |
| 74 |
|
| 75 |
|
| 76 |
!!!BEGIN: To set up warm-standby for dcs1 on dcs2:: |
| 77 |
Assumes no DB for dcs1 running on dcs2. If so, |
| 78 |
pg_ctl -D /var/lib/pgsql/dcs1_data/ stop -m fast |
| 79 |
Make sure no /tmp/pgsql.trigger.5431 on dcs2 |
| 80 |
|
| 81 |
Login to dcs1 as postgres |
| 82 |
> ls -1 / | grep WALarchive | wc -l |
| 83 |
1 |
| 84 |
|
| 85 |
NOTE: the original dcs0 will return 1. Orig dcs2 converted to dcs0 returns 0. |
| 86 |
|
| 87 |
If (0) then |
| 88 |
cp configfiles/postgresql.conf.dcs1.0 dcs1_data/postgresql.conf |
| 89 |
rm -f /data/WALarchive/5431/* |
| 90 |
|
| 91 |
if (1) then |
| 92 |
cp configfiles/postgresql.conf.dcs1.1 dcs1_data/postgresql.conf |
| 93 |
rm -f /WALarchive/5431/* |
| 94 |
|
| 95 |
Restart the dcs1 database to create WAL archives: |
| 96 |
> pg_ctl -D /var/lib/pgsql/dcs1_data -m fast stop |
| 97 |
> pg_ctl -D /var/lib/pgsql/dcs1_data/ start |
| 98 |
|
| 99 |
Begin the backup process on dcs1: |
| 100 |
> psql -p 5431 -c "select pg_start_backup('2010-03-15')" |
| 101 |
|
| 102 |
Create a tar archive of the system data directory: |
| 103 |
> tar -czvf dcs1_backup.tar.gz dcs1_data/ |
| 104 |
|
| 105 |
Stop the backup process on dcs0: |
| 106 |
> psql -p 5431 -c "select pg_stop_backup()" |
| 107 |
|
| 108 |
Copy the archive file from dcs1 to dcs2: |
| 109 |
> scp dcs1_backup.tar.gz postgres@dcs2:/var/lib/pgsql/ |
| 110 |
|
| 111 |
Now login to dcs2 as postgres: |
| 112 |
> rm -rf dcs1_data |
| 113 |
|
| 114 |
Uncompress the archive file: |
| 115 |
> tar -xzvf dcs1_backup.tar.gz |
| 116 |
|
| 117 |
> ls -1 / | grep WALarchive | wc -l |
| 118 |
0 |
| 119 |
|
| 120 |
if(0) then |
| 121 |
cp configfiles/recovery.conf.dcs1.0 dcs1_data/recovery.conf |
| 122 |
cp configfiles/postgresql.conf.dcs1.ws dcs1_data/postgresql.conf |
| 123 |
if(1) then |
| 124 |
cp configfiles/recovery.conf.dcs1.1 dcs1_data/recovery.conf |
| 125 |
cp configfiles/postgresql.conf.dcs1.ws dcs1_data/postgresql.conf |
| 126 |
|
| 127 |
Remove the postmaster.pid file: |
| 128 |
> rm dcs1_data/postmaster.pid |
| 129 |
|
| 130 |
> chmod 700 /var/lib/pgsql/dcs1_data/ |
| 131 |
|
| 132 |
Start the dcs1 warm standby server, execute the command: |
| 133 |
> pg_ctl -D /var/lib/pgsql/dcs1_data/ start |
| 134 |
|
| 135 |
The server should go into recovery mode and stay there. |
| 136 |
After everything looks good, remove the tar file from dcs2: |
| 137 |
> rm dcs1_backup.tar.gz |
| 138 |
!!!END: To set up warm-standby for dcs1 on dcs2:: |
| 139 |
========================================================================== |
| 140 |
|
| 141 |
!!!BEGIN: dcs0 fails. Switch over to warm-standby on dcs2: |
| 142 |
Ensure that the database on the dcs0 server is down: |
| 143 |
> pg_ctl -D /var/lib/pgsql/dcs0_data -m fast stop |
| 144 |
|
| 145 |
Login to dcs2 as postgres. |
| 146 |
Make sure no dcs1 warm-standby is running: |
| 147 |
> pg_ctl -D /var/lib/pgsql/dcs1_data/ stop -m fast |
| 148 |
|
| 149 |
To bring the dcs0 warm standby out of recovery mode and make it live: |
| 150 |
NOTE: run psql -p 5430 -l before and after this to see the db come online |
| 151 |
> touch /tmp/pgsql.trigger.5430 |
| 152 |
|
| 153 |
!!TBD find out when to rm /tmp/pgsql.trigger.5430 |
| 154 |
|
| 155 |
Now make dcs2 the new dcs0. Run as root: |
| 156 |
(MAKE sure old dcs0 is down. Do not bring it up again with 2 dcs0's) |
| 157 |
> /etc/sysconfig/network/AIA-up |
| 158 |
|
| 159 |
To restore to dcs2 again: |
| 160 |
> /etc/sysconfig/network/MYSELF-up |
| 161 |
|
| 162 |
|
| 163 |
As user production on the new dcs0: |
| 164 |
> vncserver -geometry 1280x1024 |
| 165 |
On j0: |
| 166 |
j0:/home/production> vncviewer |
| 167 |
answer dcs0.jsoc.Stanford.EDU:1 |
| 168 |
|
| 169 |
Apply crontab -e from /home/production/crontab.dcs0 |
| 170 |
|
