Falla la restauración, se cuelga, tiempos de espera en ALTER TABLE

Tuve una falla de disco en un servidor de Discourse y estoy intentando restaurar una copia de seguridad desde S3 en una instancia AWS EC2 t2.small. Estoy obteniendo consistentemente el siguiente error.

Usando este flujo:

git clone https://github.com/discourse/discourse_docker /var/discourse

Estoy usando este app.yml

./launcher bootstrap app
./launcher start app

Luego hago scp de la copia de seguridad, ./launcher enter app, discourse enable_restore discourse restore <el archivo>.

Entonces ocurre esto, comenzando truncado:

14
(1 fila)
setval
--------
1
(1 fila)
setval
--------
1
(1 fila)
ALTER TABLE
ALTER TABLE
ALTER TABLE
ALTER TABLE
ALTER TABLE
ALTER TABLE
ALTER TABLE
ALTER TABLE
ALTER TABLE
ALTER TABLE
ALTER TABLE
ALTER TABLE
ALTER TABLE
ALTER TABLE
ALTER TABLE
ALTER TABLE
ALTER TABLE
ALTER TABLE
ALTER TABLE
ALTER TABLE
ALTER TABLE
ALTER TABLE
ALTER TABLE
ALTER TABLE
ALTER TABLE
ALTER TABLE
ALTER TABLE
ALTER TABLE
ALTER TABLE
ALTER TABLE
ALTER TABLE
ALTER TABLE
ALTER TABLE
ALTER TABLE
ALTER TABLE
ALTER TABLE
ALTER TABLE
ALTER TABLE
ALTER TABLE
ALTER TABLE
ALTER TABLE
ALTER TABLE
ALTER TABLE
ALTER TABLE
ALTER TABLE
ALTER TABLE
ALTER TABLE
ALTER TABLE
ALTER TABLE
ALTER TABLE
ALTER TABLE
ALTER TABLE
ALTER TABLE
ALTER TABLE
ALTER TABLE
ALTER TABLE
ALTER TABLE
ALTER TABLE
ALTER TABLE
ALTER TABLE
ALTER TABLE
ALTER TABLE
ALTER TABLE
ALTER TABLE
ALTER TABLE
ALTER TABLE
ALTER TABLE
ALTER TABLE
 #<Thread:0x0000562604925448 /var/www/discourse/lib/sidekiq/pausable.rb:79 run> terminated with exception (report_on_exception is true):
Traceback (most recent call last):
        33: from /var/www/discourse/lib/sidekiq/pausable.rb:83:in `block (2 levels) in extend_lease_thread'
        32: from /var/www/discourse/lib/sidekiq/pausable.rb:83:in `synchronize'
        31: from /var/www/discourse/lib/sidekiq/pausable.rb:84:in `block (3 levels) in extend_lease_thread'
        30: from /usr/local/lib/ruby/2.7.0/set.rb:328:in `each'
        29: from /usr/local/lib/ruby/2.7.0/set.rb:328:in `each_key'
        28: from /var/www/discourse/lib/sidekiq/pausable.rb:85:in `block (4 levels) in extend_lease_thread'
        27: from /var/www/discourse/vendor/bundle/ruby/2.7.0/gems/rails_multisite-2.5.0/lib/rails_multisite/connection_management.rb:76:in `with_connection'
        26: from /var/www/discourse/lib/sidekiq/pausable.rb:86:in `block (5 levels) in extend_lease_thread'
        25: from /var/www/discourse/lib/discourse_redis.rb:59:in `block (2 levels) in <class:DiscourseRedis>'
        24: from /var/www/discourse/lib/discourse_redis.rb:29:in `ignore_readonly'
        23: from /var/www/discourse/lib/discourse_redis.rb:59:in `block (3 levels) in <class:DiscourseRedis>'
        22: from /var/www/discourse/lib/discourse_redis.rb:59:in `public_send'
        21: from /var/www/discourse/vendor/bundle/ruby/2.7.0/gems/redis-4.2.5/lib/redis.rb:425:in `expire'
        20: from /var/www/discourse/vendor/bundle/ruby/2.7.0/gems/redis-4.2.5/lib/redis.rb:69:in `synchronize'
        19: from /usr/local/lib/ruby/2.7.0/monitor.rb:202:in `mon_synchronize'
        18: from /usr/local/lib/ruby/2.7.0/monitor.rb:202:in `synchronize'
        17: from /var/www/discourse/vendor/bundle/ruby/2.7.0/gems/redis-4.2.5/lib/redis.rb:69:in `block in synchronize'
        16: from /var/www/discourse/vendor/bundle/ruby/2.7.0/gems/redis-4.2.5/lib/redis.rb:426:in `block in expire'
        15: from /var/www/discourse/vendor/bundle/ruby/2.7.0/gems/rack-mini-profiler-2.3.1/lib/mini_profiler/profiling_methods.rb:85:in `block in profile_method'
        14: from /var/www/discourse/vendor/bundle/ruby/2.7.0/gems/redis-4.2.5/lib/redis/client.rb:131:in `call'
        13: from /var/www/discourse/vendor/bundle/ruby/2.7.0/gems/redis-4.2.5/lib/redis/client.rb:237:in `process'
        12: from /var/www/discourse/vendor/bundle/ruby/2.7.0/gems/redis-4.2.5/lib/redis/client.rb:325:in `logging'
        11: from /var/www/discourse/vendor/bundle/ruby/2.7.0/gems/redis-4.2.5/lib/redis/client.rb:238:in `block in process'
        10: from /var/www/discourse/vendor/bundle/ruby/2.7.0/gems/redis-4.2.5/lib/redis/client.rb:389:in `ensure_connected'
         9: from /var/www/discourse/vendor/bundle/ruby/2.7.0/gems/redis-4.2.5/lib/redis/client.rb:248:in `block (2 levels) in process'
         8: from /var/www/discourse/vendor/bundle/ruby/2.7.0/gems/redis-4.2.5/lib/redis/client.rb:131:in `block in call'
         7: from /var/www/discourse/vendor/bundle/ruby/2.7.0/gems/redis-4.2.5/lib/redis/client.rb:279:in `read'
         6: from /var/www/discourse/vendor/bundle/ruby/2.7.0/gems/redis-4.2.5/lib/redis/client.rb:268:in `io'
         5: from /var/www/discourse/vendor/bundle/ruby/2.7.0/gems/redis-4.2.5/lib/redis/client.rb:280:in `block in read'
         4: from /var/www/discourse/vendor/bundle/ruby/2.7.0/gems/redis-4.2.5/lib/redis/connection/ruby.rb:378:in `read'
         3: from /var/www/discourse/vendor/bundle/ruby/2.7.0/gems/redis-4.2.5/lib/redis/connection/ruby.rb:45:in `gets'
         2: from /var/www/discourse/vendor/bundle/ruby/2.7.0/gems/redis-4.2.5/lib/redis/connection/ruby.rb:52:in `_read_from_socket'
         1: from /var/www/discourse/vendor/bundle/ruby/2.7.0/gems/redis-4.2.5/lib/redis/connection/ruby.rb:52:in `loop'
/var/www/discourse/vendor/bundle/ruby/2.7.0/gems/redis-4.2.5/lib/redis/connection/ruby.rb:56:in `block in _read_from_socket': Redis::TimeoutError (Redis::TimeoutError)
        33: from /var/www/discourse/lib/sidekiq/pausable.rb:83:in `block (2 levels) in extend_lease_thread'
        32: from /var/www/discourse/lib/sidekiq/pausable.rb:83:in `synchronize'
        31: from /var/www/discourse/lib/sidekiq/pausable.rb:84:in `block (3 levels) in extend_lease_thread'
        30: from /usr/local/lib/ruby/2.7.0/set.rb:328:in `each'
        29: from /usr/local/lib/ruby/2.7.0/set.rb:328:in `each_key'
        28: from /var/www/discourse/lib/sidekiq/pausable.rb:85:in `block (4 levels) in extend_lease_thread'
        27: from /var/www/discourse/vendor/bundle/ruby/2.7.0/gems/rails_multisite-2.5.0/lib/rails_multisite/connection_management.rb:76:in `with_connection'
        26: from /var/www/discourse/lib/sidekiq/pausable.rb:86:in `block (5 levels) in extend_lease_thread'
        25: from /var/www/discourse/lib/discourse_redis.rb:59:in `block (2 levels) in <class:DiscourseRedis>'
        24: from /var/www/discourse/lib/discourse_redis.rb:29:in `ignore_readonly'
        23: from /var/www/discourse/lib/discourse_redis.rb:59:in `block (3 levels) in <class:DiscourseRedis>'
        22: from /var/www/discourse/lib/discourse_redis.rb:59:in `public_send'
        21: from /var/www/discourse/vendor/bundle/ruby/2.7.0/gems/redis-4.2.5/lib/redis.rb:425:in `expire'
        20: from /var/www/discourse/vendor/bundle/ruby/2.7.0/gems/redis-4.2.5/lib/redis.rb:69:in `synchronize'
        19: from /usr/local/lib/ruby/2.7.0/monitor.rb:202:in `mon_synchronize'
        18: from /usr/local/lib/ruby/2.7.0/monitor.rb:202:in `synchronize'
        17: from /var/www/discourse/vendor/bundle/ruby/2.7.0/gems/redis-4.2.5/lib/redis.rb:69:in `block in synchronize'
        16: from /var/www/discourse/vendor/bundle/ruby/2.7.0/gems/redis-4.2.5/lib/redis.rb:426:in `block in expire'
        15: from /var/www/discourse/vendor/bundle/ruby/2.7.0/gems/rack-mini-profiler-2.3.1/lib/mini_profiler/profiling_methods.rb:85:in `block in profile_method'
        14: from /var/www/discourse/vendor/bundle/ruby/2.7.0/gems/redis-4.2.5/lib/redis/client.rb:131:in `call'
        13: from /var/www/discourse/vendor/bundle/ruby/2.7.0/gems/redis-4.2.5/lib/redis/client.rb:237:in `process'
        12: from /var/www/discourse/vendor/bundle/ruby/2.7.0/gems/redis-4.2.5/lib/redis/client.rb:325:in `logging'
        11: from /var/www/discourse/vendor/bundle/ruby/2.7.0/gems/redis-4.2.5/lib/redis/client.rb:238:in `block in process'
        10: from /var/www/discourse/vendor/bundle/ruby/2.7.0/gems/redis-4.2.5/lib/redis/client.rb:389:in `ensure_connected'
         9: from /var/www/discourse/vendor/bundle/ruby/2.7.0/gems/redis-4.2.5/lib/redis/client.rb:248:in `block (2 levels) in process'
         8: from /var/www/discourse/vendor/bundle/ruby/2.7.0/gems/redis-4.2.5/lib/redis/client.rb:131:in `block in call'
         7: from /var/www/discourse/vendor/bundle/ruby/2.7.0/gems/redis-4.2.5/lib/redis/client.rb:279:in `read'
         6: from /var/www/discourse/vendor/bundle/ruby/2.7.0/gems/redis-4.2.5/lib/redis/client.rb:268:in `io'
         5: from /var/www/discourse/vendor/bundle/ruby/2.7.0/gems/redis-4.2.5/lib/redis/client.rb:280:in `block in read'
         4: from /var/www/discourse/vendor/bundle/ruby/2.7.0/gems/redis-4.2.5/lib/redis/connection/ruby.rb:378:in `read'
         3: from /var/www/discourse/vendor/bundle/ruby/2.7.0/gems/redis-4.2.5/lib/redis/connection/ruby.rb:45:in `gets'
         2: from /var/www/discourse/vendor/bundle/ruby/2.7.0/gems/redis-4.2.5/lib/redis/connection/ruby.rb:52:in `_read_from_socket'
         1: from /var/www/discourse/vendor/bundle/ruby/2.7.0/gems/redis-4.2.5/lib/redis/connection/ruby.rb:52:in `loop'
/var/www/discourse/vendor/bundle/ruby/2.7.0/gems/redis-4.2.5/lib/redis/connection/ruby.rb:56:in `block in _read_from_socket': Connection timed out (Redis::TimeoutError)
^[[A  ^[[A^[[B       server closed the connection unexpectedly
This probably means the server terminated abnormally
before or while processing the request.
connection to server was lost
EXCEPTION: psql failed: connection to server was lost

Consistentemente, el proceso llega hasta este punto en el proceso de restauración y luego se queda colgado en el último ALTER TABLE (el 85.º). Extraí el dump.sql de mi copia de seguridad y el 85.º es, si estoy contando bien, este de «public.reviewables»:

¿Alguna idea de lo que podría estar saliendo mal según lo anterior?

Creo que lo he resuelto. Debería haber ejecutado ./discourse-setup una vez en el nuevo host para (1) crear un archivo de intercambio y (2) establecer un valor razonable para db_shared_buffers. Uno de esos dos, o quizás ambos, eran incorrectos, lo que causaba el bloqueo constante.

¡Buen hallazgo! ¡Bienvenido de nuevo y bienvenido al club de “responde a tu propia pregunta”! :wink: Tengo un montón de esos temas, aunque estoy seguro de que muchos han sido eliminados.

La otra opción sería copiar el archivo app.yml, pero aún habrías tenido que recordar el intercambio. Excepto por el problema del intercambio, el archivo yml es todo lo que necesitas si la configuración de S3 está en el yml.

¡Me alegra que hayas vuelto a estar operativo!