내용으로 건너뛰기

postgreSQL aurora serverless v1 upgrade from 11.21 to 13.12

0

Since postgreSQL 11.21 has reached EOL, we are trying to upgrade it to 13.12. Initially the upgrade was failing at the precheck stage due to postgis. Resolved that, the upgrade is proceeding past the precheck, but it's still failing. From the CloudWatch logs:

*failure*
There were problems executing ""/rdsdbbin/aurora-13.12.13.12.7.14976.0/bin/pg_restore" --host /rdsdbdata/log/upgrade --port 5433 --username **myuser** --create --exit-on-error --verbose --jobs 1 --dbname template1 "pg_upgrade_dump_16408.custom" >> "pg_upgrade_dump_16408.log" 2>&1"
Consult the last few lines of "pg_upgrade_dump_16408.log" for
the probable cause of the failure.
"/rdsdbbin/aurora-13.12.13.12.7.14976.0/bin/pg_ctl" -w -D "/rdsdbdata/db" -o "--config_file=/rdsdbdata/config_new/postgresql.conf --survivable_cache_mode=off" -m fast stop >> "pg_upgrade_server.log" 2>&1

Is there any way to Consult the last few lines of "pg_upgrade_dump_16408.log" for the probable cause of the failure.? Or any idea on how to solve the issue?

Also, I masked the --username **myuser** bit from the log, but not the --host /rdsdbdata/log/upgrade, that looks strange/wrong (or is it some automatic masking?)

1개 답변
0
수락된 답변

This seems resolved now,

not sure if ti was just a temporary issue that got resolved after a new attempt, like in https://repost.aws/questions/QUdN7XUmg8R9u-qtZNy_T-_w/access-log-file-from-failed-upgrade-from-postgresql-10-7-to-11-2#ANjDKCqn5qRqCssfxMEqzJaQ

or if it was still an issue related to postgis_raster extension. In fact after updating the postgis extension, the postgis_raster was still trying to use some old 2.* system libraries, and I followed steps reported in https://repost.aws/questions/QUP1Suk7DnSMqEZMCBkOGoBQ/rds-postgresql-engine-upgrade-from-10-21-to-14-4-version#ANpno0pqFPQXO9HtAgs4Tmqw

but maybe that was not enough, so I also tried https://postgis.net/stuff/uninstall_rtpostgis.sql before triggering again the cluster upgrade, that this time worked without errors.

We'll check how it goes, doing the same process in the other environments.

답변함 2년 전

전문가

검토됨 2년 전

  • This worked also with another cluster in another environment/account

로그인하지 않았습니다. 로그인해야 답변을 게시할 수 있습니다.

좋은 답변은 질문에 명확하게 답하고 건설적인 피드백을 제공하며 질문자의 전문적인 성장을 장려합니다.

관련 콘텐츠