cardano-db-sync icon indicating copy to clipboard operation
cardano-db-sync copied to clipboard

Restoring from snapshot error - role "cexplorer" does not exist

Open dmitrystas opened this issue 1 year ago • 3 comments

Versions The db-sync version (eg cardano-db-sync --version): 13.2.0.1 PostgreSQL version: 16

Build/Install Method The method you use to build or install cardano-db-sync: static binaries

Problem Report When restoring a snapshot (new version https://update-cardano-mainnet.iohk.io/cardano-db-sync/index.html#13.2/) postgresql-setup.sh fails with

pg_restore: error: could not execute query: ERROR:  role "cexplorer" does not exist
Command was: ALTER DOMAIN public.addr29type OWNER TO cexplorer;

since there is no cexlporer user in the system

The --no-owner flag (and maybe --role too?) needs to be added. Something like that

pg_restore \
  --schema=public \
  --jobs="${numcores}" \
  --format=directory \
  --dbname="${PGDATABASE}" \
  --no-owner \
  --exit-on-error \
  "${tmp_dir}/db/"

dmitrystas avatar Mar 19 '24 14:03 dmitrystas

I just ran into the same issue and had to use an older snapshot. Thanks for reporting!

119836792-2b11e4c6c8-474.lstate.gz
DROP OWNED
pg_restore: error: could not execute query: ERROR:  role "cexplorer" does not exist
Command was: ALTER DOMAIN public.addr29type OWNER TO cexplorer;

liv-io avatar Mar 27 '24 07:03 liv-io

I'm running Postgres through a container and I'm experiencing an inability to delete the DB PG: postgres:14.10-alpine

bash-5.2# 
pg_restore --schema=public --no-owner --jobs="4" --format=directory --dbname="${PGDATABASE}" --exit-on-error /tmp/db-sync-snapshot-gnn0s1exBH/db/
pg_restore: [archiver] unsupported version (1.14) in file header

Is the version of pg_restore too low?

https://update-cardano-mainnet.iohk.io/cardano-db-sync/13.2/db-sync-snapshot-schema-13.2-block-10102605-x86_64.tgz

YuXiaoCoder avatar Mar 27 '24 13:03 YuXiaoCoder

Yes, adding --exit-on-error in https://github.com/IntersectMBO/cardano-db-sync/commit/bfe1133195c10c26d8b12bcd7a92492eb8230d61 introduced that issue.

I've just finished restoration of latest available snapshot with --no-owner added to the script and it ended successfully. Once restoration ended I started cardano-db-sync and have it running for around ~1h - no issues.

ArturWieczorek avatar Apr 10 '24 13:04 ArturWieczorek

Fixed in 13.2.0.2 via 1f26396e91a0124a545201d4c6519d015b24fa6e

sorki avatar May 11 '24 10:05 sorki