Mirror: fix common issues of a replayer session

When running a replay session of the staging area storage, there a few common issues that occur, these should be fixed before starting a real-life mirror session.

---

Migrated from T3197 (view on Phabricator)

added Mirror priority:High labels

marked this issue as related to #3198

marked this issue as related to #3199

marked this issue as related to #3200

marked this issue as related to #3201 (closed)

Currently, the mirror test session is running with:

ii  python3-swh.core        0.12.0-1~swh1~bpo10+1 all          Software Heritage core utilities
ii  python3-swh.journal     0.7.1-1~swh1~bpo10+1  all          Software Heritage Journal utilities
ii  python3-swh.model       2.3.0-1~swh1~bpo10+1  all          Software Heritage data model
ii  python3-swh.objstorage  0.2.2-1~swh1~bpo10+1  all          Software Heritage Object Storage
ii  python3-swh.scheduler   0.10.0-1~swh1~bpo10+1 all          Software Heritage Scheduler
ii  python3-swh.storage     0.27.0-1~swh1~bpo10+1 all          Software Heritage storage utilities

Using unprivileged access to the journal.

The number of error after ~20h is:

$ docker service logs swh_storage  2>&1 | grep " ERROR:"  | sort | uniq -c
     20 swh_storage.1.yjtfpr7ambpm@libra    | ERROR:root:duplicate key value violates unique constraint "person_fullname_idx"
      1 swh_storage.1.yjtfpr7ambpm@libra    | ERROR:root:server closed the connection unexpectedly
   1679 swh_storage.1.yjtfpr7ambpm@libra    | ERROR:root:unsupported Unicode escape sequence
   1087 swh_storage.1.yjtfpr7ambpm@libra    | ERROR:root:year 292278994 is out of range

marked this issue as related to #3218 (closed)

marked this issue as related to #3303 (closed)