Вооот, кое-что прояснилось.
DDM сказал: Warning: Fixup purged corrupt document UNID (EEB0348B:6C229047:432580B5:002D923D) from XXXXXX.nsf
По этому поводу IBM пишет
Changes were made to fixup to help with the scenario where a corrupted document is found and deleted with no deletion stub. This is the common practice of fixup in order to allow a good copy of this note to replicate in from another replica. Along with deleting the document, fixup would also clear the replication history to allow a full search to occur in order to find these documents on next replication. With a PIRC enabled database, this procedure would no longer work if the documents modified time is older than the PIRC interval.
In order to solve this problem fixup was modified to no longer clear the replication history in this case. Fixup will now put the UNID of the deleted document onto a list contained in a special object in the database. Then when the next scheduled replication occurs, if this list exists, these documents are added to the list of documents to replicate, allowing replication to recover these deleted documents while not having to clear the history. PIRC will also allow these documents to replicate into the database even if they are older than the PIRC date. A DDM event is generated when a document is deleted from a database and its UNID is added to this list. UNID's will remain on this list only until the are replicated back in or they have been on the list for longer than the PIRC interval. In either case, when they are removed from the list another DDM event is generated.
New DDM Messages related to fixup:
"Warning: Fixup purged corrupt document UNID
) from testserver/IBM"
"Warning: Replication restored document UNID) from testserver/IBM"
"Warning: Replication did not restore corrupt document, time expired, UNID) from testserver/IBM"
Поскольку в алерте есть время, когда он попуржился из базы, то уже лучше, тк в письме есть время создания дока, а в ддм время его убития, и если есть на это время бекап, то можно было бы поднять.
Но в данном конкретном случае момент создания и убития совпадают.
Что сделала - в fixup поставила -N - не убивать поврежденные доки. Чтобы на них посмотреть, что с ними не так.
Больше ничего в голову не приходит.
DDM сказал: Warning: Fixup purged corrupt document UNID (EEB0348B:6C229047:432580B5:002D923D) from XXXXXX.nsf
По этому поводу IBM пишет
Ссылка скрыта от гостей
Changes were made to fixup to help with the scenario where a corrupted document is found and deleted with no deletion stub. This is the common practice of fixup in order to allow a good copy of this note to replicate in from another replica. Along with deleting the document, fixup would also clear the replication history to allow a full search to occur in order to find these documents on next replication. With a PIRC enabled database, this procedure would no longer work if the documents modified time is older than the PIRC interval.
In order to solve this problem fixup was modified to no longer clear the replication history in this case. Fixup will now put the UNID of the deleted document onto a list contained in a special object in the database. Then when the next scheduled replication occurs, if this list exists, these documents are added to the list of documents to replicate, allowing replication to recover these deleted documents while not having to clear the history. PIRC will also allow these documents to replicate into the database even if they are older than the PIRC date. A DDM event is generated when a document is deleted from a database and its UNID is added to this list. UNID's will remain on this list only until the are replicated back in or they have been on the list for longer than the PIRC interval. In either case, when they are removed from the list another DDM event is generated.
New DDM Messages related to fixup:
"Warning: Fixup purged corrupt document UNID
) from testserver/IBM""Warning: Replication restored document UNID
) from testserver/IBM""Warning: Replication did not restore corrupt document, time expired, UNID
) from testserver/IBM"Поскольку в алерте есть время, когда он попуржился из базы, то уже лучше, тк в письме есть время создания дока, а в ддм время его убития, и если есть на это время бекап, то можно было бы поднять.
Но в данном конкретном случае момент создания и убития совпадают.
Что сделала - в fixup поставила -N - не убивать поврежденные доки. Чтобы на них посмотреть, что с ними не так.
Больше ничего в голову не приходит.