Hi all,
some questios regarding HANA logging behavior:
- How can I determine the filling level of an online log (how long do we have until standstill)?
- How can I find the logs needed for a recovery? (Please assume the logs are moved to tapes already and I don’t have the FS timestamps to make a guess).
- If the system can’t write out the archive logs, because the filesystem is full, it writes 0-byte-files like MaxDB. Unlike MaxDB it doesn’t simply count up a log sequence number, but uses the filename “log_backup_<id>_<id>_<start_scn>_<end_scn>. Mapping the last 2 numbers to SCN’s (system change numbers) is an assumption. Now to the problem: If the system can’t write out the file, it will generate the next file with another set of SCN’s, beginning after the last try. Example: “log_backup_2_0_0201_0215” with size 0 byte is followed by “log_backup_2_0_0215_0248”. So what happened to the changes 201-215? I assume, the numbers don’t tell us which changes are really located in the archive log, but perhaps someone knows the internals…
- Did anyone see a HANA archiverstuck? Because I’m not able to see the online log filling level (see question one), it’s hard to say when or if I hit the point. Will the system standstill (as expected), or will it behave like after the installation, when no initial backup is available? (In this case it will not write out archivelogs, but it will also not standstill. The online logs are overwritten regardless the log parameter settings and the transactions are not recoverable, that's my observations at least.)
Thanks a lot in advance.
Best regards,
Dirk
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.