Subject | Re: consistency check for nbackup |
---|---|
Author | heilsteffen |
Post date | 2007-11-20T21:26:16Z |
Hi
However, I consider this a bug of gfix -v or at least a reason for a
important feature request. A validate function should validate as
strict as possible. It might differentiate between warnings and
errors though.
However let me be more specific about my intention: I do backups and
test restores with nbackup and plan on running gfix -v on the result.
If gfix -v returns with zero exit code, can I be REALLY sure then,
that the result is a structurally intact database. I must be
absolutely sure, that I can connect to it and read data out of it.
I can accept null values in non-null value fields, but I cannot
accept any condition which would prevent me to do "select *" on any
table. And I would absolutely pefer a state, where will never get
data back, that no client ever wrote to the database.
Is gfix -v enough for that?
Which safety constrains can gfix -v give me?
Regards,
Steffen
> If you wish a quick check for internal structure erros you couldThe facts are also my impressions so far.
> try gfix, but if you want a full check (null fields in not
> null columns and the like) the only way I am aware of is to
> use gbak (not only back-up but *restore*).
However, I consider this a bug of gfix -v or at least a reason for a
important feature request. A validate function should validate as
strict as possible. It might differentiate between warnings and
errors though.
However let me be more specific about my intention: I do backups and
test restores with nbackup and plan on running gfix -v on the result.
If gfix -v returns with zero exit code, can I be REALLY sure then,
that the result is a structurally intact database. I must be
absolutely sure, that I can connect to it and read data out of it.
I can accept null values in non-null value fields, but I cannot
accept any condition which would prevent me to do "select *" on any
table. And I would absolutely pefer a state, where will never get
data back, that no client ever wrote to the database.
Is gfix -v enough for that?
Which safety constrains can gfix -v give me?
Regards,
Steffen