[Crash-utility] [PATCH 1/1 V2] crash: initial note of excluded page structures
Dave Anderson
anderson at redhat.com
Fri Jan 10 14:02:30 UTC 2014
----- Original Message -----
>
> Dave, the excluding vmemmap obviously affect a user's investigation,
> is this really acceptable for you ?
>
> There is no chance to re-capture the same dump image,
> I think we should be more carefully about filtering out
> since it's an irreversible change.
> How many users want to get such a broken dump image even if
> it could be gotten faster ?
>
> Why we capture dump images, it's for analyzing, of course.
> At least, we should supply alternatives to the affected commands.
>
> Thanks
> Atsushi Kumagai
No, I don't like this at all. That is why I immediately objected
to doing the exclusion by default based upon memory size.
But I understand Cliff's request. I also don't think too much of
the eraseinfo/eppic capability either, but I understand why it's
considered necessary.
And so if you do accept it as a feature, I want the up-front warning
message crash utility message to explicitly state that the page-exclusion
is as a direct result of user/administrator actions. Perhaps the warning
message should also indicate that the new "makedumpfile -x" option (I forget
the option letter) was used. At least then the end-user will have
somebody to "blame", and will not consider it a problem with the tools
that were used.
And I absolutely do not want the crash source code sprinkled with a
bunch of alternative code paths to handle what will be a very
rarely-used option.
Thanks,
Dave
More information about the Crash-utility
mailing list