You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@geode.apache.org by Kenneth Howe <kh...@pivotal.io> on 2016/09/09 20:40:03 UTC
Gfsh Command enhancement - show missing-disk-stores
GEODE-1128 <https://issues.apache.org/jira/browse/GEODE-1128> requests the addition of missing colocated region information to the gfsh show missing-disk-stores command. This is of course doable, however with additional information not directly related to disk stores, the command name would be misleading. You may have missing colocated regions without missing disk stores, or the converse, missing stores without missing regions.
So my question is: Should the command be renamed to better indicate the types of information reported? While working on this Jira, I have been using the new command name ‘show persistent-recovery-failures’. (Please suggest a better name!)
Alternatives to renaming the command are
1) Do nothing with the command name. Add the missing colocated region information, but leave the command name as is.
2) Add a new command with both missing disk stores and missing colocated regions, and leave the existing missing-disk-stores command as is.
Looking for a consensus on the best approach.
Thanks,
Ken Howe
Re: Gfsh Command enhancement - show missing-disk-stores
Posted by Darrel Schneider <ds...@pivotal.io>.
+1 for a single command
I'm ok with keeping the old command name and just having it show more info.
On Fri, Sep 9, 2016 at 2:53 PM, Dan Smith <ds...@pivotal.io> wrote:
> I think it should be a single command because the user is trying to
> diagnose the same problem - what persistent data is missing that is
> preventing system recovery? I'm not sure what the best name of the command
> is.
>
> -Dan
>
> On Fri, Sep 9, 2016 at 1:40 PM, Kenneth Howe <kh...@pivotal.io> wrote:
>
> > GEODE-1128 <https://issues.apache.org/jira/browse/GEODE-1128> requests
> > the addition of missing colocated region information to the gfsh show
> > missing-disk-stores command. This is of course doable, however with
> > additional information not directly related to disk stores, the command
> > name would be misleading. You may have missing colocated regions without
> > missing disk stores, or the converse, missing stores without missing
> > regions.
> >
> > So my question is: Should the command be renamed to better indicate the
> > types of information reported? While working on this Jira, I have been
> > using the new command name ‘show persistent-recovery-failures’. (Please
> > suggest a better name!)
> >
> > Alternatives to renaming the command are
> > 1) Do nothing with the command name. Add the missing colocated region
> > information, but leave the command name as is.
> > 2) Add a new command with both missing disk stores and missing colocated
> > regions, and leave the existing missing-disk-stores command as is.
> >
> > Looking for a consensus on the best approach.
> >
> > Thanks,
> > Ken Howe
> >
>
Re: Gfsh Command enhancement - show missing-disk-stores
Posted by Dan Smith <ds...@pivotal.io>.
I think it should be a single command because the user is trying to
diagnose the same problem - what persistent data is missing that is
preventing system recovery? I'm not sure what the best name of the command
is.
-Dan
On Fri, Sep 9, 2016 at 1:40 PM, Kenneth Howe <kh...@pivotal.io> wrote:
> GEODE-1128 <https://issues.apache.org/jira/browse/GEODE-1128> requests
> the addition of missing colocated region information to the gfsh show
> missing-disk-stores command. This is of course doable, however with
> additional information not directly related to disk stores, the command
> name would be misleading. You may have missing colocated regions without
> missing disk stores, or the converse, missing stores without missing
> regions.
>
> So my question is: Should the command be renamed to better indicate the
> types of information reported? While working on this Jira, I have been
> using the new command name ‘show persistent-recovery-failures’. (Please
> suggest a better name!)
>
> Alternatives to renaming the command are
> 1) Do nothing with the command name. Add the missing colocated region
> information, but leave the command name as is.
> 2) Add a new command with both missing disk stores and missing colocated
> regions, and leave the existing missing-disk-stores command as is.
>
> Looking for a consensus on the best approach.
>
> Thanks,
> Ken Howe
>