You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@hbase.apache.org by Sergey Shelukhin <se...@hortonworks.com> on 2014/01/07 00:32:42 UTC
multi() contract or lack thereof
Hi.
Looking at code, the way multi currently works is that either the entire
thing fails with an exception, or multiresponse is returned with exactly
one result/exception per action in multiaction, in corresponding order.
This sounds like a reasonable API contract, but it could be "by accident".
Do you think we can rely on this remaining true?
That way I can remove passing action index from client back and forth thru
the server (with backward compat ofc).
--
CONFIDENTIALITY NOTICE
NOTICE: This message is intended for the use of the individual or entity to
which it is addressed and may contain information that is confidential,
privileged and exempt from disclosure under applicable law. If the reader
of this message is not the intended recipient, you are hereby notified that
any printing, copying, dissemination, distribution, disclosure or
forwarding of this communication is strictly prohibited. If you have
received this communication in error, please contact the sender immediately
and delete it from your system. Thank You.
Re: multi() contract or lack thereof
Posted by Sergey Shelukhin <se...@hortonworks.com>.
The index is always with respect to the original list on client; server
just sends it back.
Actually, nevermind, I noticed for each action what I said is not true.
Non-atomic ops can reorder, and atomic ops appear to add nothing to
response (I wonder how this works...)
On Mon, Jan 6, 2014 at 3:45 PM, Jimmy Xiang <jx...@cloudera.com> wrote:
> Suppose some actions in a multi call fail. Do we change the index during
> retry?
>
>
> On Mon, Jan 6, 2014 at 3:32 PM, Sergey Shelukhin <sergey@hortonworks.com
> >wrote:
>
> > Hi.
> > Looking at code, the way multi currently works is that either the entire
> > thing fails with an exception, or multiresponse is returned with exactly
> > one result/exception per action in multiaction, in corresponding order.
> > This sounds like a reasonable API contract, but it could be "by
> accident".
> > Do you think we can rely on this remaining true?
> > That way I can remove passing action index from client back and forth
> thru
> > the server (with backward compat ofc).
> >
> > --
> > CONFIDENTIALITY NOTICE
> > NOTICE: This message is intended for the use of the individual or entity
> to
> > which it is addressed and may contain information that is confidential,
> > privileged and exempt from disclosure under applicable law. If the reader
> > of this message is not the intended recipient, you are hereby notified
> that
> > any printing, copying, dissemination, distribution, disclosure or
> > forwarding of this communication is strictly prohibited. If you have
> > received this communication in error, please contact the sender
> immediately
> > and delete it from your system. Thank You.
> >
>
--
CONFIDENTIALITY NOTICE
NOTICE: This message is intended for the use of the individual or entity to
which it is addressed and may contain information that is confidential,
privileged and exempt from disclosure under applicable law. If the reader
of this message is not the intended recipient, you are hereby notified that
any printing, copying, dissemination, distribution, disclosure or
forwarding of this communication is strictly prohibited. If you have
received this communication in error, please contact the sender immediately
and delete it from your system. Thank You.
Re: multi() contract or lack thereof
Posted by Jimmy Xiang <jx...@cloudera.com>.
Suppose some actions in a multi call fail. Do we change the index during
retry?
On Mon, Jan 6, 2014 at 3:32 PM, Sergey Shelukhin <se...@hortonworks.com>wrote:
> Hi.
> Looking at code, the way multi currently works is that either the entire
> thing fails with an exception, or multiresponse is returned with exactly
> one result/exception per action in multiaction, in corresponding order.
> This sounds like a reasonable API contract, but it could be "by accident".
> Do you think we can rely on this remaining true?
> That way I can remove passing action index from client back and forth thru
> the server (with backward compat ofc).
>
> --
> CONFIDENTIALITY NOTICE
> NOTICE: This message is intended for the use of the individual or entity to
> which it is addressed and may contain information that is confidential,
> privileged and exempt from disclosure under applicable law. If the reader
> of this message is not the intended recipient, you are hereby notified that
> any printing, copying, dissemination, distribution, disclosure or
> forwarding of this communication is strictly prohibited. If you have
> received this communication in error, please contact the sender immediately
> and delete it from your system. Thank You.
>