[prev in list] [next in list] [prev in thread] [next in thread]
List: evms-devel
Subject: Re: [Evms-devel] Re: [Linux-ha-dev] ccm tests
From: Lars Marowsky-Bree <lmb () suse ! de>
Date: 2003-02-28 13:20:08
[Download RAW message or body]
On 2003-02-26T21:39:49,
Ram Pai <linuxram@us.ibm.com> said:
> > If the CRM wanted to trigger EVMS recovery and wait for it to complete,
> > would it have to be triggered on all nodes or just one? Do you offer some
> > blocking call like "int evms_do_recover()", returning success or failure?
>
> It should be on all nodes. yes we do have a recovery entry point.
>
> But would'nt it help if evms registers a recovery function with CRM, so that
> CRM can call it when evms has to recover?
>
> So when the function returns, it implies that our recovery is done.
Yup, that's what I meant to have. I was just asking whether that function
would need to be called on all nodes or whether one node would act as a proxy
for all EVMS recovery. Ok, calling it on all nodes should be doable, too.
> Did your CRM paper proposal get accepted for OLS?
I didn't submit a full paper, because the schedules were too tight for me this
year; I'd have had to have the full paper ready too early, my work doesn't
allow for that. I hope I can still make it to OLS, though, and think a BOF or
WIP session would be appropriate by then.
Sincerely,
Lars Marowsky-Brée <lmb@suse.de>
--
Principal Squirrel
SuSE Labs - Research & Development, SuSE Linux AG
"If anything can go wrong, it will." "Chance favors the prepared (mind)."
-- Capt. Edward A. Murphy -- Louis Pasteur
-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf
_______________________________________________
Evms-devel mailing list
Evms-devel@lists.sourceforge.net
To subscribe/unsubscribe, please visit:
https://lists.sourceforge.net/lists/listinfo/evms-devel
[prev in list] [next in list] [prev in thread] [next in thread]
Configure |
About |
News |
Add a list |
Sponsored by KoreLogic