[sheepdog] [PATCH] sheep: fix explanation of SD_RES_OLD_NODE_VER and SD_RES_NEW_NODE_VER

[MORITA Kazutaka]

At Fri, 07 Jun 2013 12:27:35 +0800,
Liu Yuan wrote:
> 
> On 06/07/2013 11:32 AM, MORITA Kazutaka wrote:
> > From: MORITA Kazutaka <morita.kazutaka at lab.ntt.co.jp>
> > 
> > Sheep returns SD_RES_OLD_NODE_VER when the sent request contains an
> > older epoch, and SD_RES_NEW_NODE_VER when it contains an newer one.
> > 
> > Signed-off-by: MORITA Kazutaka <morita.kazutaka at lab.ntt.co.jp>
> > ---
> >  include/internal_proto.h |    4 ++--
> >  include/sheep.h          |    4 ++--
> >  2 files changed, 4 insertions(+), 4 deletions(-)
> > 
> > diff --git a/include/internal_proto.h b/include/internal_proto.h
> > index bddb5dc..24d5519 100644
> > --- a/include/internal_proto.h
> > +++ b/include/internal_proto.h
> > @@ -85,8 +85,8 @@
> >  #define SD_FLAG_CMD_DEL      0x0400
> >  
> >  /* internal error return values, must be above 0x80 */
> > -#define SD_RES_OLD_NODE_VER  0x81 /* Remote node has an old epoch */
> > -#define SD_RES_NEW_NODE_VER  0x82 /* Remote node has a new epoch */
> > +#define SD_RES_OLD_NODE_VER  0x81 /* Request has an old epoch */
> > +#define SD_RES_NEW_NODE_VER  0x82 /* Request has a new epoch */
> >  #define SD_RES_NOT_FORMATTED 0x83 /* Sheepdog is not formatted yet */
> >  #define SD_RES_INVALID_CTIME 0x84 /* Creation time of sheepdog is different */
> >  #define SD_RES_INVALID_EPOCH 0x85 /* Invalid epoch */
> > diff --git a/include/sheep.h b/include/sheep.h
> > index e917064..0d3fae4 100644
> > --- a/include/sheep.h
> > +++ b/include/sheep.h
> > @@ -206,8 +206,8 @@ static inline const char *sd_strerror(int err)
> >  		[SD_RES_READONLY] = "Object is read-only",
> >  
> >  		/* from internal_proto.h */
> > -		[SD_RES_OLD_NODE_VER] = "Remote node has an old epoch",
> > -		[SD_RES_NEW_NODE_VER] = "Remote node has a new epoch",
> > +		[SD_RES_OLD_NODE_VER] = "Request has an old epoch",
> > +		[SD_RES_NEW_NODE_VER] = "Request has a new epoch",
> >  		[SD_RES_NOT_FORMATTED] = "Cluster has not been formatted",
> >  		[SD_RES_INVALID_CTIME] = "Creation times differ",
> >  		[SD_RES_INVALID_EPOCH] = "Invalid epoch",
> > 
> 
> This isn't a fix. Just switch the view from remote to request sender.
> Seems not necessary to me.

But we see "Remote node has an old epoch" in sheep.log even though the
local node has an old epoch actually.  It's confusing.

Thanks,

Kazutaka