[prev in list] [next in list] [prev in thread] [next in thread] 

List:       linux-ha-dev
Subject:    [Linux-ha-dev] RE: [PATCH] Cases to CTS
From:       "Mi, Jun" <jun.mi () intel ! com>
Date:       2003-10-15 3:35:43
[Download RAW message or body]

Alan Robertson wrote:
> Mi, Jun wrote:
> > Hi, Alan
> >
> >          I've added three cases to CTS; the three cases are about
> > testing the bandwidth which the heartbeat used, heartbeat API
including
> > Zhuyi's ordered message API and reflection to the other node mainly
"up"
> > and "down". Please review.
> >
> >          I find two bugs. One is about log message. The patch has
shown
> > the bug clearly. The other one has been fixed in the cts.patch.
> 
> I believe the 30 second sleep is the wrong kind of fix for this
problem.  It
> already loops waiting for the resource to become available.  The other
one
> fixes a problem already fixed in the tree (or at least in my tree).
> 
Yes, It loops 5 times waiting for the resource to become available. But
looping 5 times is not enough time. Is there any other method to this
problem? Loop more times?

> >          I also want to ask you some questions about CTS. The first
one
> > is when running the CTS, it always let me input the password of the
> > exerciser node. It doesn't need password to go back to the
exerciser,
> > does it? Is it the requirement of ssh?
> 
> Yes.  But, you can configure ssh to not require this.  This is
described in
> the README file.
> 
Ok. 

> > The second one is why you did
> > import the CTS module to CTStests,py, and import CTStests module to
> > CTS.py? As I know, it is not reasonable to do it according to the
> > oriented object programming rules. I have studied the CTS; I think
we
> > can put the class of RandomTest to CTS.py.
> 
> Maybe I made a mistake ;-).  Originally everything was in one file,
then I
> split things out one way, then another.  I'll check on this when I get
home.

If you think there is need to do it, I will do it.

> >
> > The first patch is about log message. The second one is for CTS.
> >
> >
> >
> > Thanks,
> >
> > Mi, Jun
> >
> >
> >
------------------------------------------------------------------------
> >
> > diff -Nura heartbeat-1.1.2/heartbeat/hb_api.c
> heartbeat-1.1.2-cts/heartbeat/hb_api.c
> > --- heartbeat-1.1.2/heartbeat/hb_api.c	2003-07-14
12:19:07.000000000
> +0800
> > +++ heartbeat-1.1.2-cts/heartbeat/hb_api.c	2003-09-19
> 19:50:55.000000000 +0800
> > @@ -571,7 +571,7 @@
> >  		return I_API_BADREQ;
> >  	}
> >  	if ((pvalue = GetParameterValue(pname)) != NULL) {
> > -		if (ha_msg_mod(resp, F_PVALUE, pvalue)) {
> > +		if(!ha_msg_mod(resp, F_PVALUE, pvalue)){
> >  			ha_log(LOG_ERR
> >  			,	"api_parameter: cannot add "
> >  			F_PVALUE " field to message");
> 
> 
> Right.  But that's not what my tree has in it...  Mine looks like
this...
> 
>          }
>          if ((pvalue = GetParameterValue(pname)) != NULL) {
>                  if (ha_msg_mod(resp, F_PVALUE, pvalue) != HA_OK) {
>                          ha_log(LOG_ERR
> 
> 
> which I think is slightly better...
> 
Yes, it's better. The patch is just the one which I sent to mail list
:-), I pasted the original one with wrong. 


>      Alan Robertson <alanr@unix.sh>
> 
> "Openness is the foundation and preservative of friendship...  Let me
claim
> from you at all times your undisguised opinions." - William
Wilberforce

Alan, I am a student in Nanjing University. I work on Linux-ha in Intel
as an intern. Now I am focusing on CTS. 

Do you have any requirement about CTS? Should we test DRBD? What is the
detail requirement?

Thanks,
Mi, Jun

Not speaking for Intel, all opinions are my own. 
_______________________________________________________
Linux-HA-Dev: Linux-HA-Dev@lists.linux-ha.org
http://lists.linux-ha.org/mailman/listinfo/linux-ha-dev
Home Page: http://linux-ha.org/

[prev in list] [next in list] [prev in thread] [next in thread]