OS/2 client DEAD????

I attempted to restart seti on my TPad600, downloading the latest 
version that Yuri had released last year.

The existing OS/2 client was complaining that it had an old account key, 
but it was working fine

I got far enough into it to see that I needed to detach and reattach the 
TPad.   With a Skype session I found out how to get a seti account key, 
detached and reattached the client to seti.

The client does not resume work, nor does it attempt to go get work.   
Using BoincMenu, I send RPC requests which do get acknowleged, yet al it 
will do is say "Scheduler request succeeded" and "platfrom 
'i386-pc-os2-emx' not found" and dies there.

I tried this again on the existing os/2 box, detach, reattach.   Same 
result.  It cannot go get any new work.

Seti OS/2 client is now dead!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!

 

Re: OS/2 client DEAD????

On Wed, 04 Apr 2007 09:21:36 -0500, Robert Gammon wrote:

>I attempted to restart seti on my TPad600, downloading the latest 
>version that Yuri had released last year.

Works fine here. I download work and upload results very regularly.

Cheers
Andreas

--

Andreas Ludwig
Point Grey, Vancouver, BC
using PMMail/2 on eComStation 1.2MR on a Dual Opteron 244

 

Re: OS/2 client DEAD????

Andreas Ludwig wrote:
>
> On Wed, 04 Apr 2007 09:21:36 -0500, Robert Gammon wrote:
>
> >I attempted to restart seti on my TPad600, downloading the latest
> >version that Yuri had released last year.
>
> Works fine here. I download work and upload results very regularly.
>
> Cheers
> Andreas
>
> --
>
> Andreas Ludwig
> Point Grey, Vancouver, BC
> using PMMail/2 on eComStation 1.2MR on a Dual Opteron 244
>

Mine worked fine on my main box except for a note that the account key 
was out of date.  Downloads, uploads, no problem.  So we are agreed, so 
long as you do NOT attempt to update the account key, its fine

Detach and reattach with a new account key, and DEAD in the water

 

Re: OS/2 client DEAD????

** Reply to message from Robert Gammon <rgammon51 <at> yahoo.com> on Wed, 04 Apr
2007 09:21:36 -0500

>I got far enough into it to see that I needed to detach and reattach the 
>TPad.   With a Skype session I found out how to get a seti account key, 
>detached and reattached the client to seti.
>
>The client does not resume work, nor does it attempt to go get work.   
>Using BoincMenu, I send RPC requests which do get acknowleged, yet al it 
>will do is say "Scheduler request succeeded" and "platfrom 
>'i386-pc-os2-emx' not found" and dies there.
>
>I tried this again on the existing os/2 box, detach, reattach.   Same 
>result.  It cannot go get any new work.
>
>Seti OS/2 client is now dead!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!

How absolutely ridiculous!!!!!!!!!!!!!!!!!!!!!!!!!!!

Without any confirmation you unilaterally declare the software dead because you
ran into a problem?

Well, Chicken Little, did you bother to read the readme.txt file? Under
"Installation" it says...

Install the Boinc client code.
Attach to the Seti project.
Terminate boinc client. Remove <min_rpc_time> tag in client_state.xml,
otherwise you have
to wait 24 hours for the first work unit.

Re: OS/2 client DEAD???? maybe not

OK, JOHN,

I terminated BOINC

I removed the min_rpc_time line from client_state on BOTH machines

Restarted BOINC

Both machines say Scheduler request succeeded, no work from Project

At least two cycles of this so far.

Cumbersome process

Yes, I may be WRONG about the client being DEAD

At least they are both TRYING to get work, so far unsuccessful

Another report tomorrow

Server status page says there are problems at the host right now

 

Re: OS/2 client DEAD???? maybe not

Re: OS/2 client DEAD???? maybe not, DEFINITELY NOT

OK, all is working correctly on all machines

 

woun't finish in time

At one time I could keep 4 or 5 work units downloaded for any seti outage but
now I can not get the system to download any ahead.  I have been trying to find
out why I could not download work units so I could run them whenever the seti
servers were down or unavailable.

Looking at the log file I find an entry 2007-02-04
Computer on 100% of time, BONIC on 0.1% of that, this project gets 100% of that
the next entry on 2007-02-15 says
Computer on 3.0% of time, BONIC on 0.1% of that, this project gets 100% of that
the last entry was on 2007-04-02 was the same as above.

I have BONIC set for 100% in my profile.  The computer is on 24/7.

Where is this information kept and how do I reset it?

--

-- 
Robert Blair

 

Re: woun't finish in time

--- In seti-warp <at> yahoogroups.com, "Bob" <0mtblth02 <at> ...> wrote:
>
> At one time I could keep 4 or 5 work units downloaded for any seti
outage but
> now I can not get the system to download any ahead.  I have been
trying to find
> out why I could not download work units so I could run them whenever
the seti
> servers were down or unavailable.
> 
> Looking at the log file I find an entry 2007-02-04
> Computer on 100% of time, BONIC on 0.1% of that, this project gets
100% of that
> the next entry on 2007-02-15 says
> Computer on 3.0% of time, BONIC on 0.1% of that, this project gets
100% of that
> the last entry was on 2007-04-02 was the same as above.
> 
> I have BONIC set for 100% in my profile.  The computer is on 24/7.
> 
> Where is this information kept and how do I reset it?
> 
> -- 
> Robert Blair
>

Generally, the new larger WU's have a TTL (Time to Live) of about 3
weeks when issued.  I don't know how fast you process a WU.  Seti
won't issue you any more WU's if your machine is overcommitted,
meaning that a WU sitting in your cache would expire before your
machine got around to finish processing it.  Are you sure that your WU
completion time matches what the Seti benchmark says your machine is
capable of? If by chance you never changed your Seti preferences for
network connection interval for the old smaller WU's to a longer
interval necessary for the longer WU's to cache enough work to allow
for outages, then login to your account at the Seti web site and go to
your general preferences and increase your connect interval in the
network usage area.  I complete a WU in around half to three-quarters
of a day and I have generally around 6 WU's cached at any time.  My
network connect interval is every 2 days.  The Seti guys are doing
some major server shifting around this past week and things have been
intermittent.  I'd wait a week for the dust to settle down before
becoming alarmed about lack of work to process.

Keith

 

Re: Re: woun't finish in time

** Reply to message from "Keith Myers keithmyers <at> ca.rr.com" on Sat, 07 Apr 2007
16:02:04 -0000

> Generally, the new larger WU's have a TTL (Time to Live) of about 3
> weeks when issued.  I don't know how fast you process a WU.

The new work units take about 14 hours.

>  Seti
> won't issue you any more WU's if your machine is overcommitted,
> meaning that a WU sitting in your cache would expire before your
> machine got around to finish processing it.  Are you sure that your WU
> completion time matches what the Seti benchmark says your machine is
> capable of?

No, how can I tell?

> If by chance you never changed your Seti preferences for
> network connection interval for the old smaller WU's to a longer
> interval necessary for the longer WU's to cache enough work to allow
> for outages, then login to your account at the Seti web site and go to
> your general preferences and increase your connect interval in the
> network usage area.

When I noticed that I had no work units downloaded ahead I changed the
connection time from 2.5 days to 10 days and still seti world not download
additional work units.

I think that the problem is because seti only thinks that my computer is on
about 3% of the time and BONIC only gets 0.1% of that time.

From the seti log
Computer on 3.0% of time, BONIC on 0.1% of that, this project gets 100% of that

If I can reset this so it indicates that the computer is on 100% of the time
and BONIC gets 100% of that there would not be a problem.  I just do not know
where that information is stored, if I did I would change it.

>  I complete a WU in around half to three-quarters
> of a day and I have generally around 6 WU's cached at any time.  My
> network connect interval is every 2 days.  The Seti guys are doing
> some major server shifting around this past week and things have been
> intermittent.  I'd wait a week for the dust to settle down before
> becoming alarmed about lack of work to process.

Maybe their server shifting has messed up my data.

--

-- 
Robert Blair