patch cluster feature request

[John Salvo]

Tony Curtis wrote:

> To Sun: how about adding a progress counter for
> install_cluster?
>
> At the moment you just see
>
> Installing 110723-06...
> Installing 110460-30...
>
> or similar. What about?
>
> Installing 110723-06 (104 of 128)...
> Installing 110460-30 (105 of 128)...
>
>
> Could install_cluster also provide a better exit code
> report than just the exit code? The various codes
> (e.g. 2, 8, ...) have meanings, so why not print out the
> explanatory text too?
>
> t

What's wrong with /var/sadm/install_data/<cluster_name>_log ?

[Tony Curtis]

>> On Tue, 29 Jul 2003 03:40:00 GMT,

>> John Salvo <jmsalvo@yahoo.com.au> said:

> Tony Curtis wrote:

>> To Sun: how about adding a progress counter for
>> install_cluster? At the moment you just see Installing
>> 110723-06... Installing 110460-30... or similar.
>> What about? Installing 110723-06 (104 of 128)...
>> Installing 110460-30 (105 of 128)... Could
>> install_cluster also provide a better exit code report
>> than just the exit code? The various codes (e.g. 2, 8,
>> ...) have meanings, so why not print out the
>> explanatory text too? t

> What's wrong with
> /var/sadm/install_data/<cluster_name>_log ?

Nothing per se, it's just that sometimes I am patching the
system in single-user mode remotely through a serial
connection and it would be nice to see a progress
indicator in that window.

t

[kgyula@mailcity.com]

Tony Curtis <tony_curtis32@yahoo.com> wrote in message news:<87r84af15u.fsf@limey.hpcc.uh.edu>...

> >> On Tue, 29 Jul 2003 03:40:00 GMT,
> >> John Salvo <jmsalvo@yahoo.com.au> said:

>
> Tony Curtis wrote:

> >> To Sun: how about adding a progress counter for
> >> install_cluster? At the moment you just see Installing
> >> 110723-06... Installing 110460-30... or similar.
> >> What about? Installing 110723-06 (104 of 128)...
> >> Installing 110460-30 (105 of 128)... Could
> >> install_cluster also provide a better exit code report
> >> than just the exit code? The various codes (e.g. 2, 8,
> >> ...) have meanings, so why not print out the
> >> explanatory text too? t

>

> > What's wrong with
> > /var/sadm/install_data/<cluster_name>_log ?

>
> Nothing per se, it's just that sometimes I am patching the
> system in single-user mode remotely through a serial
> connection and it would be nice to see a progress
> indicator in that window.
>
> t
>

tail -f /var/sadm/install_data/<cluster_name>_log

and you can see the progress...

[John Salvo]

Tony Curtis wrote:

>>>On Tue, 29 Jul 2003 03:40:00 GMT,
>>>John Salvo <jmsalvo@yahoo.com.au> said:

>
>

>>Tony Curtis wrote:
>>

>>>To Sun: how about adding a progress counter for
>>>install_cluster? At the moment you just see Installing
>>>110723-06... Installing 110460-30... or similar.
>>>What about? Installing 110723-06 (104 of 128)...
>>>Installing 110460-30 (105 of 128)... Could
>>>install_cluster also provide a better exit code report
>>>than just the exit code? The various codes (e.g. 2, 8,
>>>...) have meanings, so why not print out the
>>>explanatory text too? t

>
>

>>What's wrong with
>>/var/sadm/install_data/<cluster_name>_log ?

>
>
> Nothing per se, it's just that sometimes I am patching the
> system in single-user mode remotely through a serial
> connection and it would be nice to see a progress
> indicator in that window.
>
> t

../install_<cluster_name> &
tail -f /var/sadm/install_data/<cluster_name>log

?

[Tony Curtis]

>> On Wed, 30 Jul 2003 02:19:30 +1000,

>> John Salvo <jmsalvo@yahoo.com.au> said:

> Tony Curtis wrote:

>>>> On Tue, 29 Jul 2003 03:40:00 GMT, John Salvo
>>>> <jmsalvo@yahoo.com.au> said:

>>

>>> Tony Curtis wrote:
>>>> To Sun: how about adding a progress counter for
>>>> install_cluster? At the moment you just see
>>>> Installing 110723-06... Installing 110460-30... or
>>>> similar. What about? Installing 110723-06 (104 of
>>>> 128)... Installing 110460-30 (105 of 128)... Could
>>>> install_cluster also provide a better exit code
>>>> report than just the exit code? The various codes
>>>> (e.g. 2, 8, ...) have meanings, so why not print out
>>>> the explanatory text too? t

>>

>>> What's wrong with
>>> /var/sadm/install_data/<cluster_name>_log ?

>> Nothing per se, it's just that sometimes I am patching
>> the system in single-user mode remotely through a
>> serial connection and it would be nice to see a
>> progress indicator in that window. t

> ./install_<cluster_name> & tail -f
> /var/sadm/install_data/<cluster_name>log

Yes, but neither output shows how *far* into the patch
sequence you are. Currently I compare the output with the
patch_order file, but adding a count-o-meter to
install_cluster is trivial and would make it obvious how
many patches remain.

[John]

> Yes, but neither output shows how *far* into the patch

> sequence you are. Currently I compare the output with the
> patch_order file, but adding a count-o-meter to
> install_cluster is trivial and would make it obvious how
> many patches remain.

I must say that I like the idea, it gives you a good overview how far
the patching is.
When you need to install patches on mutiple sun's this would be very
handy!

[Oscar del Rio]

> To Sun: how about adding a progress counter for install_cluster?

>
> At the moment you just see
>
> Installing 110723-06...
> Installing 110460-30...
>
> or similar. What about?
>
> Installing 110723-06 (104 of 128)...
> Installing 110460-30 (105 of 128)...
>
>
> Could install_cluster also provide a better exit code
> report than just the exit code? The various codes
> (e.g. 2, 8, ...) have meanings, so why not print out the
> explanatory text too?

I think the MU installation scripts do both progress
and error report. I do not know why this has not
been ported to the install_cluster.
If the install_cluster script is always the same (I think
all it does is basically run patchadd for every patch in
patch_order) you could edit the script, save it elsewhere
and run your copy for every patch cluster instead of the
bundled one.

[Tony Curtis]

>> On Wed, 30 Jul 2003 12:49:33 GMT,

>> "Oscar del Rio" <delrio@mie.utoronto.ca> said:

> I think the MU installation scripts do both progress and
> error report. I do not know why this has not been
> ported to the install_cluster. If the install_cluster
> script is always the same (I think all it does is
> basically run patchadd for every patch in patch_order)
> you could edit the script, save it elsewhere and run
> your copy for every patch cluster instead of the bundled
> one.

It does seem to be fairly standard, but I thought I'd
spread the joy amongst y'all :-)

It's such a trivial addition, and at least IMHO would add
so much good-vibe-feeling to the process.

(Someone from Sun already contacted me and said they'd
submitted a feature request for it.)

t