2003 server 100% cpu, unmanageable.

[gscanga]

I have a system (dc actually) that spikes at 100% cpu and stays that way. I
can't get any information whatsoever on the system when it does this (ie,
pslist to check cpu processes, mmc to check event logs, RDP to remotely view,
etc.) Several times, I've waited 15-20 minutes for a console session to
display with no luck.

So, after I cold-boot the machine, I check the event logs and can't find
anything. is there something that I can proactively setup and let run that
will grant me better insight into what may be happening when or prior to cpu
spikes? Perhaps perfmon?

Thanks.

 

[Greg O]

Sometimes WSUS and updates will cause this, if svchost is doing it. It's a
well known bug Microsoft recently provided a fix for,


"gscanga" <gscanga@discussions.microsoft.com> wrote in message
news:85F78934-3965-4299-9490-EEAF5FF0C900@microsoft.com...
>I have a system (dc actually) that spikes at 100% cpu and stays that way. I
> can't get any information whatsoever on the system when it does this (ie,
> pslist to check cpu processes, mmc to check event logs, RDP to remotely
> view,
> etc.) Several times, I've waited 15-20 minutes for a console session to
> display with no luck.
>
> So, after I cold-boot the machine, I check the event logs and can't find
> anything. is there something that I can proactively setup and let run that
> will grant me better insight into what may be happening when or prior to
> cpu
> spikes? Perhaps perfmon?
>
> Thanks.

 

[gscanga]

Thanks Greg,

I have seen that issue you've mentioned but not necessarily on this server.
When I've had that wsus issue on other systems, the cpu was pegged and
interaction was dreadfully slow.

With this issue however, I can't get any response no matter how long I wait,
and I've waited as long as 30 minutes. So it's beyond dreadfully slow and
more like comepletely unmanagable.



"Greg O" wrote:

> Sometimes WSUS and updates will cause this, if svchost is doing it. It's a
> well known bug Microsoft recently provided a fix for,
>
>
> "gscanga" <gscanga@discussions.microsoft.com> wrote in message
> news:85F78934-3965-4299-9490-EEAF5FF0C900@microsoft.com...
> >I have a system (dc actually) that spikes at 100% cpu and stays that way. I
> > can't get any information whatsoever on the system when it does this (ie,
> > pslist to check cpu processes, mmc to check event logs, RDP to remotely
> > view,
> > etc.) Several times, I've waited 15-20 minutes for a console session to
> > display with no luck.
> >
> > So, after I cold-boot the machine, I check the event logs and can't find
> > anything. is there something that I can proactively setup and let run that
> > will grant me better insight into what may be happening when or prior to
> > cpu
> > spikes? Perhaps perfmon?
> >
> > Thanks.
>
>
>

 

[Greg O]

Do you know which process is running at 100%? It should tell you that in
task manager. Sometimes this can also happen from using the software
mirroring or raid 5 in Computer Management.


"gscanga" <gscanga@discussions.microsoft.com> wrote in message
news:F3CBBB97-69DD-4496-BD4D-E7AA956DE473@microsoft.com...
> Thanks Greg,
>
> I have seen that issue you've mentioned but not necessarily on this
> server.
> When I've had that wsus issue on other systems, the cpu was pegged and
> interaction was dreadfully slow.
>
> With this issue however, I can't get any response no matter how long I
> wait,
> and I've waited as long as 30 minutes. So it's beyond dreadfully slow and
> more like comepletely unmanagable.
>
>
>
> "Greg O" wrote:
>
>> Sometimes WSUS and updates will cause this, if svchost is doing it. It's
>> a
>> well known bug Microsoft recently provided a fix for,
>>
>>
>> "gscanga" <gscanga@discussions.microsoft.com> wrote in message
>> news:85F78934-3965-4299-9490-EEAF5FF0C900@microsoft.com...
>> >I have a system (dc actually) that spikes at 100% cpu and stays that
>> >way. I
>> > can't get any information whatsoever on the system when it does this
>> > (ie,
>> > pslist to check cpu processes, mmc to check event logs, RDP to remotely
>> > view,
>> > etc.) Several times, I've waited 15-20 minutes for a console session to
>> > display with no luck.
>> >
>> > So, after I cold-boot the machine, I check the event logs and can't
>> > find
>> > anything. is there something that I can proactively setup and let run
>> > that
>> > will grant me better insight into what may be happening when or prior
>> > to
>> > cpu
>> > spikes? Perhaps perfmon?
>> >
>> > Thanks.
>>
>>
>>