[Info-vax] DNFS1ACP using 100% of CPU

[Sum1]

On 2013-10-13 21:57:44 +0000, David Froble said:

> Sum1 wrote:
>> On 2013-10-13 06:56:10 +0000, JF Mezei said:
>> 
>>> On 13-10-12 23:16, Sum1 wrote:
>>> 
>>>> Seems possible - no idea how to track them down as it just started "out
>>>> of the blue" and has survived multiple reboots.
>>> 
>>> dir/full dnfsX:[000000...]*.*.* of the  NFS served area. See if that
>>> triggers the 100% and if so, at which file/directory
>>> 
>>> then, write a DCL script to open/read each file in that NFS served area.
>>> See at which file it starts to go 100%
>>> 
>>> If you have a file that is a finder alias (not a unix alias), it is not
>>> clear how the OSX NFS server will feed it, and if the VMS NFS client
>>> will understand it.
>>> 
>>> If it is something which used to work, and the only thing that is
>>> changed is your web content, the problem is likely there.
>> 
>> Sigh…a little over 6,500,000 files….test timings suggest I will have to 
>> watch monitor for about 103 hours……
>> 
>> It is a good suggestion, but not feasible in this case….
>> 
> 
> Well, it's a computer, you should not have to watch anything ....
> 
> I will first admit that 6.5 million files is substantial, beyond my 
> experiences.
> 
> You could devise some procedure to access files and also monitor CPU 
> usage.  Set it up to report only positive hits.  This just might give 
> you a smaller subset of files to consider.  Then again, the problem 
> might be elsewhere, and every file might cause the same CPU usage.
> 
> Don't know your application, so this suggestion may not be feasible.  
> If it was my problem, I'd just get some large storage for the VMS 
> system and store everything there.  However, if some services occur off 
> the VMS system, this would not be a solution.  Nor does the suggestion 
> address your question.

Thanks all who have contributed with some positive thoughts and suggestions.

Quick testing by doing dir $dnfsX:[*…] shows the ACP hitting 100% 
almost immediately and staying there.  "Regular" file activity is OK 
until the NFS requests load up.  So I suspect that, as the serverr is 
under more load than ever before, it is just a limitation of the NFS 
client.

I have obtained a copy of OVMS8.4 and MUP 0800. I will do the upgrade 
and test what happens.  I will then apply the upgrade and test again.  
I will then report back if there is any improvement!