[Info-vax] DIRCACHE hit rate.

[Jan-Erik Söderholm]

Richard B. Gilbert wrote:
> Jan-Erik Söderholm wrote:
>> IanMiller wrote:
>>> On 20 Jan, 17:08, Jan-Erik Söderholm <jan-erik.soderh... at telia.com>
>>> wrote:
>>>> Hi.
>>>> I have an 7.3-2 Alpha system where I think that
>>>> the hit rate on the DIR-cache is way to low.
>>>>
>>>> DS20 with 9GB disks in BA356 shelfs.
>>>>
>>>> Here is what it typicaly looks like :
>>>>
>>>>                         CUR     AVE   MIN      MAX
>>>>   Dir Data (Hit %)     3.00    3.20  0.00    30.00
>>>>       (Attempt Rate) 312.00  548.90  0.00  2460.00
>>>>
>>>> The ACP_DIRCACHE is at the moment at 4000 (blocks)
>>>> but I have raised it to 10000 without any measurable
>>>> improvment.
>>>>
>>>> Yes, there are a few DIR's that are "large", but not
>>>> *that* large, maybe 5-10.000 files and a few DIR files
>>>> of 1-2000 blocks. Another thing is that all files
>>>> are timestamped in the filename, so there are
>>>> normaly only ;1 files.
>>>>
>>>> My batch jobs create a number of temp files during
>>>> processing and I've got a feeling that this is
>>>> slowing them down a bit.
>>>>
>>>> Now, what I'd like to ask, is if anyone knows if
>>>> this part of VMS has had any major improvments in
>>>> the 8.x versions ? We are currently thinking of
>>>> upgrading anyway, and it would be nice to know
>>>> if these hit rates would improve simply by upgrading.
>>>>
>>>> Best Regards,
>>>> Jan-Erik.
>>>
>>>
>>>
>>> What's the real issue? Is the time taken for your batch jobs too long?
>>
>> Well, It's manageable right now, but I think that the jobs runs
>> aprox 2 times long as needed. I see some run-time reduction when
>> I create a new temp-directory and re-define the temp-logical to
>> that empty directory, but after a while it's back where
>> it begun, of course.
>>
>>  From some other replies in this thread I've come to this conslusion:
>>
>> If I want a timestamp in the filename, it's better to have it in the
>> beginning, so all new file creations will happend at the end of
>> the DIR file. Say I have 5 temp files called A, B, C, D and E.
>> I have a timestamp "ts" like "yyyymmddhhmmsshh". Then files called
>>
>> dev:[dir]'ts'_A.TMP
>> dev:[dir]'ts'_B.TMP
>> dev:[dir]'ts'_C.TMP
>> dev:[dir]'ts'_D.TMP
>> dev:[dir]'ts'_E.TMP
>>
>> are better then
>>
>> dev:[dir]A_'ts'.TMP
>> dev:[dir]B_'ts'.TMP
>> dev:[dir]C_'ts'.TMP
>> dev:[dir]D_'ts'.TMP
>> dev:[dir]E_'ts'.TMP
>>
>> Lets also say that there are 1-2000 A-files, B-files
>> ans so on.
>>
>> Right ?
>>
>> Another way would be to have temp directories called [dir.yyymmdd]
>> so that there is a new directory for each day, or something lie that.
>>
> 
> A directory per day or per week or even per month, sounds like a 
> workable solution to me.  How long do you have to keep this stuff on 
> line?  How often do you need to access it?
> 

Well, they are logs from processing of mails comming in
fron all around the world. At least I have to be able
to check logs and data from jobs arriving Friday evening
on Monday. But usualy a couple of more days to handle
hollidays. I had 14 days when the application was "new"
(over 10 yrs now), then 7 days for some time and now
I save 3 or 4 days I think.

Jan-Erik.