| T.R | Title | User | Personal
Name | Date | Lines |
|---|
| 2392.1 | | OSOSPS::KAGEYAMA | Trust, but Verify | Thu Feb 20 1997 19:16 | 7 |
|
Which part of memory in the graph generated by the Motif interface is
greater than the equivalent one by the character cell interface?
Is it at certain time period or always over 100%?
- Kazunori
|
| 2392.2 | Always over 100% | TAV02::ZVI_P | Here we are | Sun Feb 23 1997 08:09 | 5 |
| I produced a graph of memory utilization for one day. Memory Usage for the
whole day was around 132%, the graph average was 132%. I.e. it was always
over 100%
Zvi
|
| 2392.3 | | BSS::JILSON | WFH in the Chemung River Valley | Sun Feb 23 1997 11:31 | 11 |
| Look at the PHYSICALPAGES parameter from the following command, does it
match reality?
$ ADV COLL REP DUMP_DATA/BEG=09:00/END=09:02/NODE={nodename}/OUTPUT=DATA.TXT
$ SEARCH DATA.TXT PHYSICALPAGES
This data cell is used to store the actual amount of physical memory and
NOT what is in SYSGEN. Then this data cell is used in the calculation of
memory utilization.
Jilly
|
| 2392.4 | PHYSICALPAGES == memory size | TAV02::ZVI_P | Here we are | Tue Feb 25 1997 11:11 | 13 |
| Hi jilly
Here is what I get:
SYSGEN_PHYSICALPAGES 131072 --> in alpha pages of 8Mb each.
Their memory is 1024Mb.
131072*8 = 1048576 = 1024Mb, so this eems fine.
Zvi
|
| 2392.5 | | BSS::JILSON | WFH in the Chemung River Valley | Tue Feb 25 1997 12:37 | 6 |
| OK now dump the same time period for DUMP_METRIC and look at the memory
usage values (ie Free List Page Cnt) and see if anything is out of whack.
If nothing looks bad here then IPMT this case and make sure you have a .CPD
file available for engineering.
Jilly
|
| 2392.6 | Memory usage values seem fine - give it a look | TAV02::ZVI_P | Here we are | Wed Feb 26 1997 03:05 | 25 |
| Jilly give it a look please. to me it seems fine:
------------------------------- Pool & Memory ---------------------------
NP Pool Maximum 36306944 NP Pool Blocks 5096
Largest NP Block 953344 Smallest NP Block 64
PG Pool Maximum 3997696 PG Pool Blocks 108
Largest PG Block 1711088 Smallest PG Block 16
NP Free Blocks =< 32 byte 0 NP Free Bytes 4420736
PG Free Blocks =< 32 byte 74 PG Free Bytes 1716368
User Memory Pages 40976 Page Size 8192
IPMT is a real bother. I had a bad experinec with that. They insisted on
getting the stuff on physical media and not by the net. Do you know how I
can pass them a CPD file on the net. As you know we are located in Israel.
Thanks
Zvi
Thanks
Zvi
|
| 2392.7 | | BSS::JILSON | WFH in the Chemung River Valley | Wed Feb 26 1997 08:45 | 11 |
| Zvi,
The easiest way to get a .CPD file to CA is to put it in a saveset,
zip it and the FTP it to them. With ZIP use the "-V9" options. If you
need help in getting the zipped file to CA I would be glad to help when CA
asks for the file.
You need more than what you posted. The Free List Page Cnt and Modifed
List Page Cnt are in a different location in the DUMP_METRIC report.
Jilly
|
| 2392.8 | The page counts seem weird to me | TAV02::ZVI_P | Here we are | Wed Feb 26 1997 11:23 | 31 |
|
Free List Page Cnt 14137 Modified List Page Cnt 1822
This is ok compared to:
User Memory Pages 40976
But, if page is 8192 bytes , i.e. 8Mb then all these numbers are much
greater than the memory size which is 1024.
If these are in pagelets then they seem too small. e.g. user memory of
40976 pagelets means nearly 20Mb out of 1024 - these are too small.
Can you explain that ?
> The easiest way to get a .CPD file to CA is to put it in a saveset,
>zip it and the FTP it to them. With ZIP use the "-V9" options. If you
>need help in getting the zipped file to CA I would be glad to help when CA
>asks for the file.
Thanks, I'll be glad to get that service from you.
BTW: Here it is 18:00 and I have to leave.
Tomorrow i am out of office.
I'll be seeing NOTES only on Sunday 2-Mar-97.
Have a nice weekend
Zvi
|