on 11-30-2007 2:49 PM
Hi,
I have a server with st02 like that :
SAP Memory Curr.Use % CurUse[KB] MaxUse[KB] In Mem[KB] OnDisk[KB] SAPCurCach HitRatio %
Roll area 0,78 8.206 9.760 1.048.576 0 IDs 95,94
Page area 0,70 7.302 91.960 524.288 524.288 Statement 89,00
Extended memory 53,81 5.320.704 7.442.432 9.887.744 0 0,00
Heap memory 0 0 0 0 0,00
But there is very few people connected : 30 users .
In st03g , i see that it s not a single specific programm witch is responsible :
-
-
Programme/Transaction | Nom du job | 1. étapes | Ø mémoire | Ø mém. ét. | Maxi ME | 1. rés. PT | Ø mémoire priv. (KB) | Mém.privée maxi (KB) | Relances |
-
-
<b>MIR4 </b> | 18 | 204.377 | 11.366 | <b>249.590</b> | 0 | 0 | 216.858 | 0 | |
F871 | 8 | 201.536 | 4.092 | 249.590 | 0 | 0 | 204.583 | 0 | |
FMKFR01 | 3 | 200.513 | 4.092 | 249.590 | 0 | 0 | 200.491 | 0 | |
MIR7 | 27 | 187.478 | 7.880 | 249.590 | 0 | 0 | 208.674 | 0 | |
S_AHR_61016404 | 87 | 139.467 | 6.819 | 257.775 | 0 | 0 | 241.408 | 0 | |
S_AHR_61016405 | 51 | 138.977 | 8.023 | 245.498 | 0 | 0 | 139.116 | 0 | |
PR05 | 22 | 138.766 | 7.997 | 249.590 | 0 | 0 | 167.758 | 0 | |
ME23N | 82 | 136.094 | 11.826 | 274.142 | 0 | 0 | 208.674 | 0 | |
ME22N | 1 | 135.046 | 12.275 | 245.499 | 0 | 0 | 135.025 | 0 | |
FBL1N | 28 | 133.291 | 11.106 | 245.499 | 0 | 0 | 135.025 | 0 | |
PRF0 | 2 | 133.001 | 6.138 | 249.590 | 0 | 0 | 135.025 | 0 | |
S_AHR_61016402 | 12 | 132.659 | 5.797 | 245.498 | 0 | 0 | 135.025 | 0 | |
MIGO | 6 | 128.908 | 8.183 | 245.499 | 0 | 0 | 130.933 | 0 | |
VA01 | 2 | 128.908 | 6.138 | 245.498 | 0 | 0 | 130.933 | 0 | |
SESSION_MANAGER | 47 | 127.907 | 127.886 | 249.590 | 0 | 0 | 196.400 | 0 | |
SMEN | 17 | 122.771 | 122.750 | 274.142 | 0 | 0 | 122.750 | 0 | |
SP01 | 1 | 122.771 | 4.092 | 245.498 | 0 | 0 | 122.749 | 0 | |
MIRO | 175 | 119.216 | 10.124 | 274.142 | 0 | 0 | 204.583 | 0 |
Hi,
The extended memory was full but not in the screen shot.
I find the solution of my problem :
For the moment we use general role Z_ALL_ACTION for the users in Formation system. ( the final role are in developpement)
This role is very very large.
When you use this role for a normal transaction, the extended memory needed is about 200Mo / 250Mo .
If you use a profil SAP_ALL , the extended memory needed is less than 50Mo.
The problem was a problem of role.
I didnt't think that role could make a so heavy load of memory.
regards,
Patrick
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
Patrick,
Your title says "Extended memory full" yet the results you posted only show it using 53% and no heap memory being used(which is a good thing). Or am I missing something ?
If you were running out of extended memory you would find dialog work processes entering into PRIV mode which means those work processes will be locked to one particular user resulting in very poor performance. You would see this in SM50. Also, when you run out of extended memory it starts to use heap memory and you're not using any. You can check what users are using heap memory via ST02 -> Detail analysis menu -> SAP memory -> mode list(This is path in ECC5, may be different for ECC6).
If you do have an extended memory problem which I can't see you can adjust it by setting em/initial_size_MB and em/address_space_MB but only if you have enough physical memory available.
Regards,
Nelis
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
Check transaction SM04, add the "KB" column to the display and sort them. Maybe someone is consuming a lot of memory.
--
Markus
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
Hi markus,
In tx sm04 :
-
-
Programme/Transaction | Nom du job | 1. étapes | Ø mémoire | Ø mém. ét. | Maxi ME | 1. rés. PT | Ø mémoire priv. (KB) | Mém.privée maxi (KB) | Relances |
-
-
MIR4 | 18 | 204.377 | 11.366 | 249.590 | 0 | 0 | 216.858 | 0 | |
F871 | 8 | 201.536 | 4.092 | 249.590 | 0 | 0 | 204.583 | 0 | |
FMKFR01 | 3 | 200.513 | 4.092 | 249.590 | 0 | 0 | 200.491 | 0 | |
MIR7 | 27 | 187.478 | 7.880 | 249.590 | 0 | 0 | 208.674 | 0 | |
S_AHR_61016404 | 87 | 139.467 | 6.819 | 257.775 | 0 | 0 | 241.408 | 0 | |
S_AHR_61016405 | 51 | 138.977 | 8.023 | 245.498 | 0 | 0 | 139.116 | 0 | |
PR05 | 22 | 138.766 | 7.997 | 249.590 | 0 | 0 | 167.758 | 0 | |
ME23N | 82 | 136.094 | 11.826 | 274.142 | 0 | 0 | 208.674 | 0 | |
ME22N | 1 | 135.046 | 12.275 | 245.499 | 0 | 0 | 135.025 | 0 | |
FBL1N | 28 | 133.291 | 11.106 | 245.499 | 0 | 0 | 135.025 | 0 | |
PRF0 | 2 | 133.001 | 6.138 | 249.590 | 0 | 0 | 135.025 | 0 | |
S_AHR_61016402 | 12 | 132.659 | 5.797 | 245.498 | 0 | 0 | 135.025 | 0 | |
MIGO | 6 | 128.908 | 8.183 | 245.499 | 0 | 0 | 130.933 | 0 | |
VA01 | 2 | 128.908 | 6.138 | 245.498 | 0 | 0 | 130.933 | 0 | |
SESSION_MANAGER | 47 | 127.907 | 127.886 | 249.590 | 0 | 0 | 196.400 | 0 | |
SMEN | 17 | 122.771 | 122.750 | 274.142 | 0 | 0 | 122.750 | 0 | |
SP01 | 1 | 122.771 | 4.092 | 245.498 | 0 | 0 | 122.749 | 0 | |
MIRO | 175 | 119.216 | 10.124 | 274.142 | 0 | 0 | 204.583 | 0 | |
F-53 | 4 | 118.679 | 4.092 | 245.499 | 0 | 0 | 122.750 | 0 |
There is a lot of different standard transaction with high load of EM. It 's seems to not be a problem of specific programm.
I could not find more exactly how this memory is used.
regards,
patrick
If you post outputs like this, please mark the table and press the "code" button, it's almost impossible to read something (because of proportional font).
In SM04 one usually see the users and all the memory they use.
I'm not sure what you select but the first column is usually the user name. I doubt that MIR7 will consume 200.000 MB of memory, that may be just the users active mode/program.
--
Markus
Yes, sorry .
I was a mistake in copy/paste.
This is the good copy :
30.11.2007 Edition de liste dynamique 1
-
-
-
-
Mdt | Utilis. | Transaction | Roll | Page | Mém.(total) | Mém.(privé |
-
-
400 | ADMIN | SM04 | 229.376 | 65.536 | 7.002.036 | 0 |
400 | ADMIN | SM04 | 1.081.344 | 352.256 | 58.794.116 | 0 |
000 | SESSION_MANAGER | 122.880 | 0 | 825.544 | 0 | |
420 | FORMATION5 | PR05 | 229.376 | 417.792 | 94.297.540 | 0 |
420 | FORMATION2 | 122.880 | 16.384 | 1.181.052 | 0 | |
420 | FORMATION1 | PR05 | 229.376 | 98.304 | 86.807.372 | 0 |
000 | SESSION_MANAGER | 122.880 | 0 | 825.544 | 0 |
But now there is no enough users connected..... (it 's friday in france )
I will paste a better view monday when they will be connected again.
regards,
Patrick
There is a lot of standard transaction with very high extended memory consumption .
We wait in prod System more than 200 users , we could not stay in this situation.
Have you got an idea ??
regards ,
patrick
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.