cancel
Showing results for 
Search instead for 
Did you mean: 

abap/heap_area_total greater than maximum 2147483647

Former Member
0 Kudos

Hello.

I made changes for parameters abap/heap_area_nondia, abap/heap_area_total, abap/heap_area_dia but when i checked values i recieved message:

<b>abap/heap_area_dia 3000000000

E:abap/heap_area_dia greater than maximum 2147483647

ztta/roll_area 6500000

abap/heap_area_nondia 3000000000

E:abap/heap_area_nondia greater than maximum 2147483647

abap/heap_area_total 4000000000

E:abap/heap_area_total greater than maximum 2147483647 </b>

whats whrong?

Accepted Solutions (0)

Answers (1)

Answers (1)

Former Member
0 Kudos

Hi,

Change these values so that they don't cross the maximum values using rz10.

Then restart the cluster for the changes to take affect.

regards and reward with good points of helpful

Former Member
0 Kudos

hi.

I'd like set abap/heap_area_total = 4Gb specially

our system 64-bit, 8Gb physical memory, 20GB swap.

regards,

Denis

vincentlim826
Employee
Employee
0 Kudos

It is fine with that setting because you're in 64 bits. See SAP Notes:146289 & 425207.

Former Member
0 Kudos

so i can enlarge em/initial_size_MB and em/roll_extension only?

regards,

Denis

Satyabrata_Basu
Contributor
0 Kudos

Hi Denis,

I will start with indeed increasing em/initial_size_MB and ztta/roll_extension.

Most of the cases that should abap/heap_area_* default value 2Gb is ok if you have em/initial_size_MB and ztta/roll_extension with enough value.

btw: what is you SAP and Kernel version ?

Cheers !!

Satya.

Former Member
0 Kudos

hello Satya,

we're using SAP BW 3.5 , patch sapkw35018, sap kernel 64-bit , hp-ux 11.11, oracle 920.

memory 8 Gb

swap 20 Gb

em/initial_size_MB = 4092

ztta/roll_extension = 2000000000 (2 Gb)

We need use more memory, because encountered dumps:

SYSTEM_NO_ROLL, TSV_TNEW_PAGE_ALLOC_FAILED.

I have enlarged em/initial_size_MB to 5120 and ztta/roll_extension to 3000000000(3 Gb) and trying restart instance. But instance not started. Then i return value to back instance are started.

regards,

Denis

Satyabrata_Basu
Contributor
0 Kudos

Hi Denis,

after you increased the em/initial_size_MB and ztta/roll_extension did you run a 'sappfpar pf=<INSTANCE PROFILE NAME> check' to adjust ipc/shm_psize_10 and ipc/shm_psize_40 ?

also, make sure to clean share memory segment via ipcrm before you restart the instance ...

Cheers !!

Satya.

Former Member
0 Kudos

Satya,

could you tell me how clean share memory segment ?

is ipc/shm_psize_10 and ipc/shm_psize_40 parameters of hp-ux kernel?

and what consequence of checking ipc/shm_psize_10 and ipc/shm_psize_40 :

first - change parameters

second - check by running sappfpar

third - restart instance?

regards,

Denis

Satyabrata_Basu
Contributor
0 Kudos

Hi Denis,

ipc/shm_psize_10 and ipc/shm_psize_40 is parameter in SAP Instance profile. They are kind of shared memory segment. So, if you increase any memory parameter, they need to be adjusted accordingly..

Steps:

1. Increase em/initial_size_MB and ztta/roll_extension from RZ10 inside SAP and activate the profile

2. Stop SAP instance (stopsap r3)

3. then logon as <SID>adm and go to Profile directory (/sapmnt/<SID>/profile)

4. run from HPUX level: sappfpar check pf=<Your SAP instance profile name>. this should show some errors/warnings and suggest a new value for Shared memory key 10 and 40. Adjust those value for ipc/shm_key_10 and 40 accordingly

5. Check for active shared memory segment ipcs | grep sap

6. delete the active shared memory segments (if any) via ipcrm -m <number from ipcs output> -s <number from ipcs output> (this step is optional) More info see 'man ipcrm' under HPUX

7. restart SAP

Cheers !!

Satya.

Former Member
0 Kudos

Hello Satya,

I've made settings of parameters:

ztta/roll_extension = 3000000000

em/initial_size_MB = 5120

but when checked profile instance recieved message :

<b>ztta/roll_extension grater than 2147483647</b>.

Where is the restrictions 2147483647 of ztta/roll_extension ?

and <b>sappfpar</b> about too:

for example i have:

<b> Key: 10 Pool

Size configured.....: 124000000 ( 118.3 MB)

Size min. estimated.: 119695426 ( 114.1 MB)

Advised Size........: 122000000 ( 116.3 MB)

Key: 40 Pool for database buffers

Size configured.....: 114000000 ( 108.7 MB)

Size min. estimated.: 108975040 ( 103.9 MB)

Advised Size........: 112000000 ( 106.8 MB)

</b>

what does it mean? Should i set up ipc/shm_psize_10 = 116.3 MB and

ipc/shm_psize_40 = 106.8 MB?

regards,

Denis

Satyabrata_Basu
Contributor
0 Kudos

Hi Denis,

It's strange. For 6.40 kernel this ztta/roll_extension should not have any more 2GB limit...

Anyway, could you check if your HP-UX kernel paramaters are set properly as per recommendation from note: 172747. That might cause the bottleneck.

However, as you can increase em/initial_size_MB, I propsoe to increase it even further (e.g. 7168 ) and keep ztta/roll_extension as 2GB as now.

That already should help you...

regarding ipc/shm_psize_* looks your configured value s already bigger than they advise. So, no changes would be required.

Cheers !!

Satya.

Former Member
0 Kudos

Hello Satya.

i found this in tr.OS04:

Parameter Value defined Value used

acctresume 4

acctsuspend 2

bufpages 0

dbc_max_pct 8

dbc_min_pct 5

dst 1

fs_async 0

maxdsiz 966.852

maxfiles 2.048

maxfiles_lim 2.048

maxmem 1.892.672

maxssiz 81.920

maxswapchunks 16.384

maxtsiz 65.536

maxuprc 400

maxusers 1.024

<b>msgmap 2.048 </b>

msgmax 8.192

msgmnb 65.535 65.535

msgmni 50

msgseg 32.767 32.767

msgssz 32 32

msgtql 2.046 2.046

nbuf 0

nfile 15.195 15.205

nflocks 800

ninode 9.404

nproc 8.212

npty 60

nstrpty 60

nswapdev 10

nswapfs 10

<b>physmem 2.096.640 </b>

scroll_lines 100

semaem 16.384

semmap 522

semmni 520 520

semmns 1.024 1.024

semmnu 256 256

semume 100 100

semvmx 32.767

shmmax 16.777.216 16.777.216

shmmni 256 256

shmseg 200 200

swchunk 2.048

timeslice 10

timezone 420

unlockable_mem 0

<b>maxdsiz_64bit 16777216 </b>

what's wrong? physmem or msgmap or maxdsiz_64bit ?

regards,

Denis

Satyabrata_Basu
Contributor
0 Kudos

Hi Denis,

You need to increase atleast....

<b>shmmax = >=17,179,869,184

maxdsiz_64bit >=17,179,869,184</b>

msgmap should take from msgtql which needs to be set at 4096.

pls go thru note 172747 and set all the parameter accordingly.

you will then need to regenerate the kernel and reboot the machine.

Cheers !!

Satya.

Former Member
0 Kudos

Hello Satya,

We can't change parameters of hp-ux kernel, because strong reglament there is, i.e. it's work of other men - specialist of hp-ux.

But, you now, this some strange, we''ve occupied 3GB &#1086;f extended memmory howewer and what is more we've occupied 5Gb. although during checking profile instance (tr. RZ10) we are getting message: ztta/roll_extension grater than 2147483647.

Regards,

Denis

Satyabrata_Basu
Contributor
0 Kudos

Hi Denis,

You have to ask your HPUX administrators to prepare Unix server as required by SAP

I can't see why you cannot increase ztta/roll_extension. From 6.40 kernel (64-bit) it's up to 16,384 petabytes. For 32-bit the limit is at 2 GB.

I guess you first make sure you are using indeed 64-bit kernel and then there is no limitation imposed by OS.

Cheers !!

Satya.

Former Member
0 Kudos

Hello Satya,

Thanks for help.

We connected with SAP Support. They clear we can ignore message ztta/extension about.

regards,

Denis.

Satyabrata_Basu
Contributor
0 Kudos

Hi Denis,

Thank you for the feedback. Glad to know that the problem is solved.

Cheers !!

Satya.