3000 locks in system?

[Former Member]

Dear Team,

At times we are getting locks up to 3000 in PRD system. we check from SM12. Due to this users can't update. we have to forcefully delete lock entries from SM12

Lock table memory is of 4 MB.

Please suggest ?

[former_member4224 · Accepted solution]

HI,

4Mb for enqueue table size usually is too small.

You can increase this size until 100Mb without problem. If you want to find out the program responsible for the overflow, take a look at a file called ENQHISOVR in the work folder of the instance where the enqueue service runs.

The enqueue service logs table overflows there. You will do find the 3 users with the respective program with the higher number of locks at the moment of the overflow in this file.

Clebio

[Former Member · Accepted solution]

HI,
> 
> 4Mb for enqueue table size usually is too small.
> You can increase this size until 100Mb without problem.  If you want to find out the program responsible for the overflow, take a look at a file called ENQHISOVR  in the work folder of the instance where the enqueue service runs.
> The enqueue service logs table overflows there. You will do find the 3 users with the respective program with the higher number of locks at the moment of the overflow in this file.
> 
> Clebio

Hi,

After analyzing lock table, SM12. I have some queries:

1. i found this file under ASCS01, however we have DVEBMGS00, but there is no enqueue process under DVEBMGS00.

Enqueue process running under ASCS01. Is this correct.

2. parameter rdisp/enqname not in profile parameter list.

3. parameter enque/table_size is under default.pfl

4. In sm12->extras-> diagnosis, output is:

Lock management operation mode

Lock management in a standalone server

HOST: SAPPRDCI

INST: 01

Diagnosis

Error: Local lock management process despite external lock management

Error: System contains more than one lock management SAPPRDCI_PRD_00

Test lock operations do not contain errors.

Optimistic locks are implemented in the kernel

Test lock operations do not contain errors.

Test lock operations do not contain errors.

What does this message mean?

[Former Member · Accepted solution]

Hi Sumit,

1. i found this file under ASCS01, however we have DVEBMGS00, but there is no enqueue process under DVEBMGS00.

Enqueue process running under ASCS01. Is this correct.

A. You are right. Standalone equeue server (ASCS) is setup in your environment. Enqueue process will not run on DVEBMGS.

2. parameter rdisp/enqname not in profile parameter list.

A. You can find something like rdisp/myname in ASCS profile

3. parameter enque/table_size is under default.pfl

A. You can either remove the parameter enque table size from default profile and keep the parameter in ASCS profile and restart your SAP system or simply make a new entry in ASCS profile.

4. In sm12->extras-> diagnosis, output is:

Lock management operation mode

Lock management in a standalone server

HOST: SAPPRDCI

INST: 01

A. What is the value of rdisp/wp_no_enq in DVEBMGS profile? It seems that the parameter rdisp/enqname is mentioned wrongly.

Number of servers: there is exactly one enqueue server in the system (system parameter rdisp/enqname)

        Number of processes: One ENQ process must run on the enqueue server. It may make sense only in certain special cases (extremely large systems) to have more than one ENQ process running (on the same instance).

Br,

Venky

[Former Member · Accepted solution]

This message was moderated.

[Former Member · Accepted solution]

Hi venky,

1. i found this file under ASCS01, however we have DVEBMGS00, but there is no enqueue process under DVEBMGS00.

Enqueue process running under ASCS01.

A. You are right. Standalone equeue server (ASCS) is setup in your environment. Enqueue process will not run on DVEBMGS.

Sumit Reply: ok,,,,, Thanks a lot. previously, i couldn't recognize standalone enqueue server. But if we have standalone enqueue server, then why it's showing DVEBMGS00 intead of DVBMGS00

2. parameter rdisp/enqname not in profile parameter list.

A. You can find something like rdisp/myname in ASCS profile

Sumit Reply: No, I haven't find parameter like rdisp/myname or rdisp/enqname. Moreover, there's no ASCS profile in system. i find only default.pfl, PRD_DVEBMGS00_SAPPRDCI, START_DVEBMGS00_SAPPRDCI

3. parameter enque/table_size is under default.pfl

A. You can either remove the parameter enque table size from default profile and keep the parameter in ASCS profile and restart your SAP system or simply make a new entry in ASCS profile.

4. In sm12->extras-> diagnosis, output is:

Lock management operation mode

Lock management in a standalone server

HOST: SAPPRDCI

INST: 01

A. What is the value of rdisp/wp_no_enq in DVEBMGS profile? It seems that the parameter rdisp/enqname is mentioned wrongly.

Number of servers: there is exactly one enqueue server in the system (system parameter rdisp/enqname)

Number of processes: One ENQ process must run on the enqueue server. It may make sense only in certain special cases (extremely large systems) to have more than one ENQ process running (on the same instance).

Sumit Reply: rdisp/wp_no_enq in DVEBMGS profile is 1. enque/table_size is 1000 in default.pfl.

[Former Member · Accepted solution]

Hi Sumit,

Could you pleas post the instance and default profiles. Don't forget to rename original hostnames.

Br,

Venky

[Former Member · Accepted solution]

Hi Sumit,

Could you pleas post the instance and default profiles. Don't forget to rename original hostnames.

Br,

Venky

[Former Member · Accepted solution]

Hi Sumit,
> 
> Could you pleas post the instance and default profiles. Don't forget to rename original hostnames.
> 
> Br,
> Venky

Hi venky,

i am afraid, iI can't provide all values of profiles. Please name a profile or two, i will available that profiles value

[former_member4224 · Accepted solution]

Lets make the things easy.

4. In sm12->extras-> diagnosis, output is:
Lock management operation mode
Lock management in a standalone server
HOST: SAPPRDCI
INST: 01

This means that you are running an ASCS (standalone enqueue server) in the hostname SAPPRDCI. The instance number is 01. If there is a sapstartsrv running for this instance you can check with "sapcontrol -nr 01 -function GetProcessList", and you should be able to check the message server and enqueue services in this isntance.

You also can check the parameter enque/process_location. If it is set to REMOTESA means that you are running an ASCS.

Go to server SAPPRDCI and try to find out the instances running in this host.

Go to SM50 of instance DVEBMGS00 and check if there is an ENQ work process. If you do find ENQ work process this means that the parameter rdisp/wp_no_enq is > 0.

If you are running and standalone instance for enqueue and message server you do not need to run ENQ work process in this instance.

Clébio

[Former Member · Accepted solution]

Lets make the things easy.

4. In sm12->extras-> diagnosis, output is:
> Lock management operation mode
> Lock management in a standalone server
> HOST: SAPPRDCI
> INST: 01

This means that you are running an ASCS (standalone enqueue server) in the hostname SAPPRDCI. The instance number is 01. If there is a sapstartsrv running for this instance you can check with "sapcontrol -nr 01 -function GetProcessList", and you should be able to check the message server and enqueue services in this isntance.

Sumit Reply: Correct output for SM12->extras->diagnosis is-

Lock management operation mode

Lock management in a standalone server

HOST: SAPPRDCI

INST: 01

Diagnosis

Error: Local lock management process despite external lock management

Error: System contains more than one lock management SAPPRDCI_PRD_00

Test lock operations do not contain errors.

Optimistic locks are implemented in the kernel

Output for "sapcontrol -nr 01 -function GetProcessList" is-

prdadm> sapcontrol -nr 01 -function GetProcessList

28.02.2012 06:43:16

GetProcessList

OK

name, description, dispstatus, textstatus, starttime, elapsedtime, pid

msg_server, MessageServer, GREEN, Running, 2012 02 19 08:09:14, 214:34:02, 19977

enserver, EnqueueServer, GREEN, Running, 2012 02 19 08:09:14, 214:34:02, 19978

You also can check the parameter enque/process_location. If it is set to REMOTESA means that you are running an ASCS.

Go to server SAPPRDCI and try to find out the instances running in this host.

Sumit Reply: Parameter enque/process_location set to REMOTESA,enque/table_size- 100000;enque/serverhost -SAPPRDCI;enque/serverinst -01 . All these parameters are set in default.pfl. Please let me know how to check which instance running on particular host. I ain't able to see instance ASCS at sap level, from OS level it is visible.

Go to SM50 of instance DVEBMGS00 and check if there is an ENQ work process. If you do find ENQ work process this means that the parameter rdisp/wp_no_enq is > 0.

Sumit Reply: i found 1 ENQ process in DVEBMGS00 , parameter rdisp/wp_no_enq is set to 1 in DVEBMGS00. But problem is that when I execute ps -ef | grep sap and sapcontrol -nr 00 -function GetProcessList for the instance DVEBMGS00, i don't find any ENQ process. Please see output from sapcontrol -nr 00 -function GetProcessList:

prdadm> sapcontrol -nr 00 -function GetProcessList

28.02.2012 06:44:13

GetProcessList

OK

name, description, dispstatus, textstatus, starttime, elapsedtime, pid

disp+work, Dispatcher, GREEN, Running, Message Server connection ok, Dialog Queue time: 0.00 sec, 2012 02 19 08:09:36, 214:34:37, 20572

rslgcoll, Central Syslog Collector, GREEN, Running, 2012 02 19 08:09:36, 214:34: 37, 20573

rslgsend, Central Syslog Sender, GREEN, Running, 2012 02 19 08:09:36, 214:34:37, 20574

igswd_mt, , GREEN, Running, 2012 02 19 08:09:36, 214:34:37, 20575

Moreover, when I check /usr/sap/PRD/DVEBMGS00/work, don't find any enqueue process. If we have a standalone ENQ server, then instance name should be DVBMGS00 instead of DVEBMGS00

Dear Clébio,Please help and let me know if any details required.......

Edited by: Sumit Jain1 on Feb 29, 2012 5:52 PM

[Former Member · Accepted solution]

Lets make the things easy.
> 
> 
> 
4. In sm12->extras-> diagnosis, output is:
> Lock management operation mode
> Lock management in a standalone server
> HOST: SAPPRDCI
> INST: 01
> 
> This means that you are running an ASCS (standalone enqueue server) in the hostname SAPPRDCI. The instance number is 01. If there is a sapstartsrv running for this instance you can check with "sapcontrol -nr 01 -function GetProcessList", and you should be able to check the message server and enqueue services in this isntance.
> 
> You also can check the parameter enque/process_location. If it is set to REMOTESA means that you are running an ASCS.
> Go to server SAPPRDCI and try to find out the instances running in this host.
> 
> Go to SM50 of instance DVEBMGS00 and check if there is an ENQ work process. If you do find ENQ work process this means that the parameter rdisp/wp_no_enq is > 0.
> 
> If you are running and standalone instance for enqueue and message server you do not need to run ENQ work process in this instance.
> 
> Clébio

Hi,

Please respond!!!!!!!!!!!

[former_member4224 · Accepted solution]

Sumit Reply: i found 1 ENQ process in DVEBMGS00 , parameter rdisp/wp_no_enq is set to 1 in DVEBMGS00. But problem is that when I execute ps -ef | grep sap and sapcontrol -nr 00 -function GetProcessList for the instance DVEBMGS00, i don't find any ENQ process. Please see output from sapcontrol -nr 00 -function GetProcessList:

You do not need the parameter rdisp/wp_no_enq i=1 in the instance DVEBMGS00 because the enqueue and message server is running on the standalone instance 01. This is the prove:

Sumit Reply: Correct output for SM12->extras->diagnosis is-
Lock management operation mode 
Lock management in a standalone server 
HOST: SAPPRDCI 
INST: 01

and

enque/process_location set to REMOTESA,

refer to the documentation:

http://help.sap.com/saphelp_nw70ehp1/helpdata/en/47/ea3ef600e83b8be10000000a421937/frameset.htm

Moreover, when I check /usr/sap/PRD/DVEBMGS00/work, don't find any enqueue process. If we have a standalone ENQ server, then instance name should be DVBMGS00 instead of DVEBMGS00

What you expect to found in /usr/sap/PRD/DVEBMGS00/work? this is only the work folder where the traces are stored.

Regarding the instance name, no!

Definitely, you are running a standalone enqueue service and you should be able to find out a profile called PRD_ASCS01_SAPPRDCI or something.

Clébio

[jay_b2]

Is your enqueue process running ? Please post your enqueue server logs.

[Former Member]

Is your enqueue process running ? Please post your enqueue server logs.

Hi,

prdadm> cat dev_enqsrv

-

---

trc file: "dev_enqsrv", trc level: 1, release: "701"

-

---

sysno 01

sid PRD

systemid 390 (AMD/Intel x86_64 with Linux)

relno 7010

patchlevel 0

patchno 32

intno 20020600

make: multithreaded, Unicode, 64 bit, optimized

pid 19978

[Thr 139669584586496] Sun Feb 19 08:09:15 2012

[Thr 139669584586496] profile /usr/sap/PRD/SYS/profile/PRD_ASCS01_SAPPRDCI

[Thr 139669584586496] hostname SAPCI

[Thr 139669584586496] initialize_global: enqueue server without replication

[Thr 139669584586496] Enqueue: EnqMemStartupAction Utc=1329619155

[Thr 139669584586496] Enqueue Info: replication enabled

[Thr 139669584586496] Enqueue Info: enque/replication_dll not set

[Thr 139669584586496] Sun Feb 19 08:09:16 2012

[Thr 139669584586496] ***LOG GEZ=> Server start [encllog.cpp 493]

[Former Member]

Hello,

Have you applied the latest oracle bundle patch?

if not then apply it.

Regards

Abhishek

[Former Member]

> Is your enqueue process running ? Please post your enqueue server logs.{q

hi,

Please reply!!!!!!!!!!!!

Edited by: Sumit Jain1 on Mar 1, 2012 7:51 AM

[Former Member]

Hi Sumit,

In addition to the above, please go through the note

Note 746138 - Analyzing lock table overflows

Br,

Venky

[rupali_karbhari3]

Hi,

Pls analyze the cause of the locks. it might occure due to batch job , 3rd party interface ect causing sudden high locks to any table.

refer [http://www.sdn.sap.com/irj/scn/go/portal/prtroot/docs/library/uuid/b0869e29-1405-2e10-27be-db1dad7703c5?QuickLink=index&overridelayout=true]

Regards,

Edited by: Rupali B on Feb 24, 2012 7:31 PM