Stopsap all taking long time to stop the server

[Former Member]

hi all,

i tried to stop my SAP server from <sid>adm with the command: stopsap all.

however it is still a the stage :

Checking IXP Database

-

---

ABAP Database is running

Stopping the SAP instance DVEBMGS00

-

---

Shutdown-Log is written to /home/ixpadm/stopsap_DVEBMGS00.log

Instance on host sapxiprod stopped

Waiting for cleanup of resources......................................

.......

Stopping the SAP instance SCS01

-

---

Shutdown-Log is written to /home/ixpadm/stopsap_SCS01.log

Instance on host sapxiprod stopped

Waiting for cleanup of resources.........

Running /usr/sap/IXP/SYS/exe/run/stopdb

Trying to stop IXP database ...

Log file: /home/ixpadm/stopdb.log

its not going ahead from this stage,

when i checked alert_SID.log from /oracle/SID/saptrace/background, i am getting following output:

Wed Aug 25 05:15:17 2010

Beginning log switch checkpoint up to RBA [0x17d.2.10], SCN: 3003194

Thread 1 advanced to log sequence 381

Current log# 1 seq# 381 mem# 0: /oracle/IXP/origlogA/log_g11m1.dbf

Current log# 1 seq# 381 mem# 1: /oracle/IXP/mirrlogA/log_g11m2.dbf

Wed Aug 25 05:15:17 2010

Shutting down instance: further logons disabled

Wed Aug 25 05:15:17 2010

Completed checkpoint up to RBA [0x17d.2.10], SCN: 3003194

Wed Aug 25 05:15:17 2010

Stopping background process QMNC

Wed Aug 25 05:15:17 2010

Stopping background process CJQ0

Wed Aug 25 05:15:20 2010

Stopping background process MMNL

Wed Aug 25 05:15:21 2010

Stopping background process MMON

Wed Aug 25 05:15:22 2010

Shutting down instance (immediate)

License high water mark = 26

Wed Aug 25 05:15:22 2010

Stopping Job queue slave processes

Wed Aug 25 05:15:22 2010

Job queue slave processes stopped

Wed Aug 25 05:17:01 2010

Incremental checkpoint up to RBA [0x17d.6.0], current log tail at RBA [0x17d.6.0

]

Wed Aug 25 05:20:21 2010

Active call for process 418110 user 'oraixp' program 'oracleIXP@sapxiprod'

SHUTDOWN: waiting for active calls to complete.

Wed Aug 25 05:35:22 2010

MMNL absent for 1227 secs; Foregrounds taking over

sapxiprod:ixpadm 34>

Can some body please help me.

In ps -ef | grep sap

i can still see the process

ixpadm 672088 225614 0 05:13:29 pts/0 0:00 /bin/sh /usr/sap/IXP/SYS/exe/ru

n/stopsap all

please help me.

Regards,

Priya

[Former Member]

Hi Priya,

Wed Aug 25 05:20:21 2010
Active call for process 418110 user 'oraixp' program 'oracleIXP@sapxiprod'
SHUTDOWN: waiting for active calls to complete.

Reason for above issue:Oracle is rolling back all active transactions. It has disabled all current session from performin any further actions, it has also disabled any new connections to the database, but in order to shut down a database in consistent state, it must perform a rollback of all the transactions that have not yet been commited at the time when shutdown has begun. If those transaction were "huge" it might take a while before the rollback (and thus shutdown) completes.

If this takes too long or in the event that database for some reason realy "hangs", you still have the option to perform a SHUTDOWN ABORT.

You can always shutdown your DB first if you are facing issue like hang of stopsap... I know its not a right way but sometimes we need to do it... first try with shutdown immediate else... go with abort option... and later you can kill sap process.... I hope your none of the datafile is in backup mode.. check if any backup is running....

PS: Before issuing shutdown abort always issue "alter system checkpoint"

Also have a look at below post

All the best !

[Former Member]

hi i tried stopping the database manually using :

shutdown immediate,

even that has gone in hang stage.

tell me one thing if i so shutdown abort, would i need to do recover.

I have no backup, as i had just installed this server and were performing installlation steps,

when i tried to restart the server.

regards,

priya

[Former Member]

Priya,

Check whether u have any BRTools sessions opened and connected to this database. Sometimes this will let the shutdown command to hang for long time.

Issuing shutdown abort is not a good practice as the transaction roll-out and comits will not happen when this is issued. Thus leading to DB inconsistency.

As it is a new server, I guess thr shudn't be any problem while issuing shutdon abort.

Good Luck.

Regards,

Varadharajan M

[Former Member]

how do i check whether any br tool session is opened or not?

with ps -ef | grep br m not getting anything.

when i issues ps -ef | grep ora i got following output:

ixpadm 73844 1 0 02:16:35 - 0:00 ora_arc1_IXP

oraixp 196800 507954 0 07:31:54 pts/0 0:00 sqlplus

oraixp 233500 512156 0 08:15:11 pts/3 0:00 -csh

root 278698 454962 0 08:24:49 pts/1 0:00 grep ora

oraixp 299194 1 0 Aug 23 - 0:01 /oracle/IXP/102_64/bin/tnslsnr

LISTENER -inherit

ixpadm 385044 1 0 02:16:30 - 0:00 ora_mman_IXP

ixpadm 495812 1 0 02:16:30 - 0:00 ora_smon_IXP

oraixp 507954 647648 0 07:30:58 pts/0 0:00 -csh

ixpadm 528506 1 0 02:16:35 - 0:00 ora_arc0_IXP

ixpadm 557088 1 0 02:16:30 - 0:00 ora_psp0_IXP

oraixp 561276 618550 0 08:15:18 - 0:00 oracleIXP (DESCRIPTION=(LOCAL=Y

ES)(ADDRESS=(PROTOCOL=beq)))

ixpadm 589970 1 0 02:16:30 - 0:00 ora_lgwr_IXP

oraixp 606332 196800 0 07:31:54 - 0:00 oracleIXP (DESCRIPTION=(LOCAL=Y

ES)(ADDRESS=(PROTOCOL=beq)))

oraixp 618550 233500 0 08:15:18 pts/3 0:00 sqlplus

oraixp 131414 1 0 Aug 21 - 0:00 /bin/sh /oracle/IXP/102_64/jdk/

bin/java -Djava.awt.headless=true -Doracle.oc4j.localhome=/oracle/IXP/102_64/sql

plus -Djava.security.properties=/oracle/IXP/102_64/oc4j/j2ee/home/config/jazn.se

curity.props -jar /oracle/IXP/102_64/oc4j/j2ee/home/oc4j.jar -config /oracle/IXP

/102_64/oc4j/j2ee/isqlplus/config/server.xml

oraixp 246254 131414 0 Aug 21 - 1:44 /oracle/IXP/102_64/jdk/bin/java

.bin -Dibm.stream.nio=true -Djava.awt.headless=true -Doracle.oc4j.localhome=/ora

cle/IXP/102_64/sqlplus -Djava.security.properties=/oracle/IXP/102_64/oc4j/j2ee/h

ome/config/jazn.security.props -jar /oracle/IXP/102_64/oc4j/j2ee/home/oc4j.jar -

config /oracle/IXP/102_64/oc4j/j2ee/isqlplus/config/server.xml

oraixp 418110 1 120 02:17:12 - 366:59 oracleIXP (LOCAL=NO)

ixpadm 438534 1 0 02:16:30 - 0:00 ora_ckpt_IXP

ixpadm 459244 1 0 02:16:30 - 0:01 ora_pmon_IXP

ixpadm 557446 1 0 02:16:30 - 0:00 ora_reco_IXP

ixpadm 610746 1 0 02:16:30 - 0:00 ora_dbw0_IXP

#

if i do shutdown abort at this stage, what prerequisite should i take? how should i start my server again post that and will DB require any recovery?

kindly suggets..

Regards,

Priya

[Former Member]

Hey Priya...

As its a new system there is no issue in giving the command with abort option.. else you can issue a checkpoint on oracle to get the consistency... followed by shutdown immediate. Pls be aware of issuing abort on live system where in you have customer data and users are working esp. Prod systems.

For your better understaing; read at below link....

[Oracle start and stop process|http://www.dba-oracle.com/t_oracle_shutdown_immediate_abort.htm]

All the best !

[Former Member]

i have got following error finally after issuing the shutdown immediate command 1 hour ago: ORA-01013: user requested cancel of current operation.

although there is no user logged into the system.

do you think i should go for shutdown abort.

Regards,

Priya

[Former Member]

Hi,

and how do i start it back after shutdown abort,

should i start the database with startup command and then issue startsap command from <sid>adm or

shopuld i directly execute startsap all

Regards,

Priyanshu Srivastava

[Former Member]

hi all,

thanks a lot for your replies.. i used shutdown abort and then started the server using startsap all. it started.

replied,

Priya

[MarcelloUrbani]

Just bumped into this problem, in my case it was just waiting on a zombie process.

Fixed replacing

while $ps_command | awk '{print $1}'...

with

while $ps_command | awk '$3~/^[^Z]/ {print $1}'...

...which basically filters out zombies

Ideally I would like to fix the underlying issue, but it's good enough for now