Troubleshooting
Problem
The Tivoli Storage Manager server is very slow. A simple select command take about 5 minutes to complete.
Cause
An empty /etc/netsvc.conf or entry with hosts=local,bind can trigger DNS lookup for IPv6 and can cause the delay.
Environment
IBM Spectrum Protect Server on AIX
Diagnosing The Problem
1. Run the "show thread" command to view the call stack of the administrative command running the select command. The stack shows a DNS lookup is being done. For example :
Thread 3508888, Parent 3508864: TsmSqlPrepareData, Storage 33600,
AllocCnt 5 HighWaterAmt 33600
tid=d098, ptid=11180, det=0, zomb=0, join=0, result=0, sess=0
Stack trace:
0x090000000012cdd4 __fd_poll
0x0900000000102a30 poll
0x0900000000101a58 res_nsend
0x090000000012c2dc res_nquery
0x090000000012b818 res_nquerydomain
0x090000000012bc0c res_nsearch
0x09000000001083d4 res_search
0x090000000013b088 ho_byname2
0x090000000016478c ho_byname2
0x0900000000106970 gethostbyname2
0x0900000000103424 getaddrinfo2
0x0900000000105628 getaddrinfo
0x0900000002037014 sqloPdbTcpIpGetAddrInfo 0x0900000001a2e8f0sqlcctcpconnr__FP15SQLCC_CONNECT_TP17SQLCC_COMHANDLE_TP12SQL 0x090000000213bd5c sqlccconnr__FP15SQLCC_CONNECT_TP17SQLCC_COMHANDLE_T8 0x090000000213b7d4 sqljcCommConnect__FP10sqljCmnMgr
0x09000000021394e4 sqljcCmnMgrInit__FP10sqljCmnMgrP17SQLCC_COMHANDLE_
0x0900000002138e74 sqljrCreateTransport__FP14db2UCconHandle
0x0900000002139114 sqljrAssociateTransport__FP14db2UCconHandlePlbP19SQLE_SRVLST 0x090000000152b2e0 sqljrConnectBasedOnCatalog__FP14db2UCconHandlePibPb
0x090000000152a47c sqljrInternalConnect__FP14db2UCconHandlePbPib
0x0900000001f7e47c sqljrInternalConnect__FP14db2UCconHandlePbPib
0x0900000001f7dc90 sqljrDrdaArAuthenticate__FP14db2UCinterfacelPUi
0x090000000423db98 ICCC_BN_bn2bin@AF391_141
0x0900000001f86b08 sqlexAppAuthenticate__FP14db2UCinterface
0x0900000001f85074 sqljrDrdaArConnect__FP14db2UCinterface
0x0900000001f857ec sqleUCdrdaARinit__FP14db2UCconHandle
0x0900000001fc8dfc sqleUCappConnect
0x0900000001fac268CLI_sqlConnect__FP15CLI_CONNECTINFOP5sqlcaP19CLI_ERRORHEADERINF0x09000000021483ecSQLDriverConnect2__FP15CLI_CONNECTINFOPvPUcsT3T4PsUsUcT9P19
0x0900000002108d98 SQLDriverConnect2__FP15CLI_CONNECTINFOPvPUcsT3T4PsUsUcT9P19
0x0900000002146c50 SQLConnect1__FP15CLI_CONNECTINFOPUcsT2T3T2T3
0x090000000214727c SQLConnect
0x000000010008d044 RdbCreateConnection
0x000000010008b830 DbiGetConnectionTracked
0x000000010014e54c AllocTxnDesc
0x000000010014ea74 DbiParticipate@AF32_6
0x00000001000dfea8 tbOpenX
0x0000000100846868 TsmSqlPrepareData
0x000000010000d5f4 StartThread
Holding mutex txnP->participateMutex (0x156e06fc8), acquired at dbitxn.c(1511) Thread context:
COMMMETHOD: Tcp/Ip
COMMAND: SELECT
THREAD_TYPE: SESSION
SESSION_TYPE: ADMIN
ADMIN_NAME: ADMIN
2. Verify the database connection is using TCP/IP. For example :
Node name = TSMNODE1
Comment =
Directory entry type = LOCAL
Protocol = TCPIP
Hostname = localhost
Service name = 51500
Remote instance name =
System =
Operating system type = None
[/code>]
3. The /etc/hosts file has the localhost entry :
/
4: The /etc/netsvc.conf file has the following bind entry :
[
127.0.0.1 loopback localhost
hosts=local,bind
Resolving The Problem
Modify the /etc/netsvc.conf file to disable IPv6 lookup. For example :
hosts=local4,bind4
Was this topic helpful?
Document Information
Modified date:
17 June 2018
UID
swg21978954