执行idsicrt创建SDS实例失败：DB2数据库实例初始化出错

Let's break down the issues you're facing—those DB2-related errors during SDS instance creation often stem from resource limits, misconfigured environment variables, or system-level restrictions, even if db2val reports a successful installation. Here are actionable steps to resolve this:

1. Verify and Adjust System Resource Limits (ulimits) #

DB2 requires higher resource limits than RHEL's defaults, especially for file descriptors and process counts.

• Edit /etc/security/limits.conf as root and add/update these lines for the db2inst1 user:

  db2inst1 soft nofile 65536
  db2inst1 hard nofile 65536
  db2inst1 soft nproc 16384
  db2inst1 hard nproc 16384
  

• Switch to the db2inst1 user and confirm the changes with:

  ulimit -n
  ulimit -u
  

• Log out and back in to apply the limits, then retry the idsicrt command.

2. Ensure DB2 Environment Variables Are Loaded #

The idsicrt process relies on proper DB2 environment setup.

• Before running idsicrt, switch to the db2inst1 user and source the DB2 profile:

  . /home/db2inst1/sqllib/db2profile
  

• Verify the instance variable is set correctly:

  echo $DB2INSTANCE
  

  This should return db2inst1 (the default DB2 instance associated with SDS).

3. Tune DB2 Database Manager Configuration #

Incorrect DB2 manager settings can cause initialization timeouts.

• As db2inst1, check current configuration:

  db2 get dbm cfg
  

• Focus on these key parameters:
  • SVCENAME: Confirm the DB2 service port is correctly set and not in use (check with netstat -tulpn | grep db2).
  • MAXAGENTS: Increase this to at least 200 if it's lower:

    db2 update dbm cfg using MAXAGENTS 200
    

  • INSTANCE_MEMORY: Set it to automatic to let DB2 allocate memory dynamically:

    db2 update dbm cfg using INSTANCE_MEMORY AUTOMATIC
    

• Apply changes by stopping and restarting DB2:

  db2 force applications all
  db2 stop
  db2 start
  

4. Check Firewall and SELinux Restrictions #

RHEL's firewalld or SELinux might block DB2/SDS port communication.

• Firewall: Temporarily disable it to test:

  systemctl stop firewalld
  

  If this fixes the issue, add permanent rules for your ports:

  firewall-cmd --add-port=389/tcp --permanent
  firewall-cmd --add-port=686/tcp --permanent
  firewall-cmd --add-port=3538/tcp --permanent
  firewall-cmd --add-port=3539/tcp --permanent
  firewall-cmd --reload
  

• SELinux: Temporarily set it to permissive mode:

  setenforce 0
  

  If successful, update /etc/selinux/config to set SELINUX=permissive and restart the system to make it permanent.

5. Clean Up Residual Configuration and Reinitialize #

Partial rollbacks from failed attempts can leave invalid state.

• Stop all DB2 processes:

  db2 force applications all
  db2 stop
  

• Drop the incomplete SDS instance:

  idsidrop -I idsldap
  

• Drop the DB2 instance (backup any existing data first):

  db2idrop db2inst1
  

• Recreate the DB2 instance:

  db2icrt -u db2fenc1 db2inst1
  

• Now re-run your original idsicrt command to create the SDS instance.

6. Dig Deeper with Logs #

If the issue persists, check detailed logs for clues:

• DB2 trace logs: Look in /home/db2inst1/sqllib/db2dump for trace files and error logs (filter for DBI1281E or timeout-related entries).
• System logs: Check /var/log/messages for permission errors, resource exhaustion, or DB2-related failures.

内容的提问来源于stack exchange，提问作者Marco Ordoñez Mendoza