== DEBUGGING CORE FILES  == WHAT IS A CORE FILE?
A core file is the (almost complete) image of the running process in memory.
Core files can be created on demand using gencore, gcore or other commands
but most of the time they are the result of a fault in the executing program.
They are useful for finding what was wrong at the time a process crashed
because they provide us with access to most of the data that the program
was using when it crashed.
While the internal format of a core file will depend based on the machine
Can I use online backup to encrypt an existing database ? The answer is YES . Converting a database to be encrypted is a kind of big change. Therefore, I still strongly suggest to take an offline backup before the change if business permits the outage time to do that. On the DB2 knowledge center page " Encrypting an existing database ", it shows the example steps using an offline backup. But that does not necessarily mean we can only use the offline backup. On this blog, I will show some sample steps using online... [More]
Recently following simple guidelines helped customers to get rid of corrupted pages easily. So, just putting it here in case it helps. Customer's db2 inspect output was showing following kind of corruptions, DATABASE: TESTDB VERSION : SQL09053 2018-02-01-14.45.44.077383 Action: CHECK TABLE Schema name: DB2INST1 Table name: MY_TAB Tablespace ID: 3 Object ID: 580 Result file name: badpage.out Table phase start (ID Signed: 580,... [More]
Question: Why does db2icrt fail on Linux? It was able to run without any error and an instance was created one year ago. Answer: Here is the steps and result. As root user: --- [root]# useradd testuser [root]# su - testuser [testuser]$ exit [root]# /opt/ibm/db2/V10.5/instance/db2icrt -d -u testuser testuser DBI1446I The db2icrt command is running. DBI1264E This program failed. Errors encountered during execution were written to the installation log file. Program name: db2icrt. Log file name: ... [More]
It is a HADR system set up with near_sync mode. This server has been running well for several years with version V9.7 FixPack7 even though the user already know the storage on standby side could be a bottleneck. At the beginning of this year 2018, the user upgraded Db2 to V10.5, the problem happened since then. The log gap kept on increasing, and standby can never catch up with primary. Collected stacks from standby, found hadrs EDU was busy on writing log to disk: 0x0900000000037B18 pwrite64 + 0x38 0x090000002817CF4C... [More]
Sometimes db2 stack traces on Solaris can be misleading as the function
name we see on the stack is not the name of the function that was called.
This might happen when a 'static' function is called. It has to do with
the fact that 'static' functions do not have 'global' entries in the
For example, if you have a source file containing functions like this:
static int function2()
If some function calls 'function2()'... [More]
As an experienced DB2 DBA, you are familiar with HADR and applying DB2 fixpack by rolling maintenance. Now pureScale with HADR clusters system. What are you going to do ? From DBA’s perspective, it’s always better to have enough service down maintenance time window. But if your critical business does not permit this way for 24/7 operation requirement, we will have to implement appropriate steps without service interruption. So I’m writing this blog assuming you have to complete the mission without taking down... [More]
We have seen some cases recently where the we see following messages frequently written into the db2diag.log for purescale environments : 2018-01-10-03.35.00.656437+540 I8443A1349 LEVEL: Error PID : 18153532 TID : 3193 PROC : db2sysc 1 INSTANCE: db2inst1 NODE : 001 HOSTNAME: host0005 EDUID : 3232 EDUNAME: db2pdbc 1 FUNCTION: DB2 UDB, fast comm manager, sqkfChannel::DeliverInboundBuffer , probe:4717 DATA #1 : String, 51 bytes Invalid Sequence No. Detected = 2. Expected No. = 1 DATA #2 :... [More]
When your SSL certificate is about to expire or already expired you generate/create a new one. To make this new/revised certificate effective and for DB2 to use it, DB2 needs to be recycled (db2stop/ db2start) . References : Infocenter link : https://www.ibm.com/support/knowledgecenter/en/SSEPGG_11.1.0/com.ibm.db2.luw.admin.sec.doc/doc/t0025241.html Step by step instructions : Secure Sockets Layer (SSL) support in DB2 for Linux, UNIX, and Windows... [More]
In general, the KMIP config file is read on each major command to retrieve a key from the keystore. But If Db2 is still attempting to execute the first command, unsuccessfully, the default number of retry times. Every time Db2 requires the master key it makes the trip to the server. A single Db2 command may result in multiple trips to the keystore. The kmip.cfg is only read when a new trip to the keystore is initiated, NOT a retry because the first trip failed. Once max retries has been reached, the new kmip.cfg values come into effect... [More]
If we update configuration file for Db2 with wrong IP address or Port,i.e not the IP and Port Key-store is configured on.We may notice Connection to Database appear as hang. We are actually experiencing a 'connection timeout' issue that is taking a while due to multiple attempts to reconnect. By default retry value set to the default 50, and each reconnect takes time based upon network settings to report a timeout,then it will take 50x time window to go through the attempts, during which the system will appear like it is doing... [More]
In case Master Node is down and Clone is set up in cfg file. We can notice few of the Read only Query to Keystore goes to Clone server even after Master Node is up. In v22.214.171.124 we introduced a “blacklist” that servers are placed on when we get an error from them. The default time they are on the blacklist is 5 minutes, after which we will try them again. There is no command to reset the black list. Any commands that require creating a new key, such as create database, are sent to the Master key manager... [More]
A key rotation on the primary will drive a key rotation automatically on the standby. The key rotation on the standby is not instantaneous. The key rotation is written on regular log records, and is not sent to the standby until a other log records are sent. So it’s possible and expected that you could do a key rotation on the primary, and then immediately check for the key rotation on the standby and not see it. Some write activity at the primary is required to drive sending the log records. There is no explicit command to force... [More]
The same master key must be used on both primary and standby. A key rotation on the primary will drive a key rotation automatically on the standby. It’s possible that they can start out with different keys, but they will come into sync during the first key rotation. We can not keep different Key in Primary and Standby as a long term solution.
Our recommendation is that both primary and standby are encrypted. However, running with an encrypted primary and a non-encrypted standby is supported, but only for enabling Native encryption in existing setup without complete outage.(i.e. Online implementation) It’s not intended to be a long term solution. if we enforced that both primary and standby had to be encrypted, you would not be able to use this online method of enabling encryption. You’d have to shutdown both primary and standby (whole system offline), enable... [More]