Djelibeybi
commented
5 years ago @sjvp could this be caused by the changed to harden the middleware/fmw-infrastructure image?
Closed marvinamorim closed 5 years ago
Djelibeybi
commented
5 years ago @sjvp could this be caused by the changed to harden the middleware/fmw-infrastructure image?
sjvp
commented
5 years ago What DB are you using? Anything of interest in the DB logs?
Is there a more complete stack trace in /u01/oracle/.inventory/logs/config2019-05-18_11-44-25AM/wlst_essbase.log?
Is it reproducible for you?
The most likely cause if you are using non-RAC Oracle DB is lack of entropy. Typical error in the log files are:
Root Cause: java.sql.SQLRecoverableException: IO Error: Connection reset by peer
com.oracle.cie.domain.security.DomainSecurityService - Time required to process encryption was longer than expected
So check you have enough entropy. Official docs are: https://docs.oracle.com/middleware/1221/wls/WLUPG/compat.htm#WLUPG650.
marvinamorim
commented
5 years ago Hello.
But reading wlst_essbase.log, there is the line: "com.oracle.cie.domain.security.DomainSecurityService - Time required to process encryption was longer than expected: 336477ms; possible lack of entropy might result in poor performance (getEncryptionContext)". Going to check this entropy matter with the company that is responsible for our database, and going to check the official docs link provide by you.
Thanks a lot for the response.
sjvp
commented
5 years ago As you have found, the logs indicate a lack of entropy.
This is on the client side (BI/Weblogic running in docker on your host) - it is not a problem with the database.
I think it is the host's /dev/random that it is used for docker containers. See: https://stackoverflow.com/questions/52699235/how-are-requests-to-dev-urandom-etc-handled-in-docker
But 'host' could be the hyperkit VM if using docker-for-mac, boot2docker etc. See: https://stackoverflow.com/questions/26021181/not-enough-entropy-to-support-dev-random-in-docker-containers-running-in-boot2d
marvinamorim
commented
5 years ago So I see. Going to check those threads on stackoverflow and look a way to solve this problem. If I solve this, I give a feedback here.
Again, thank's a lot for the information.
marvinamorim
commented
5 years ago Hello again.
So we tried manual installation, and got stuck at the same point. But we solved by changing the line: securerandom.source=file:/dev/random to: securerandom.source=file:/dev/urandom After this, everything went smooth.
Where in the process of the installation of the 3 dockerfiles required by this guide (installation of OracleJava, OracleFMWInfrastructure, OracleBI) can I find the file java.security?
marvinamorim
commented
5 years ago So, I got the BI working, but only a part of it.
What I did was:
On the Dockerfile: docker-images/OracleFMWInfrastructure/dockerfiles/12.2.1.3/Dockerfile, i changed the env USER_MEM_ARGS from "-Djava.security.egd=file:/dev/./urandom" to "-Djava.security.egd=file:/dev/urandom".
And on the Dockerfile: docker-images/OracleBI/dockerfiles/12.2.1.3/Dockerfile, I add at the last line:
COPY java.security /usr/java/jdk1.8.0_211/jre/lib/security/java.security, where my local java.security had changed the line securerandom.source=file:/dev/random to securerandom.source=file:/dev/urandom.
Everything went fine, until the time to start the services.
The console on port 9500 is working fine. But the services on port 9502 are returning error 500 The server encountered an unexpected condition which prevented it from fulfilling the request..
Any idea why only the console on port 9500 is working? Here follows the Start Services log attached start_services.log.
And thanks very much for everything.
sjvp
commented
5 years ago The Dockerfiles are samples and so you can/should make modifications as needed - this is what you have done to get it working for you, hence closing this issue.
You should raise a new issue for the latest, unrelated problem.
However to get you moving...
Please check if the following error appears in /u01/oracle/user_projects/domains/bi/servers/*/logs/*:
error while loading shared libraries: libaio.so.1: cannot open shared object file: No such file or directory
If so, it appears to have already been fixed by this commit to add libaio back in: https://github.com/oracle/docker-images/pull/1293/commits/9c8cc2cfc0bc49998b30ccde677bfdb547350ab5
So - pull the changes and rebuild your FMW Infra image and then the BI image.
You may find that it fails due to Unable to create directory /u01/oracle/user_projects/domains. We are working on this but for now you could add RUN rmdir /u01/oracle/user_projects to the BI Dockerfile (not tested this, but should work if the base FMW Image has stopped creating this as a VOLUME).
marvinamorim
commented
5 years ago Everything is working now.
To solve the libaio problem I just switched the line that installed libaio to after the oracle user was set on the Dockerfile from the FMW, and everthing went fine, we are already using the BI with everything that we need.
Thanks a lot for your help @sjvp, now we can finaly work on our BI. See you around.
Hello.
Yesterday I started to deploy OracleBI on docker, and all the dockerfiles build went out OK. But when I tryed the docker run, I got an error on the step:
Anyone knows what the problem is? I can connect to the oracle DB_HOST with the password that I provied, and I'm using an alphanumeric password on the environment variables ADMIN_PASSWORD and SCHEMA_PASSWORD, and the schemas are normaly created on my DB_HOST. Thanks in advance.
Here is the output of the process:
Edited for formatting.