Take the 2-minute tour ×
Server Fault is a question and answer site for professional system and network administrators. It's 100% free, no registration required.

Is there an issue with "Ulimit -Hn" being set too low (at 1024) when (Oracle recommend 65536)? This is for Oracle 64-bit 11g on Linux 5.

It is one of the settings that appears to be woefully short of its recommendation. But I am also aware that the database server in question is an Oracle Data Guard Local Standby and should only really have a connection or two from its Primary database server (to ship the redo logs across).

The Local Standby database server has 'hung' about 3 times in as many months and then requires a reboot. I do not have access to this server, so rely on others to look at logs etc. The sanity check on kernel params uncovered the low value for "ulimit -Hn". Has anyone ever seen that 'low' value cause a hang or crash?

share|improve this question

2 Answers 2

up vote 1 down vote accepted

Linux 5 does not exists. You are using linux 2 or linux 3. Or maybe are you talking about rhel5?

Back to problem : Never seen any kernel crash due to ulimit -Hn output being too low. Merely some softwares won't work correctly or even won't start. IMHO you should raise it but it's not likely to be the root cause of your problem.

Most relational database engines (at least oracle mysql and postgresql) open a lot of files.

share|improve this answer

Does the "hang" apply to Linux kernel, or to Oracle server (i.e. the Oracle RDBMS software)?

I don't expect any hang of the kernel. But such low value of maximum hard limit of number of open file descriptors is really insufficient for standard Oracle physical standby server.

Live example:

ps -ef | grep oracle | wc -l    # how many processes approx. 
      65

lsof | grep oracle | wc -l      # how many files approx. 
    1343

Running out of open file descriptors would result in unpredictable behavior of Oracle including hanging connections to the database or inability to connect as administrator (even sqlplus / as sysdba). It is generally impossible to determine one ORA-xxxx error message for the case, many can indicate this, including ORA-01116, ORA-12535, ORA-00376.

Remember that a standby physical server is only usable if it is able to become the PRIMARY database, and get more processes from clients' connections, in which case things would get really ugly:

ps -ef | grep oracle | wc -l      # how many processes approx. 
     241

lsof | grep oracle | wc -l        # how many files approx.
    7342
share|improve this answer

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.