Category Archives: Errors

ORA-00845: MEMORY_TARGET not supported on this system

I have a RH5.0 Linux server that has oracle11g server running on it. I have just created a new database and when I try to start it I’m getting the following error message:-

ORA-00845: MEMORY_TARGET not supported on this system

I have increased the size of /dev/shm to 50GB and I’m still getting this error message, what can I do to resolve this issue?

In /etc/fstab i have made the following change:-

tmpfs                   /dev/shm                tmpfs   defaults,size=50G        0 0

and then did the following:-

mount -o remount /dev/shm

Or do I actually need to physically reboot the server?

karluk (MIS)
12 May 09 11:10
What is the output of a “df -k” command?  That should tell you whether /dev/shm is mounted and its current size.
nyck (TechnicalUser)
13 May 09 5:15

The output is below:-

tmpfs                  50G   26G   25G  52% /dev/shm

karluk (MIS)
13 May 09 14:02
My only 11g databases are on SLES 10, so I don’t have a directly comparable setup to yours.  For me automatic memory management worked without a hitch.  For comparison, this is what I see on the O/S.


Filesystem           1K-blocks      Used Available Use% Mounted on
shmfs                  8388608    300560   8088048   4% /dev/shm

How much RAM do you have on your server?  50G seems like an awfully lot to allocate to shared memory, but it certainly looks as if your system is allowing it.

karluk (MIS)
13 May 09 14:42
The following is a link to an interesting article on Oracle 11g automatic memory management.  The author claims that /dev/shm has to be configured to be at least as large as MEMORY_MAX_TARGET.  If true and you set this parameter high enough, I suppose it’s possible that you simply don’t have a large enough /dev/shm.  Also, there is apparently an alert log entry when you get the ORA-00845 error.  What message is in your alert log?  Did it say how much shared memory Oracle thinks is available?

nyck (TechnicalUser)
14 May 09 4:02
I finally got round to solving this issue. When creating the databases using dbca I did not notice that the memory being allocated to the db’s was 40% of the physical RAM ( this is 64GB ). I created four of these databases like this before I noticed what was going on. To resolve this I perfromed the following from within each database:-

Alter system set memory_target= 2000M scope=spfile;

then bounced the database and all was fixedsmile