I have already blogged on similar issue, but decided to do it once again because symptoms are the same, but the source reason is different.
So, one of our customers asked for our help in situation when Linux server rebooted by them-self and, because it was hosting standby database, it was quite urgent to get it working properly to provide protection for primary database.
I have just started managed recovery process and everything was OK for a couple of minutes, but then … console hang and server just went down for reboot… interesting 😉
Just to test if it’s consistent, I have done it once more and… it rebooted once again after applying several archive logs… 😉
Simple question from previous post: how Starting Oracle MRP may reboot Linux server ?
The answer was inside system log file(/var/log/messages) – thanks the Luck that Linux was able to write next messages before server reboot:
Jun 20 15:37:56 myserver kernel: Uhhuh. NMI received. Dazed and confused, but trying to continue
Jun 20 15:37:56 myserver kernel: You probably have a hardware problem with your RAM chips
Jun 20 15:40:44 myserver syslogd 1.4.1: restart.
So we sew mentioned server reboot because of bad RAM chip, but how is this fact connected with starting MRP ?
It’s quite simple – because starting MRP starts using/addressing more memory, particularly for Oracle buffer cache, and eventually touches bad memory range.
BTW, it was different server and Oracle version was 18.104.22.168 this time, but it doesn’t depends on it….