/Tmp partitions are full, killing oracle rac

I just got back from Shandong and had a bunch of troubles. No, it's about to get off work. The customer says oracle rac is down.

 

650) this. width = 650; "src =" http://www.bkjia.com/uploads/allimg/131229/19233051N-0.jpg "border =" 0 "alt =" "/>

This is a rac system consisting of two nodes and a shared storage. I installed oracle 11g r2. I tried to log on to the system and found that a server cannot be connected. Run df-h and find that a partition/tmp is full. Go to the subdirectory and check that a bunch of files such as 0926. dmp are located at around MB in size. It seems that the programmer is performing the exp backup operation.

 

Log on to another node from the private network of rac and go in. This indicates that the heartbeat check is normal.

 

After communicating with others, I take the following steps:

1. Clear/tmp space, manually delete old files, and release at least 50% of the space.

2. Disable srvctl stop instance-d sery_db-I sery1 and srvctl stop instance-d sery_db-I sery2.

3. Check the network status of the server that cannot be accessed from outside. The check method is as follows:

1) check whether the ip address exists by ip add

2) check whether the network cable is normal with mii-tool

3) service network start restarts the network. Hey hey, it's normal.

4. Check the asm file system by performing the following steps:

1) ps aux | grep asm check process

2) su-grid; asmcmd: Check whether the log file and data file exist. Fortunately, they are all there.

5. Start the instance, and then use sqlplus to log on and check the data files.

 

Major cause: oracle will write data to/tmp partitions as needed, such as sorting and other intermediate processes). If the file system is full, the write will fail and the session may be suspended.

This article is from the sery blog, please be sure to keep this source http://sery.blog.51cto.com/10037/1008477