VMware Networking Community
aaronkiki
Enthusiast
Enthusiast

NSX-T manger nodes' /image partition usage reached 100%

Just like the title said, the disk usage of all the 3 manager nodes' /image partition reached 100%.

There are lots of named java_pidXXXX.hprcf  files in the sub-directory ./core under the /image.  I tried to delete all the java_pidxxxx.hprcf files and all the manger nodes works well. But a few hours later, the disk usage of the /image reached 100% again.

Is there anyone has the same problem and how to solve ?

NSX-T version is 3.1.2

Reply
0 Kudos
5 Replies
aaronkiki
Enthusiast
Enthusiast

I downloaded one of the files named java_pidXXXX.hprof under the /image/core.

To open the file by the Eclipe Memory Analyizer and shows:

2021-05-10_00162.jpg

So, the cause of the disk usage of the partition /image increased quickly must out of memory, but I don't know which one caused this.

Reply
0 Kudos
ax98
Contributor
Contributor

I'm the same, and I haven't found a solution

Reply
0 Kudos
ax98
Contributor
Contributor

At present, my processing method is  crontab - e      0 * / 6 * * rm / image / core / Java_*

Add a scheduled command to clean up every 6 hours. I wonder if there is any way to update it.

Reply
0 Kudos
SrVMoussa
VMware Employee
VMware Employee

Hi 

 

Can you please post the output of 

 

/config/corfu/LAYOUT_CURRENT.ds

 

 

Regards,
Khalid Moussa
Reply
0 Kudos
aaronkiki
Enthusiast
Enthusiast

Thank you for your suggestion.

one of the local engineer had suggested me to keep just one node manager. the error disappeared.

Recently, I've upgraded the NSX-T to the newest version and added manger nodes to 3. now it's ok too.