0

I am trying to copy data files from my VM to a NFS VM- ZFS Storage(Both VM's can talk to each other). During copy sometimes I encounter error:

INFO: task cp: blocked for more than 120 seconds . "echo 0 > /proc/sys/kernel/hung_task_timeout_secs disables this message"

Both my VM's hang and I have to restart them. If I copy again it works.

I have around 233 data files to copy and its becoming difficult to restart VM's again and again.

I looked at the solutions given on internet and changed the vm.dirty_ratio to 5 and vm.dirty_background_ratio to 10 to resolve but it did not work.

I am running these VM's on virtual box and allocated around 17GB RAM for one and the NFS VM around 6GB RAM.

Any hack which could help me in copying these files to the NFS without my VM's hanging?

user1004779
  • 147
  • 1
  • 3
  • 13

1 Answers1

0

I am sorry if I am answering an answer with another answer, but this case has many variables that need exploring.

1, you have a Linux VM sharing your storage (assumption)

A. Which distro ? 32 or 64 bits ? When the problem happens what does top reports for system load ?

B. Local storage or nas ? Or San ?

C. Which version of NFS ? 3 or 4 ?

D. Can you set the variables of your mount when mapping the NFS share? You might want to play with rsize and wsize, setting them to at least 64000. I would recommend also setting noatime and nodiratime on the share.

E. From my VMware background with Gluster, there are some timeout/refresh settings you can set on the storage side. How often the storage publishes its presence, telling it is alive. A good start is 20 seconds.

F. VMware can tell you how much latency you have for read or write on a physical and on a VM level. Try to figure out those to know who to blame.

Ah, and, of course, make sure your Linux VM has the latest patches applied.

Let's see where we get from here.

Carlos Cap
  • 61
  • 3