Host: esxi hypervisor, on it 3 datastores (3 reid arrays) Every 20min eats 1m on one of the datastores. VM configuration, the disk is thick, that is, the place is allocated immediately. I thought that the problem was in the log file as it was very fast and I decided to delete it. In the hope that the place will cease to decrease. But still the situation continued.

/vmfs/volumes/51808b6b-b7ae6443-e788-10bf484ee5cb # du -sh

Datastor capacity: 470.9G

 Filesystem Bytes Used Available Use% Mounted on VMFS-5 510832672768 506097303552 4735369216 99% /vmfs/volumes/dc

and after a period of time

 Filesystem Bytes Used Available Use% Mounted on VMFS-5 510832672768 506098352128 4734320640 99% /vmfs/volumes/dc

through the console

 VMFS-5 475.8G 471.3G 4.4G 99% /vmfs/volumes/dc

through graphic tooling

enter image description here

virtualka

 /vmfs/volumes/51808b6b-b7ae6443-e788-10bf484ee5cb/dc # ls -lha total 493002784 drwxr-xr-x 1 root root 1.8K Apr 4 09:27 . drwxr-xr-t 1 root root 1.2K Dec 19 2014 .. -rw-r--r-- 1 root root 27 Jul 17 2015 dc-06badfdf.hlog -rw------- 1 root root 0 Mar 29 18:03 dc-06badfdf.vswp -rw------- 1 root root 470.0G Apr 5 06:37 dc-flat.vmdk -rw------- 1 root root 8.5K Apr 4 14:31 dc.nvram -rw------- 1 root root 492 Mar 29 18:05 dc.vmdk -rw------- 1 root root 0 Nov 24 2013 dc.vmsd -rw------- 1 root root 4.2K Mar 31 17:48 dc.vmx -rw------- 1 root root 0 Mar 29 18:03 dc.vmx.lck -rw------- 1 root root 3.2K Jul 17 2015 dc.vmxf -rw------- 1 root root 4.2K Mar 31 17:48 dc.vmx~ -rw------- 1 root root 166.0M Mar 29 18:03 vmx-dc-112910303-1.vswp

In a virtual machine, nothing concrete is noticed, that is, nothing grows. But the place is leaving. I can not understand what the problem is. I will be glad to any tips. Thank you.

  • one
    And what is the log file that you deleted and most importantly, did you restart the process that wrote this log? because in linux you can delete the open file, but this only removes the directory entry, the inode continues to be used and grow until all the processes that have opened this file are completed - Mike
  • I deleted: vmware.log in the virtual machine directory. The process did not restart. How can he grow if he has nowhere to write? - ivanzor
  • Record goes to inode. And the file name in the directory is only a pointer to the inode on which the file can be opened. find the pm of the vmvare process, go to the / proc / pid process / fd / and give ls -l there or another open file descriptors, I think you will see your log there and even the tail command will make sure that the file is still being written - Mike
  • Thanks for the tip. Through lsof, I managed to see that the processes for writing log files were running. But since this host is spinning 3 VMs, most likely I saw the very processes that are written to the other 2 VMs. What I saw in / proc, a couple of directories: bus, driver, mpt, net, scsi. - ivanzor

0