0

We have 2 servers on AWS, one running Windows 2012R2 Datacenter, the other an Amazon Linux 2017.03, each in a different VPC.

The Windows server is exporting a folder over SMB, that is mounted on the Linux server through samba using version 3.02 of the SMB protocol. The SMB ports are allowed for communication between the 2 instances, and this is working fine. Here is the output of the mount command: //XXX.XXX.XXX.127/pim/ on /pim type cifs (rw,relatime,vers=3.02,sec=ntlmssp,cache=strict,username=censhareuser,domain=W2K12-DATARELAY,uid=861,forceuid,gid=861,forcegid,addr=XXX.XXX.XXX.127,file_mode=0777,dir_mode=0777,nounix,mapposix,rsize=1048576,wsize=1048576,echo_interval=60,actimeo=1)

We recently noticed that from time to time, the share is seen as unmounted for a few seconds on the Linux instance, before being seen as mounted once again. As the echo_interval is set to 60s, I understand why in case of unavailability the mount point would become available again. However I fail to see how it would be seen as unmounted for just a few seconds, as I don't think a share would be unresponsive long enough to be unmounted while still triggering the echo happening every 60s.

Any idea? mount point disappear for a few seconds

Christophe Garrigue
  • 11
  • 3
  • have you seen this: https://serverfault.com/questions/622238/linux-cifs-samba-mount-hangs-for-several-minutes Might be helpful! – batistuta09 Mar 07 '18 at 10:13
  • Did you check other places when it is appears "not mounted"? It disappeared from /proc/mounts (aka "mount" output)? "ls -a" is empty, while the share is not? The icon in some GUI changes? As it this is kernel SMB client, I also suggest checking dmesg. – Nikita Kipriyanov Mar 07 '18 at 11:18
  • @batistuta09 I looked at that question before creating this one. I'm pondering using tcpdump to see if any packets are dropped, but do the issue happening irregularly it might imply creating a huge quantity of log. Which is why I'm keeping this solution as a last resort. – Christophe Garrigue Mar 07 '18 at 11:45
  • @NikitaKipriyanov Nothing reported by dmesg, that's one of the first thing I looked at. However I did not look at /proc/mounts or perform any other tests at that time, neither did the collegue who reported the issue who reported the issue to me. Something to look at next time. This is a headless server by the way. – Christophe Garrigue Mar 07 '18 at 11:46

0 Answers0