6

I'm using GlusterFS 3.7.9, currently on a single server with 4 bricks.

Each brick has 4TB and the volume is set up as distribute only.

The volume is mounted on a secondary server and I use it for backup purposes.

The problem is that once every couple of days the secondary server sees the volume smaller (12TB) and when it tried to access some of the files I get transport endpoint not connected.

It's like it stops seeing one of the 4 bricks, while the server where the volume is sees them all without a problem.

I have the IPs in the firewall whitelist, gluster volume status says everything is fine.

What should I check? I can't find any reason for this.

Alex Dumitru
  • 315
  • 2
  • 3
  • 8
  • 1
    You should check all of your logs for filesystem errors and run checks on your bricks. I'm not certain what you're using as a brick, but you should check for failing disks if that is applicable in your situation. Also check the filesystems that constitute the bricks. This can be done online with read-only checks using most filesystems. Gluster has a few log commands of its own, detailed in the man page: https://linux.die.net/man/8/gluster – Spooler Oct 04 '16 at 12:02
  • 1
    Each brick is a 4TB HDD. I don't think there's anything wrong with the disks, because the volume works perfectly if mounted on the server it's on. The problem is only on a remote server. – Alex Dumitru Oct 04 '16 at 12:09

1 Answers1

-1

The way to solve this is to remount the mountpoint:

sudo umount /mntpoint
sudo mount /mntpoint

user2219975
  • 205
  • 2
  • 4