[Bugs] [Bug 1248123] New: writes to glusterfs folder are not synced to other nodes unless they are explicitly read from gluster mount

[bugzilla at redhat.com]

https://bugzilla.redhat.com/show_bug.cgi?id=1248123

            Bug ID: 1248123
           Summary: writes to glusterfs folder are not synced to other
                    nodes unless they are explicitly read from gluster
                    mount
           Product: GlusterFS
           Version: 3.7.2
         Component: replicate
          Assignee: bugs at gluster.org
          Reporter: cmptuomp3 at gmail.com
                CC: bugs at gluster.org, gluster-bugs at redhat.com



Description of problem: writes to glusterfs folder are not synced to other
nodes unless they are explicitly read from glusterfs mount

for example, I can write 1000 files in a folder in the glusterfs mount, then if
I ssh to another server and query the actual brick, the files are not there, if
I `cat` each file from the glusterfs mount - files appear in the brick mount.
Aren't files supposed to be automatically copied to the brick? What if
meanwhile the server on which I performed the writes goes down, are they lost?


Version-Release number of selected component (if applicable): 3.7.2-3.el7


How reproducible: very


Steps to Reproduce:
1. touch /ssd_data/test.file on 10.0.1.3 (ssd_data is glusterfs mount)
2. try and open /ssd_storage/test.file on 10.0.1.2 (not found) (ssd_storage is
the brick mount)

Actual results:
 file not found

Expected results:
 file is present



Additional info:

Number of volumes 2
Volume Names ssd-gluster-data, sata-gluster-data
Type of volumes replica 3

Output of gluster volume info

gluster volume info

Volume Name: sata-gluster-data
Type: Replicate
Volume ID: dbb498cc-fa8f-4513-8d6c-1aa9a159b7cf
Status: Started
Number of Bricks: 1 x 3 = 3
Transport-type: tcp
Bricks:
Brick1: 10.0.1.3:/sata_storage/GLUSTER_DO_NOT_MODIFY
Brick2: 10.0.1.2:/sata_storage/GLUSTER_DO_NOT_MODIFY
Brick3: 10.0.1.1:/sata_storage/GLUSTER_DO_NOT_MODIFY
Options Reconfigured:
nfs.disable: off
cluster.server-quorum-type: server
cluster.server-quorum-ratio: 60

Volume Name: ssd-gluster-data
Type: Replicate
Volume ID: 967b2328-c5d3-4ea1-8cd0-8114796f3b50
Status: Started
Number of Bricks: 1 x 3 = 3
Transport-type: tcp
Bricks:
Brick1: 10.0.1.3:/ssd_storage/GLUSTER_DO_NOT_MODIFY
Brick2: 10.0.1.2:/ssd_storage/GLUSTER_DO_NOT_MODIFY
Brick3: 10.0.1.1:/ssd_storage/GLUSTER_DO_NOT_MODIFY
Options Reconfigured:
nfs.disable: off
cluster.server-quorum-type: server
cluster.server-quorum-ratio: 60

Output of gluster volume status

gluster volume status
Status of volume: sata-gluster-data
Gluster process                             TCP Port  RDMA Port  Online  Pid
------------------------------------------------------------------------------
Brick 10.0.1.3:/sata_storage/GLUSTER_DO_NOT
_MODIFY                                     49153     0          Y       16358
Brick 10.0.1.2:/sata_storage/GLUSTER_DO_NOT
_MODIFY                                     49155     0          Y       8691 
Brick 10.0.1.1:/sata_storage/GLUSTER_DO_NOT
_MODIFY                                     49154     0          Y       24495
NFS Server on localhost                     2049      0          Y       8705 
Self-heal Daemon on localhost               N/A       N/A        Y       8713 
NFS Server on 10.0.1.1                      2049      0          Y       24509
Self-heal Daemon on 10.0.1.1                N/A       N/A        Y       24517
NFS Server on 10.0.1.3                      2049      0          Y       30660
Self-heal Daemon on 10.0.1.3                N/A       N/A        Y       30699

Task Status of Volume sata-gluster-data
------------------------------------------------------------------------------
There are no active volume tasks

Status of volume: ssd-gluster-data
Gluster process                             TCP Port  RDMA Port  Online  Pid
------------------------------------------------------------------------------
Brick 10.0.1.3:/ssd_storage/GLUSTER_DO_NOT_
MODIFY                                      49152     0          Y       16366
Brick 10.0.1.2:/ssd_storage/GLUSTER_DO_NOT_
MODIFY                                      49152     0          Y       8698 
Brick 10.0.1.1:/ssd_storage/GLUSTER_DO_NOT_
MODIFY                                      49153     0          Y       24502
NFS Server on localhost                     2049      0          Y       8705 
Self-heal Daemon on localhost               N/A       N/A        Y       8713 
NFS Server on 10.0.1.1                      2049      0          Y       24509
Self-heal Daemon on 10.0.1.1                N/A       N/A        Y       24517
NFS Server on 10.0.1.3                      2049      0          Y       30660
Self-heal Daemon on 10.0.1.3                N/A       N/A        Y       30699

Task Status of Volume ssd-gluster-data
------------------------------------------------------------------------------
There are no active volume tasks

fstab:

10.0.1.3:/ssd-gluster-data        /ssd_data    glusterfs   
defaults,_netdev,backup-volfile-servers=10.0.1.1:10.0.1.2    0 0
10.0.1.3:/sata-gluster-data        /sata_data    glusterfs   
defaults,_netdev,backup-volfile-servers=10.0.1.1:10.0.1.2    0 0


Client Information
OS Type: Linux
Mount type: fuse.glusterfs
OS Version: CentOS 7, x86_64

-- 
You are receiving this mail because:
You are on the CC list for the bug.
You are the assignee for the bug.