Description of problem:
The logrotate configuration included with the GlusterFS client packages for Ubuntu doesn't send SIGHUP to the clients as part of its postrotate script:

/var/log/glusterfs/*.log {
…
        rotate 7
…
        postrotate
                [ ! -f /var/run/glusterd.pid ] || kill -HUP `cat /var/run/glusterd.pid`
        endscript
}

Cient logs are written to /var/log/glusterfs/….log. Upon rotation the files are renamed, for example to /var/log/glusterfs/srv-shared-foobar_test-var.log.1. In our case that file was still opened by the client:

$ ls -lh /proc/29872/fd/5
l-wx------ 1 … /proc/29872/fd/5 -> /var/log/glusterfs/srv-shared-foobar_test-var.log.1

After 7 rotations log messages end up being written to a file unreachable via the normal file system (could still be read via /proc/…/fd/…).

The logrotate configuration on CentOS/RHEL looks as follows:
/var/log/glusterfs/*.log {
…
  postrotate
  /usr/bin/killall -HUP glusterfs > /dev/null 2>&1 || true
  /usr/bin/killall -HUP glusterd > /dev/null 2>&1 || true
  endscript
}

Expected behaviour:
Postrotate script sends SIGHUP to Gluster client and client reopens log file.

Reproduced using these package versions:
glusterfs-client 3.8.12-ubuntu1~xenial1
glusterfs-client 3.10.3-ubuntu1~xenial1