Nate created KARAF-6695:
---------------------------
Summary: Karaf 4.2.8 installed as a service on CentOS 7 uses POSIX
as the locale instead of UTF-8 like in 4.0.7
Key: KARAF-6695
URL: https://issues.apache.org/jira/browse/KARAF-6695
Project: Karaf
Issue Type: Bug
Components: karaf
Affects Versions: 4.2.8
Environment: CentOS 7 Core
[user@hostname bin]$ hostnamectl
Static hostname: hostname
Icon name: computer-vm
Chassis: vm
Machine ID: f71a935b20b249c8bdfdbc7d0dfd0c99
Boot ID: d4936791e0724265aa940fdfcf5e618e
Virtualization: kvm
Operating System: CentOS Linux 7 (Core)
CPE OS Name: cpe:/o:centos:centos:7
Kernel: Linux 3.10.0-1062.12.1.el7.x86_64
Architecture: x86-64
Reporter: Nate
I'm testing Karaf 4.2.8 running as a service on CentOS 7 using the
karaf:wrapper feature. A problem I encountered is that in this new version,
SSHing into the console and running
{code:java}
exec locale{code}
returns POSIX as the locale for everything:
{code:java}
karaf@root()> shell:exec locale
LANG=
LC_CTYPE="POSIX"
LC_NUMERIC="POSIX"
LC_TIME="POSIX"
LC_COLLATE="POSIX"
LC_MONETARY="POSIX"
LC_MESSAGES="POSIX"
LC_PAPER="POSIX"
LC_NAME="POSIX"
LC_ADDRESS="POSIX"
LC_TELEPHONE="POSIX"
LC_MEASUREMENT="POSIX"
LC_IDENTIFICATION="POSIX"
LC_ALL={code}
. This is different from the locale of the user that the karaf server is
running as (which is en-us.UTF-8 or something like that), and also different
from the value in karaf 4.0.7 running as a service, the version I upgraded from
(also UTF-8) and also different from running it using /bin/karaf-server.sh
(also UTF-8). For me, this causes an issue when doing a svn checkout from a
karaf feature/bundle, because the svn repository is in UTF-8 and Subversion
can't deal with the client and server having different locales. I imagine there
will be other issues as well with other programs I'd like to run from Karaf.
For now, I can work around it by manually changing my locale through an exec,
but I can't really ask that of my end users. I feel like it's not really
intended that Karaf runs under a different locale based on whether it's been
started as a service or not. I also think this might be a regression. At any
rate, even if it's intended, I'm hoping there is a workaround like a startup
script I can edit so the locale is always UTF-8.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
