Join the Community
and take part in the story

Failed Link an rdir to a rawx


#1

Hi All, i’m just trying to install OpenIO on a three virtualized cluster just to have a test environment. I’ve followed all instructions up to the point where i need to bootstrap directory. But what i’ve got is this…

_Meta1 prefix mapping already initialized
9768 7F5DB5643550 log WARNING Failed to link an rdir to rawx 10.10.1.130:6004: found only 0 services matching the criteria: no service polled from [rdir], 1/2 services polled, 3 services in slot (HTTP 500) (STATUS 481)
9768 7F5DB5643550 log WARNING Failed to link an rdir to rawx 10.10.1.120:6004: found only 0 services matching the criteria: no service polled from [rdir], 1/2 services polled, 3 services in slot (HTTP 500) (STATUS 481)
9768 7F5DB5643550 log WARNING Failed to link an rdir to rawx 10.10.1.126:6004: found only 0 services matching the criteria: no service polled from [rdir], 1/2 services polled, 3 services in slot (HTTP 500) (STATUS 481)
Several errors encountered: {ClientException(): [‘10.10.1.126:6004’], ClientException(): [‘10.10.1.120:6004’], ClientException(): [‘10.10.1.130:6004’]}
_

All nodes are up and running… What i’m missing?
[root@localhost ~]# gridinit_cmd status
KEY STATUS PID GROUP
OPENIO-account-0 UP 642 OPENIO,account,account-0
OPENIO-beanstalkd-0 UP 894 OPENIO,beanstalkd,beanstalkd-0
OPENIO-conscience-0 UP 1057 OPENIO,conscience,conscience-0
OPENIO-conscienceagent-0 UP 643 OPENIO,conscienceagent,conscienceagent-0
OPENIO-meta0-0 UP 1658 OPENIO,meta0,meta0-0
OPENIO-meta1-0 UP 641 OPENIO,meta1,meta1-0
OPENIO-meta2-0 UP 1062 OPENIO,meta2,meta2-0
OPENIO-oio-blob-indexer-0 UP 636 OPENIO,oio-blob-indexer,oio-blob-indexer-0
OPENIO-oio-event-agent-0 UP 637 OPENIO,oio-event-agent,oio-event-agent-0
OPENIO-oioproxy-0 UP 1058 OPENIO,oioproxy,oioproxy-0
OPENIO-rawx-0 UP 1056 OPENIO,rawx,rawx-0
OPENIO-rdir-0 UP 1060 OPENIO,rdir,rdir-0
OPENIO-redis-0 UP 1061 OPENIO,redis,redis-0
OPENIO-redissentinel-0 UP 1059 OPENIO,redissentinel,redissentinel-0
OPENIO-zookeeper-0 UP 632 OPENIO,zookeeper,zookeeper-

Thanks in advance.


#2

Hello @warfox,

You need to make sure that all your machines have unique hostnames. This is due to the way load balancer locations work in the default install. Once done changing the hostnames, apply the puppet file, and retry the bootstrap once again.


#3

Thanks Vladimir. Now i’ve got a cluster online and perfectly working.

Got a problem with puppet which seems to be updated in version and missing --no-stringy-facts for Swift installation.
Is there a workaround?


#4

Hello @warfox,

Have you tried downgrading your puppet version?


#5

As stated by @Philippe, puppet gets updated to major version 4 when OpenStack repos are installed, which we don’t support. To fix the problem, do the following:

puppet --version
#  4.8.2
yum-config-manager --disable CentOS-OpenStack-pike
yum remove -y puppet && yum install -y puppet
yum-config-manager --enable CentOS-OpenStack-pike
puppet --version
# 3.6.2

You should not be able to apply once again

EDIT

Simply yum downgrade puppet=3.6.2 works aswell.