Known Issues

Navigation:  Troubleshooting and Support >

Known Issues

Previous pageReturn to chapter overviewNext page

In some cases KuberDock doesn`t work properly on Amazon Web Services. If you got a problem with cluster on AWS then contact our support team to help you solve the problem.

 

Web-interface in KuberDock version 1.0 doesn`t work in Internet Explorer.

 

If during KuberDock updating, upgrade log is frozen for more than 30 minutes on running docker pull action, for example:

 

[node_hostname] run: docker pull fluentd.

 

Then go to this node console and run the command: 

 

systemctl restart docker.service.

 

Upgrade process will continue after these actions. If not, then stop upgrade process and run command to resume upgrade process:

 

kuberdock-upgrade resume-upgrade 

 

If WHMCS does not send invoices to the customers then check prices in KuberDock package configuration. At least one value must equal more than zero.

 

In some cases users' pods can be frozen in Pending status during container update or upgrade process, starting pod or other actions with pod and container. In this case you should restart docker on the node where this pod is allocated, performing the following steps:

 

oCopy pending pod UUID. (You can find UUID in browser address bar on pod page. For example: https://masterip/#pods/bd1218c0-0321-4821-873f-8fc48638dfdc, where UUID is bd1218c0-0321-4821-873f-8fc48638dfdc).

 

oRun the command on master server to find a node with this pod:

 

kubectl get pods --namespace=UUID_HERE -o yaml

 

You will get a YAML specification of this pod.

 

oFind "hostIP" parameter - which is node IP.

 

oGo to console of this node and run command systemctl restart docker.service

 

In some cases container logs does not appear for a long time.

 

After adding memory to a node KuberDock web-interface does not show new value, but KuberDock core uses it.

 

In case when KuberDock server and WHMCS server have unsyncronized time and date then user will get an error “Link expired” after click on Pay & Start button. You need to follow WHMCS installation guide step 3 and do synchronization between KuberDock master server and WHMCS server.

 

After executing upgrade script the pods that were pending at the moment of upgrade may have no SSH access. Such pods should be restarted manually if SSH access is needed.

 

If WHMCS does not send invoices to your customers, please follow the instructions (step 3) to solve the problem .

 

If you get the following error in WHMCS:

 

Exception with message 'DateTime::__construct(): Is not safe to rely on the system's time zone settings'

 

Do the following:

 

1. Open your php.ini file on WHMCS server.

 

2. Find there a string which says:

 

;date.timezone =

 

3. Remove semicolon “;” in the beginning to uncomment it and add the appropriate timezone for you which can be selected from here http://php.net/manual/en/timezones.php.

 

For example: date.timezone = America/New_York

 

If in WHMCS during KuberDock package edit or create process of Trial package input fields with prices are not hidden, then go to Standard package settings and click Save changes. It must reactivate all the needed hooks in KuberDock plugin.

 

If a user terminates pod in KuberDock, then this item will be deleted from user profile in billing system only the next day. Until the next day WHMCS admin will see this billable item in user profile.

 

For WHMCS plugin version 1.0.8.1 and lower where product setting in Module settings section is set to Automatically setup the product as soon as the first payment is received will cause a problem with manual mark of invoice as Mark Paid. The problem is that users` application will not be created. This will be solved with the upcoming version of WHMCS plugin. Note that there is no behavior to solve that problem at the moment.

 

After establishing SFTP connection using our SFTP direct access feature, you may experience troubles copying local directories into container's file system.

 

sftp> put -r localDirectory

Uploading localDirectory/ to /root/localDirectory

Couldn't canonicalize: No such file or directory

Unable to canonicalize path "/root/localDirectory"

 

 

This is a known issue of OpenSSH https://bugzilla.mindrot.org/show_bug.cgi?id=2150, which is not related to KuberDock itself.

 

One of the suggested workarounds is to create the target directory manually.

 

sftp> mkdir localDirectory

sftp> put -r localDirectory

Uploading localDirectory/ to /root/localDirectory

Entering localDirectory/

 

However in case of any subdirectories present, the command will fail again.

 

Other suggested workarounds is to use SCP instead.

 

In WHMCS If user does not pay the invoice then this invoice will be canceled after suspended days configured in Automation settings. To generate new invoice a user should start unpaid pod in KuberDock.

 

When Plesk is integrated with KuberDock plugin then there is a problem with pod accessibility via Service IP.

 

This problem can be fixed by adding this line to /etc/sysconfig/flanneld config on remote host:

 

FLANNEL_OPTIONS="--iface=eno16777728"

 

Then Flanneld restart is required.

 

This problem affects KuberDock versions below 1.5.0. You can just update KuberDock to 1.5.0 or higher to solve the problem.

 

When deploying KuberDock plugin on cPanel with kcli-deploy.sh from 1.5.0, at the end of the deploy the following error message appears: "/opt/bin/calicoctl: error while loading shared libraries: libz.so.1: failed to map segment from shared object: Operation not permitted".

 

The solution is:

 

Run

 

mount /tmp -o remount,exec

 

before the deploy.