Hi Billy, > Chunghan will check for relevant section in fuel documentation
I checked fuel repo.( https://gerrit.opnfv.org/gerrit/gitweb?p=fuel.git;a=tree;f=docs;hb=HEAD) to add the troubleshooting to the fuel documentation. However, I cannot find the suitable location at the documents. >[fuel.git]/docs/installationprocedure/installation.instruction.rst The above location looks like the suitable location, but there is no mention about the virtual deployment. (It looks like the deployment procedure for bare-metal…) We’d like to ask the location to Michal Skalski ? (In my memory, he mentioned the fuel documentation at the yesterday's meeting.) # If I was wrong, sorry. Alternatively, our project has the document for the virtual deployment of old version. https://wiki.opnfv.org/display/ovsnfv/Deploy+OPNFV+with+Fuel+in+a+virtual+environment I think that it is a possible way to add the troubleshooting to the above link (wiki). Could you give me your opinion ? Regards, Chunghan Lee ---------------------------------------------------------- Chunghan Lee Cloud Network Project. Fujitsu Laboratories Ltd. 4-1-1 Kamikodanaka, Nakahara-ku, Kawasaki-shi, Kanagawa 211-8588, Japan E-mail： [email protected]<mailto:[email protected]> ---------------------------------------------------------- From: O Mahony, Billy [mailto:[email protected]] Sent: Monday, August 29, 2016 6:14 PM To: Lee, Chunghan/李 忠翰; [email protected] Cc: [email protected]; Gray, Mark D Subject: RE: [ovsnfv] Success virtual deployment (Fuel) + Report troubleshooting Hi Chunghan, That’s great work - good to know of some fixes required for using CentOS. A few other comments inline below. Thanks, Billy. From: Lee, Chunghan [mailto:[email protected]] Sent: Friday, August 26, 2016 10:58 AM To: O Mahony, Billy <[email protected]<mailto:[email protected]>>; [email protected]<mailto:[email protected]> Cc: [email protected]<mailto:[email protected]>; Gray, Mark D <[email protected]<mailto:[email protected]>>; Lee, Chunghan <[email protected]<mailto:[email protected]>> Subject: [ovsnfv] Success virtual deployment (Fuel) + Report troubleshooting Hi Billy, I can deploy the OPNFV testbed (Colorado) using Fuel on my local environment. Thank you so much! The local environment (one server) is as follow. --- OS : CentOS 7.2 (Default kernel : 3.10) CPU : Intel(R) Xeon(R) CPU E5640 @ 2.67GHz Memory : 48GB DISK : 1TB Fuel IMAGE : opnfv-2016-08-24_22-58-32.iso http://artifacts.opnfv.org/fuel.html # I tried to do the indicated image (opnfv-2016-08-03_23-58-37.iso) from Billy, # but it was already deleted. --- I’d like to report some problems to install the testbed and empirical approach to solve them. 1. If the OS language (LANG) is not English, the installation will be stopped. Fuel did not know the other languages, such as Korean or Japanese. In the local environment, LANG was ja_JP.UTF-8 unfortunately. In my case, LANG is changed from ja_JP.UTF-8 to C, and I can solve this problem. 2. If the OS is CentOS7, the location of kvm is different from Ubuntu. In Fuel, the kvm location is based on Ubuntu. I created the follow symbolic link myself. > ln -s /usr/libexec/qemu-kvm /usr/bin/kvm 3. New version of python packages are required for the installation. If the scp python package is installed by yum (epel), this version (python-scp-0.7.1-3.el7.noarch.) is old and there is no data type (sanitize) at the old version. As a result, the installation will be stopped due to no sanitize. The scp package (0.10.2) is installed by pip. 4. The following message is time-consuming. > Trying to SSH into Fuel VM 10.20.0.2 ... sleeping 10 seconds # The patience is required for this procedure. [[BO'M]] This is the deployment scripts way of figuring out when the installation of the fuel master has completed - it just keeps trying to ssh to the fuel master. After that it can track progress by invoking the fuel cli over ssh. 5. IP addresses are require for DNS and NTP. IP address (10.20.0.2) are inserted at the following points. # I think that 10.20.0.2 is invalid for actual testbed. [[BO'M]] [[BO'M]] Yes you can’t leave these blank. File location : fuel/deploy/config/labs/local/virtual_minimal/fuel/config/dea-pod-override.yaml Modified fields (dns_nameservers:, external_dns: value:, external_ntp: value:) When the above fields are null, the following ERROR is always occurred. (But, there is no problem of download.) ---- Upload network config for environment 1 Failed command: fuel network --env 1 --upload --dir /var/lib/opnfv Command returned response: 400 Client Error: Bad Request for url: http://10.20.0.2:8000/api/v1/clusters/1/network_configuration/neutron (None is not of type 'string' Failed validating 'type' in schema['properties']['networking_parameters']['properties']['dns_nameservers']['items']: {'anyOf': [{'format': 'ipv4'}, {'format': 'ipv6'}], 'type': 'string'} On instance['networking_parameters']['dns_nameservers'][0]:None) ---- > The script will create a three VMs on your host. One for fuel master and one > each for a control and compute node. I found three VMs and can access to fuel master and controller. But, I did not check the installation status of OPNFV modules (e.g. DPDK OVS and vperf…). I will check them and notify the situation. [[BO'M]] [[BO'M]] The vsperf plugin should work as it just copies vsperf python code to the compute nodes. The DPDK plugin won’t work in a virtual deployment. But this should be solved in Colorado release - at least I Fuel team have made some commits regarding that recently. Regards, Chunghan Lee ---------------------------------------------------------- Chunghan Lee Cloud Network Project. Fujitsu Laboratories Ltd. 4-1-1 Kamikodanaka, Nakahara-ku, Kawasaki-shi, Kanagawa 211-8588, Japan E-mail： [email protected]<mailto:[email protected]> ----------------------------------------------------------
_______________________________________________ opnfv-tech-discuss mailing list [email protected] https://lists.opnfv.org/mailman/listinfo/opnfv-tech-discuss
