From d4216dd28019256fd6363ed4269d540ef64ec23c Mon Sep 17 00:00:00 2001 From: Luc Provoost Date: Fri, 12 Apr 2019 11:20:58 +0200 Subject: Improved image building & new PROX Bug fixes for the script for building an image via the packer tool. This also installs a newer version of PROX with some bug fixes in PROX. The *.test files have been simplified. Also introduced a new file MachineMap.cfg to map test machines on actual VMs. Change-Id: I69445327ef0a907bc8c1566aaa60c733418c541e Signed-off-by: Luc Provoost --- .../DPPD-PROX/helper-scripts/openstackrapid/README | 51 +++++++++++----------- 1 file changed, 26 insertions(+), 25 deletions(-) (limited to 'VNFs/DPPD-PROX/helper-scripts/openstackrapid/README') diff --git a/VNFs/DPPD-PROX/helper-scripts/openstackrapid/README b/VNFs/DPPD-PROX/helper-scripts/openstackrapid/README index 0dea6565..43243a6c 100644 --- a/VNFs/DPPD-PROX/helper-scripts/openstackrapid/README +++ b/VNFs/DPPD-PROX/helper-scripts/openstackrapid/README @@ -21,19 +21,18 @@ rapid is a set of files offering an easy way to do a sanity check of the dataplane performance of an OpenStack environment. Copy the files in a directory on a machine that can run the OpenStack CLI -commands and that can reach the OpenStack public network. +commands and that can reach the OpenStack networks to connect to the VMs. You will need an image that has the PROX tool installed. -The best way to do this is to use the packer tool to build an image for a target of your choice. -You can also build this image manually by executing all the commands described in the deploycentos1.sh -and deploycentos2.sh file. -The default name of the qcow2 file should is rapidVM.qcow2 +A good way to do this is to use the packer tool to build an image for a target of your choice. +You can also build this image manually by executing all the commands described in the deploycentos.sh. +The default name of the qcow2 file is rapidVM.qcow2 When using the packer tool, the first step is to upload an -existing CentOs cloud image from the internet into OpenStack. +existing CentOS cloud image from the internet into OpenStack. Check out: https://cloud.centos.org/centos/7/images/ You should now source the proper .rc file so Packer can connect to your OpenStack. -There are 3 files: centos.json, deploycentos1.sh and deploycentos2.sh, allowing you to create +There are 2 files: centos.json and deploycentos.sh, allowing you to create an image automatically. Run # packer build centos.json Edit centos.json to reflect the settings of your environment: The following fields need to @@ -47,7 +46,7 @@ be the ID's of your system: - "security_groups": ID of the security group being used Note that this procedure is not only installing the necessary tools to run PROX, -but also does some system optimizations (tuned). +but also does some system optimizations (tuned). Check deploycentos.sh for more details. Now you can run the createrapid.py file. Use help for more info on the usage: # ./createrapid.py --help @@ -58,12 +57,11 @@ It will create a .env file containing all info that will be used by runra to actually run the tests. Logging can be found in the CREATE.log file You can use floating IP addresses by specifying the floating IP network --floating_network NETWORK -or directly connect throught the INTERNAL_NETWORK by using the following parameter: +or directly connect through the INTERNAL_NETWORK by using the following parameter: --floating_network NO /etc/resolv.conf will contain DNS info from the "best" interface. Since we are deploying VMs with multiple interface on different networks, this info might be -taken from the "wrong" network (e.g. the dataplane network). It is good to provide -also a working DNS on that network. +taken from the "wrong" network (e.g. the dataplane network). Now you can run the runrapid.py file. Use help for more info on the usage: # ./runrapid.py --help @@ -71,6 +69,7 @@ The script will connect to all machines that have been instantiated and it will PROX in all machines. This will be done through the admin IP assigned to the machines. Once that is done it will connect to the PROX tcp socket and start sending commands to run the actual test. +Make sure the security groups allow for tcp access (ssh & prox port). It will print test results on the screen while running. The actual test that is running is described in .test. @@ -114,35 +113,37 @@ achieved using createrapid.py. Fields needed for runrapid are: admin_ip = none [M1] -admin_ip = 192.168.4.130 -dp_ip = 10.10.10.6 -dp_mac = fa:16:3e:3c:1e:12 +name = rapid-VM1 +admin_ip = 10.25.1.116 +dp_ip = 10.10.10.7 +dp_mac = fa:16:3e:59:b8:28 [M2] -admin_ip = 192.168.4.140 -dp_ip = 10.10.10.9 -dp_mac = fa:16:3e:2a:00:5d +name = rapid-VM2 +admin_ip = 10.25.1.126 +dp_ip = 10.10.10.11 +dp_mac = fa:16:3e:c9:54:c7 [M3] -admin_ip = 192.168.4.138 -dp_ip = 10.10.10.11 -dp_mac = fa:16:3e:ae:fa:86 +name = rapid-VM3 +admin_ip = 10.25.1.108 +dp_ip = 10.10.10.15 +dp_mac = fa:16:3e:72:90:3e [OpenStack] stack = rapid -yaml = 3VMrapid.yaml +vms = rapidVMs key = prox -flavor = prox_flavor image = rapidVM image_file = rapidVM.qcow2 dataplane_network = dataplane-network subnet = dpdk-subnet subnet_cidr = 10.10.10.0/24 internal_network = admin_internal_net -floating_network = admin_floating_net +floating_network = floating-ip-net [rapid] loglevel = DEBUG -version = 17.10.25 -total_number_of_vms = 3 +version = 19.4.15 +total_number_of_machines = 3 -- cgit 1.2.3-korg