ngcp-upgrade optionsThe sip:carrier version mr5.5.10 has the following important changes:
Performance tuning:
Linux sysctl changes:
Please find the complete changelog in our release notes on our WEB site.
The sip:carrier software upgrade procedure to mr5.5.10 will perform several fundamental tasks:
sip:carrier is a PRO-style system that has "A" and "B" sets of nodes with specific roles. The number of nodes can differ between installations and must be clarified before the upgrade at the planning stage.
The software upgrade is performed by Sipwise engineers with the following steps:
| warning | |
The only allowed software upgrade path is the one described above. All the other theoretically possible upgrade scenarios can lead to unpredictable results. |
| warning | |
Nodes "A" and "B" MUST be used as described in this document. It is NOT allowed to swap them unless proxy replication (of MySQL on port 3308) is configured on the db01b node. |
| warning | |
Debian 9 (stretch) provides OpenSSH 7 with the ssh-dss (DSA) public key algorithm disabled (see more information here). Please ensure you are NOT using ssh-dss (DSA) based SSH keys, otherwise migrate to modern ssh-ed25519 (ED25519) algorithm BEFORE the upgrade. |
| warning | |
The license key for the sip:carrier must be installed and applied on all the servers before the software upgrade. Please contact the Sipwise support team to receive the license key. |
Specify the provided license key in /etc/ngcp-config/config.yml
general: license_key: XXX_XXX_XXX
Have the following information prepared in writing:
It is recommended to execute all preparation steps in this chapter a few days before the actual software upgrade. They do not cause service downtime, so it is safe to execute them in peak hours.
| tip | |
Use the physical IP addresses instead of the shared one, so you can jump between the nodes later on. |
Run the terminal multiplexer under the sipwise user (to reuse the Sipwise .screenrc settings that are convenient for working in multiple windows):
screen -S ngcp-upgrade
Become root inside your screen session:
sudo -s
Check the overall system status:
ngcp-status --all
Ensure that all the proxy nodes replicate the read-only DB (127.0.0.1:3308) from the db01a node. Otherwise, discuss a special plan to address this particularity.
For the following steps, investigate and make sure you understand why the custom modifications were introduced and if they are still required after the software upgrade. If the custom modifications are not required anymore, remove them (e.g. if a bug was fixed in the target release and the existing patch becomes irrelevant).
Create tickets to Sipwise developers to make relevant custom modifications part of the product in the future releases. This will allow you to get rid of the customtt files.
Find local changes to the template files by executing:
ngcp-customtt-diff-helper
ngcp-customtt-diff-helper can help you to download the templates for you target release. To do so, you can run the following command:
ngcp-customtt-diff-helper -d
In the tmp folder provided by the script you can, now, merge the current customtt with the new tt2, creating the new customtt.tt2 files. Once did that, you can TAR the new customtts in a tar.gz file and you can use that after the upgrade to deploy the new customtt.
ngcp-customtt-diff-helper -t
You can check further details regarding the ngcp-customtt-diff-help with the option "-h".
Check if there are any *.tt2.dpkg-dist files among the templates. They usually appear when tt2 files are modified directly instead of creating customtt files. If you find any *.tt2.dpkg-dist files, then treat the corresponding tt2 files as if they were custom templates and introduce the changes from the existing tt2 files into the new templates (by creating associated *.customtt.tt2) before the software upgrade.
find /etc/ngcp-config -name \*.tt2.dpkg-dist
Note that in the end all *.tt2.dpkg-dist files must be removed before the software upgrade.
If you changed the configuration (e.g. added custom templates or changed the existing ones), then the system must be thoroughly tested when these changes are applied. Continue with the software upgrade preparation only when the results of the tests are successful.
Check and remove dpkg files left from previous software upgrades.
Make sure that the list is empty before you continue:
find /etc/ngcp-config -name \*.tt2.dpkg\*
| warning | |
If the installation uses locally specified mirrors, then the mirrors must be switched to the Sipwise APT repositories (at least for the software upgrade). Otherwise, the public Debian mirrors may not provide packages for old Releases anymore or at least provide outdated ones! |
Open separate windows for all the servers inside your "screen" session.
(Press Ctrl+a + c to open a new window, Ctrl+a + a or Ctrl+a + [0-9] to
change the window. Ctrl+a + " shows the list of all your windows.
Use Ctrl+a + A to change the window names to corresponding hosts).
Check the system for locally modified files (move them to appropriate customtt.tt2 files if necessary) on all servers:
ngcp-status --integrity
Make sure the cluster status is OK: on all nodes manually run:
Check all nodes in parallel, using clish and parallel-ssh:
ngcp-clish "ngcp version package check" - ensure that all nodes have the identical Debian package installed
| info | |
Software on all nodes must be identical before and after the upgrade! |
| info | |
to exit from ngcp-clish press Ctrl+Z (or type exit): |
root@web01b:~# ngcp-clish Entering 'clish-enable' view (press Ctrl+Z to exit)... # exit root@web01b:~#
Run "apt-get update" on all nodes. Ensure that you have no warnings/errors here.
Test the cluster failover to see if everything works fine on "B" nodes as well. On all the standby nodes execute:
ngcp-make-active
Afterwards, check ngcp-status --all again.
Log in to all nodes and execute the checks from Section 12.4, “Preparing the software upgrade” again. This will ensure that nothing was broken since the preparation steps were finished. Also, execute ngcpcfg show and ngcpcfg status to check the latest configuration changes.
Perform the BFT test.
Sipwise NGCP introduced Maintenance Mode with its mr5.4.1 release. The maintenance mode of NGCP will disable some background services (for instance, mediator) during the software upgrade. It thus prevents the system from getting into an inconsistent state while the upgrade is being performed. You can activate maintenance mode by applying a simple configuration change as described later.
| important | |
For sip:carrier systems running NGCP release older than mr5.4.1: As upgrading from an earlier release to mr5.4.1 or later will result in a system being aware and making use of the maintenance mode, it is necessary to prepare the NGCP for it before the software upgrade. |
Enable maintenance mode:
ngcpcfg pull
Edit /etc/ngcp-config/config.yml file:
maintenance: yes" in the general section of the configuration file
general.maintenance
parameter to yes
general: maintenance: yes
Disabling background services. These tasks are for systems currently running NGCP release older than mr5.4.1. The maintenance mode of NGCP is not available in your system prior to the software upgrade. It is therefore necessary to manually disable the following background services to avoid a potential system inconsistency:
Execute the following:
Stop HB watchdog on all nodes:
monit stop hb_watchdog
Stop mediator and rate-o-mat services on proxy (prxNNy) nodes :
monit stop mediator monit stop rate-o-mat
Prevent background services from (re)starting. You need to change
configuration in the /etc/ngcp-config/config.yml file: set the enable
parameter to "no":
heartbeat:
hb_watchdog:
enable: no
...
mediator:
enabled: no
...
rateomat:
enable: noIn order to disable the CDR exporter process, add a new custom template file because there is no such file by default:
config="/etc/cron.d/ngcp-cdr-exporter"
customtt="/etc/ngcp-config/templates/etc/cron.d/ngcp-cdr-exporter.customtt.tt2"
[ -f "${customtt}" ] || cp "${config}" "${customtt}"
vim "${customtt}"Comment out this single line in the file:
# 25,55 * * * * root . /etc/default/ngcp-roles; if ...
Please don’t forget to reverse the manual changes after the software upgrade as described in Enabling background services section of the handbook!
For all NGCP systems, regardless of their version:
Apply configuration changes by executing s:
ngcpcfg apply 'Enabling maintenance mode before the upgrade to mr5.5.10' ngcpcfg push all
To upgrade the sip:carrier to mr5.5.10 release, execute the following commands on the standby management "A" node:
The following options in ngcp-upgrade can be specially useful in some
instances of upgrade:
--step-by-step: confirm before proceeding to next step. With this option
the upgrade operation is performed confirming every step before execution,
with the possibility to instruct to continue without confirming further steps
until the end (if confirmation is only needed for some steps at the
beginning).
--pause-before-step STEP_NAME: pause execution before step, given by the
name of the script (e.g. "backup_mysql_db"). This option can be useful in
several scenarios, for example:
--step-by-step), or just continue without stop until the end
--skip-db-backup: This will speed-up the process in cases where it’s
deemed unnecessary, and this is very likely in the upgrade of nodes other than
the first.
| info | |
Sometimes the DB and MGMT roles are assigned to the same host. This is OK. |
| warning | |
Do NOT execute the software upgrade on web01a and db01a in parallel! |
The main goal of the following commands is to download the new packages into the approx cache. So all the nodes in the cluster will get identical packages.
NGCP_CURRENT_VERSION=$(cat /etc/ngcp_version)
sed -i "s/$NGCP_CURRENT_VERSION/mr5.5.10/" /etc/apt/sources.list.d/sipwise.list
sed -i "s/jessie/stretch/g" /etc/apt/sources.list.d/sipwise.list /etc/apt/sources.list.d/debian.list
[ -r /etc/default/ngcp-proxy ] && source /etc/default/ngcp-proxy
mgmt_node=$(cat /etc/ngcp_mgmt_node)
grep -q "debian-debug" /etc/apt/sources.list.d/debian.list || \
echo "deb http://${mgmt_node:-web01}:${APPROX_RO_PORT:-9998}/debian-debug/ stretch-debug main contrib non-free" >> /etc/apt/sources.list.d/debian.list
grep -q "debian-debug" /etc/approx/approx.conf || \
echo "debian-debug https://debian.sipwise.com/debian-debug" >> /etc/approx/approx.conf
ngcp-approx-cache-helper --auto --node localhost
apt-get update
apt-get install ngcp-upgrade-pro| info | |
Don’t worry, ngcp-upgrade-carrier does not exist, use ngcp-upgrade-pro as outlined above. |
| warning | |
Do not use "ngcpcfg apply/build" after executing the steps from the above section, otherwise the changes will be overwritten and you will have to redo these steps. The same applies to similar sections below. |
Execute ngcp-upgrade on the standby node as root:
ngcp-upgrade
| info | |
sip:carrier can be upgraded to mr5.5.10 from previous release or previous build only. The script ngcp-upgrade will find all the possible destination releases for the upgrade and allow to choose the proper one. |
| info | |
If there is an error during the upgrade, the ngcp-upgrade script will
request you to solve it. Once you’ve fixed the problem, just execute
|
Merge/add the custom configuration templates if needed.
Apply the changes to configuration templates:
ngcpcfg apply 'applying customtt for new release mrX.X on node xxx01a'
Send the new templates to the shared storage and the other nodes
ngcpcfg push --nobuild --noapply all
| warning | |
Do NOT execute ngcpcfg push --shared-only at this stage, as it will affect further upgrades due to noticed outdated local ngcpcfg storage. If you did so, run ngcpcfg push --nobuild --noapply all once again to pull ngcpcfg changes on all the nodes from glusterfs. |
| info | |
If the DB and MGMT roles are assigned to the same host, then skip this step as you have already upgraded the standby MGMT node "A" above. |
Run the following commands to upgrade the standby DB node "A" (select the same release version as above and follow the on-screen recommendations):
NGCP_CURRENT_VERSION=$(cat /etc/ngcp_version) sed -i "s/$NGCP_CURRENT_VERSION/mr5.5.10/" /etc/apt/sources.list.d/sipwise.list sed -i "s/jessie/stretch/g" /etc/apt/sources.list.d/sipwise.list /etc/apt/sources.list.d/debian.list apt-get update apt-get install ngcp-upgrade-pro ngcp-upgrade
| info | |
It is important to upgrade db01a node before upgrading any proxy nodes.
Otherwise, the "local" MySQL (127.0.0.1:3308) on proxy nodes may become out of
sync in case the new release has |
Run the below commands selecting the same release version and follow the on-screen recommendations:
NGCP_CURRENT_VERSION=$(cat /etc/ngcp_version) sed -i "s/$NGCP_CURRENT_VERSION/mr5.5.10/" /etc/apt/sources.list.d/sipwise.list sed -i "s/jessie/stretch/g" /etc/apt/sources.list.d/sipwise.list /etc/apt/sources.list.d/debian.list apt-get update apt-get install ngcp-upgrade-pro ngcp-upgrade
The following options in ngcp-upgrade can be useful for this phase of
upgrades, because it is very likely that the backup was already performed:
--skip-db-backup: This will speed-up the process in cases where it’s
deemed unnecessary.
See a more detailed description of the options in: ngcp-upgrade options
| warning | |
Ensure that all standby nodes "A" are:
* upgraded to the new release (check /etc/ngcp_version or use |
On all "A" nodes run:
ngcp-make-active
Ensure that the "A" nodes became active, by executing the 'ngcp-status' and 'ngcp-clish' commands described above.
Ensure that ALL "B" nodes are standby now!
Run the following commands selecting the same release version and following the on-screen recommendations:
NGCP_CURRENT_VERSION=$(cat /etc/ngcp_version) sed -i "s/$NGCP_CURRENT_VERSION/mr5.5.10/" /etc/apt/sources.list.d/sipwise.list sed -i "s/jessie/stretch/g" /etc/apt/sources.list.d/sipwise.list /etc/apt/sources.list.d/debian.list apt-get update apt-get install ngcp-upgrade-pro ngcp-upgrade
| info | |
You can upgrade all standby "B" nodes simultaneously (including the ones with the mgmt and db roles). |
The following options in ngcp-upgrade can be useful for this phase of
upgrades:
--step-by-step: confirm before proceeding to next step.
--pause-before-step STEP_NAME: pause execution before step, given by the
name of the script (e.g. "backup_mysql_db").
See a more detailed description of the options in: ngcp-upgrade options
In order to disable the maintenance mode, do the following:
ngcpcfg pull
ngcpcfg set /etc/ngcp-config/config.yml "general.maintenance=no"
Enabling background services. If you upgraded from NGCP release earlier than mr5.4.1, you have to manually reverse the changes you applied before the upgrade in the (Disabling background services) section.
Enable background services:
ngcpcfg set /etc/ngcp-config/config.yml "heartbeat.hb_watchdog.enable=yes" ngcpcfg set /etc/ngcp-config/config.yml "mediator.enabled=yes" ngcpcfg set /etc/ngcp-config/config.yml "rateomat.enable=yes"
In order to enable the CDR exporter process, remove the corresponding customtt file:
rm /etc/ngcp-config/templates/etc/cron.d/ngcp-cdr-exporter.customtt.tt2
For all NGCP systems, regardless of their previous version:
ngcpcfg apply 'Disable the maintenance mode after the upgrade to mr5.5.10' ngcpcfg push all
When everything has finished successfully, check that replication is running.
Check ngcp-status --all.
Finally, do a basic functionality test.
Check the web interface, register two test subscribers and perform a test call
between them to ensure call routing works.
| info | |
You can find a backup of some important configuration files of your existing installation under /var/backup/ngcp-mr5.5.10-* (where * is a place holder for a timestamp) in case you need to roll back something at any time. A log file of the upgrade procedure is available at /var/backup/ngcp-mr5.5.10-*/upgrade.log. |