Home

Category: technology

  • Clear Maven Cache: Solution to transitive dependencies not updating on clean install

    Follow these steps when maven clean install just doesn’t update the transitive dependencies in nested projects.

    rm -rf .m2/repository/path-to-your-library
cd your-project-folder
find . -maxdepth 2 -name "pom.xml"  -exec mvn clean -f '{}' \;
mvn clean install -DskipTests=true
#confirm if the versions are as you wanted
mvn dependency:tree | grep -B 75 the-bugging-library

    If the above method doesn’t work, you might have missed to add the required version in the dependency management.

    Lets assume BL is the library which has been bugging you with these version issues 😉

    cd your-mvn-parent-project
vi pom.xml
#add
    <dependencyManagement>
        <dependencies>
            <dependency>
                <groupId>BL-group</groupId>
                <artifactId>BL-artefact</artifactId>
                <version>required-version-of-the-BL</version>
            </dependency>
        </dependencies>
    </dependencyManagement>

    Let me know if these steps helped.

  • Set timezone to India in Ubuntu

    Rather than a how to, written as a must do for our servers.

    Probably the most important command for 60% of developers out there 🙂

    sudo timedatectl set-timezone Asia/Kolkata

  • Fix the error: java.lang.ExceptionInInitializerError: com.sun.tools.javac.code.TypeTags during maven builds

    Problem: Maven package gives this compile time error when using Open JDK.

    Fatal error compiling: java.lang.ExceptionInInitializerError: com.sun.tools.javac.code.TypeTags -> [Help 1]

    Reason: Lombok version included in the project seemingly had a dependency on Oracle JDK, which is fixed in 1.18.2

    Fix: Upgrade lombok to 1.18.2 in the POM file.

    		<dependency>
			<groupId>org.projectlombok</groupId>
			<artifactId>lombok</artifactId>
			<version>1.18.2</version>
			<scope>provided</scope>
		</dependency>

      

  • Tat Kim Bot : A twitter bot to transliterate Sanskrit

    Tat Kim Bot : A twitter bot to transliterate Sanskrit

    Glad to write about a twitter bot I had written and launched during the last Independence day – Tat Kim. For those who are not in the know, Tat Kim means “What’s that?” in Sanskrit – a very apt name for a Bot and allied service which answers questions of Sanskrit enthusiasts.

    What does the Tat Kim Bot do?

    Currently Tat Kim bot can be used to read Sanskrit tweets in your own script (Eg, Malayalam or Kannada) – there by taking away the need to know how to read Hindi (or Devanagari) to be able to consume these tweets.

    Why?

    Sanskrit is a pan-Indian language and it should be free from the constraints of a regional Lipi like Devanagari. Also, it helps to remember that some of the seminal works in Sanskrit were recorded not in Devanagari, but in Grantha Lipi (more or less Malayalam of our times)

    Usage

    If you happen to find a Sanskrit tweet from your favourite Sanskrit tweeple, comment on that tweet in the below format

    @TatKimBot {languageCode}

Where language could be one from:
en, hi, bn, gu, kn, ml, or, pa, te, ta, mr

    Details of Language codes:

    Codescript
    enEnglish
    hiDevanagari (Kept to support translations to Hindi at a later point of time)
    bnBengali
    guGujarati
    knKannada
    mlMalayalam
    orOriya
    paPunjabi
    teTelugu
    taTamizh (The transliterations will be lossful due to Tamil script not being elaborate enough to represent all sounds of Sanskrit)
    mrMarathi (Kept to support translations to Marathi at a later point of time)

    Example:

    @TatKimBot ml
    And the bot responds in 15 seconds as in the below screenshot:

    Archives of the transliterations

    The archives of the transliterations done by Tat Kim Bot will be available at the web portal https://tat.kim

    (A sample screenshot below)

    Look forward to people finding and using it during the next few months!

  • Migrate data from one Cassandra server to another using medusa

    Here are the steps I followed to get my dev instance of cassandra copied to a new server (both single node instances)

    I’ve used medusa tool for the same. You can read more about it here

    Installation

    Install medusa on both source and destination cassandra-servers, by running the following commands on both –

    curl -1sLf   'https://dl.cloudsmith.io/public/thelastpickle/medusa/setup.deb.sh'   | sudo -E bash

sudo apt update
sudo apt-get install cassandra-medusa

    Export Data from Source DB

    #ssh into source server and do these

mkdir ~/cassandra-bkup
sudo cp /etc/medusa/medusa-example.ini /etc/medusa/medusa.ini
sudo vi /etc/medusa/medusa.ini

# Search for, and set these values in the ini file
storage_provider = local
base_path = /home/{username}/cassandra-bkup
# Save and Exit

medusa backup --backup-name=dev-db-data --mode=full

#compress output folder
tar -czf dev-db-bkup-tar ~/cassandra-bkup

# SCP the backup directory to destination server
scp -r -i <key-file> dev-deb-bkup-tar {dest-user}@{dest-server}:~/

    Import data in Target DB

    # ssh into target serer and do these
tar -xzf dev-deb-bkup-tar

ls ls /home/{username}/cassandra-bkup/cassandra_backups/
#note down the folder name parallel to the folder index, let it be {sourceHostLocalName}
sudo /etc/hosts
#add an entry with local nodes IP and host name as {sourceHostLocalName}. 
#You can remove this later
# save /etc/hosts and Exit Vim.
sudo medusa restore-node --keyspace={keyspace of intereset} --backup-name=dev-db-data

    Verify Imported data

    cqlsh {targetNodeIp}
desc keyspaces #verify you have the source keyspace created
use {keyspaceOfInterest}
desc tables; #verify all your tables in the source are recreated here.

    That’s it. You should now be able to work with your new Cassandra node!

  • Blocked when SSH’in due to startup error? SSH without running .bashrc

    In case a faulty script started by .bashrc or .profile is failing and crashing SSH initiation, you could still login and fix the issue by running shell directly via ssh.

    The command is –

    ssh -t {username}@{hostname} /bin/sh

  • A JS assistant to remind us to exercise during WFH days

    Here is a weekend hack I wrote to remind me to get up from my workstation and do some lightweight exercises during the WFH days – Fit@WFH

    Some of the features are –

    • Purely in browser side JS (your data stays at your computer)
    • Reminds you to workout during the tail end of every hour
    • Reminder both in Voice and Browser Notification
    • You can select the exercises you are comfortable with from a list
    • Built for desktop (i.e, for serious WFH’ers )

    I guess someone else might also find tremendous value from it. So if you too are caught in a sedentary lifestyle and want to move more, go ahead – book mark and use it !

    Launch WFH Fitness Utility

  • Flush DNS cache in Ubuntu 20.xx and above

    On bash, run –

    sudo systemd-resolve --flush-caches

    Should be done.

  • Change the datacenter set in cassandra-rackdc on Ubuntu

    Problem: Changing just the datacenter name in cassandra-rackdc.properties gives the error:

    Cannot start node if snitch's data center [newName] differs from previous data center [oldName]. Please fix the snitch configuration, decommission and rebootstrap this node or use the flag -Dcassandra.ignore_dc=true.

    Solution:
    If you are just beginning to setup you cassandra cluster, and is ok to reset the data to bootstrapped state, do this

    sudo service cassandra stop
sudo rm -R /var/lib/cassandra
sudo mkdir /var/lib/cassandra
sudo chown cassandra /var/lib/cassandra
sudo service cassandra start

    If you need to persist the existing data, do this

    sudo vi /etc/cassandra/cassandra-env.sh
# edit the JVM_OPTS line to make it look like below
JVM_OPTS=\"$JVM_OPTS -Dcassandra.ignore_rack=true -Dcassandra.ignore_dc=true\"

sudo service cassandra restart

    This should normally resolve the dc name correction in rack-dc properties file.

  • RabbitMQ installation fails in in the last leg of setting up in ubuntu 20.04

    Error:

    Setting up rabbitmq-server (3.8.16-1) …
Job for rabbitmq-server.service failed because the control process exited with 
error code.
See "systemctl status rabbitmq-server.service" and "journalctl -xe" for details.
invoke-rc.d: initscript rabbitmq-server, action "start" failed.

    Reason: libtinfo.so library missing.

    Solution:

    sudo apt install libncurses5

    Note: No need to reattempt the installation.