About
More than 22 years of experience in core Database Administration, consulting and…
Articles by Somdyuti
-
Oracle Streams/CDC (ORA-1461)
Oracle Streams/CDC (ORA-1461)
Source: 2 Node RAC database Target: Single instance database CDC runs in target and works as downstream capture. There…
Activity
10K followers
-
Somdyuti Paul shared thisHappy New Year to everyone. Wishing everyone great health and success in 2026 and beyond! First post of 2026 and an important one:) Tuning MySQL instances for write-intensive workloads is a persistent challenge. Cloud SQL for MySQL Enterprise Plus edition now includes optimized writes, a set of automated features that adjust MySQL configurations based on real-time workload and infrastructure metrics. This reduces write latency and increases throughput without manual intervention. Customers have started using this feature and have got upto 3x improvement on throughput. Interested to know more? Please read the Google Cloud Blog below and explore! #GCP #MySQL #CloudSQL #Database #Performance Alex Joseph Isabella Lubin Sujatha Mandava Mosè Tronci Kiran Tailor https://lnkd.in/eUSauMgFCloud SQL for MySQL introduces optimized writes | Google Cloud BlogCloud SQL for MySQL introduces optimized writes | Google Cloud Blog
-
Somdyuti Paul shared thisAre you facing challenges with creating the optimal ANN Index? You are asking yourself these questions:- What index type should I use (IVF, HNSW)? What values should I set the DB, index or query parameters to? How do I provide explainability for the results? When should I reindex? How much resources will I need to reindex? These hurdles can slow down development and make it difficult to move from prototype to production. Cloud SQL for PostgreSQL has a solution- Vector Assist Vector Assist for Cloud SQL for PostgreSQL aims to solve these challenges by providing a simplified, declarative framework. This extension helps you: Accelerate Prototyping: Go from idea to a working vector search prototype in minutes, not hours. Automate Optimization: Automatically tune index parameters (like HNSW parameters) to find the best balance of speed and recall for your needs, reducing a process that could take days to just minutes. Simplify Workflows: Use a straightforward SQL interface to create, manage, and query your vector indexes. Democratize Vector Search: Make powerful semantic search capabilities accessible to all developers, regardless of deep expertise in vector databases. With Vector Assist, you can express your intent, and Cloud SQL provides prescriptive recommendations and automates the heavy lifting. See below, you are just explaining your intent, the extension does the rest for you! Bala Narasimhan Gunjan Juyal Kiran Tailor #PostgreSQL #Vector #GenAI #GCP #CloudSQL
-
Somdyuti Paul posted thisA tricky Replication Conflict Case in PostgreSQL Streaming Replication for one of my customers. For the customer, the Standby/Read replica is mainly used to offload select queries from primary. Many of the queries in standby were getting cancelled. And it was not due to Snapshot replication conflict (hot_standby_feedback=on in standby). On querying pg_stat_database_conflicts, we could see lot of conflicts under "confl_lock". Now lock replication conflicts will happen for any operations that take ACCESS EXCLUSIVE lock (like DROP Table, TRUNCATE, DROP INDEX, ALTER TABLE, DROP TRIGGER) on the primary (which conflicts with ACCESS SHARE which is needed for SELECTs in standby) must be replayed on the standby to keep incompatible operations on the table from happening. So if primary is doing an ALTER TABLE TEST...and the standby is doing a SELECT ...FROM TEST and if the standby should replay such a lock on a table that a query uses, we have a lock conflict. We checked and verified that the failing queries on standby were primarily on 2 heavily used tables (lots of UPDATEs and DELETEs) but there was no operation in primary on these tables that could get an ACCESS EXCLUSIVE Lock (who would like to get this lock in a production on heavily used tables that would cause a business outage ??) So what was the problem? What was causing the Lock Replication conflicts in standby? We started digging into the VACUUM operations of these 2 tables in primary. There is one kind of ACCESS EXCLUSIVE lock that you cannot avoid : locks from VACUUM truncation. When VACUUM has finished processing a table(i.e. cleaning dead tuples) and the pages at the end of the table have become empty, it tries to get a short ACCESS EXCLUSIVE lock on the table. It is like lowering the HWM. Release space to OS. If that succeeds, it will truncate the empty pages and immediately release the lock. While such locks don't disrupt processing on the primary, they can cause replication conflicts on the standby. And that was exactly the customer case! Solution? ALTER TABLE busy_table SET (vacuum_truncate = off); This was done for both the tables. This is not turning off VACUUM but to avoid truncation of empty pages at the end of VACUUM. This is a lesser known case of Replication conflict in PostgreSQL but can be tricky to figure out if it happens. Hope this helps #PostgreSQL #Replication #GCP #CloudSQL
-
Somdyuti Paul shared thisMigrating your Oracle or PostgreSQL Database to GCP CloudSQL for PostgreSQL or AlloyDB using Database Migration Service (DMS)? Worried about the data transfer speed for the Initial Load and CDC? Migration window getting stretched for VLDBs? Please read the article on the various optimization that you can do to achieve data transfer speed between 120-180GB/hour using DMS. https://lnkd.in/e3qi5NVY Adi Shtatfeld #PostgreSQL #Oracle #Databases #GCP #CloudSQL #AlloyDB #DatabaseMigrationBest practices for migrating PostgreSQL to Cloud SQL with DMS | Google Cloud BlogBest practices for migrating PostgreSQL to Cloud SQL with DMS | Google Cloud Blog
-
Somdyuti Paul posted thisDoes your PostgreSQL Read Replica, using Physical Streaming Replication, has consistently high Replication lag? And most of the Replication lag is due to WAL Replay? Then read on.. Diagnosis:- 1. You have confirmed from pg_stat_replication (in source) that the Replication lag is primarily due to WAL Replay. 2. You have used the below query in Read Replica and confirmed pg_last_wal_replay_lsn is way behind pg_last_wal_receive_lsn select pg_wal_lsn_diff(pg_last_wal_receive_lsn(),pg_last_wal_replay_lsn()) replay_lag; Reason:- WAL Replay lag in Physical Streaming replication can be caused by a few factors, primarily these three- a) Replication conflicts (Snapshot Replication conflicts, Lock Replication conflicts, Buffer pin replication conflicts) if the Read Replica is also serving read-only queries. b) WAL replay(recovery) is stalled/waiting on IO for data blocks(to which WALs need to be applied) to be read from disk to shared buffers. c) Read Replica server overloaded Solution:- Based on whether the WAL Replay lag/slowness is caused by a), b), or c) or a combination of above:- 1. Are heavy read only/analytic queries being run on Read Replica? If yes, either set max_standby_streaming_delay=0 (PostgreSQL cancels queries on replica immediately, no delay for WAL replay) or set hot_standby_feedback=on. This will prevent Snapshot replication conflict but can cause Vacuum to be less aggressive on Primary instance causing some table bloats. b) Monitor the Read only queries in Replica- mainly wait events. Are they slowing down the WAL Recovery? Use pg_wait_sampling and pg_stat_statements to identify slow queries and waits. c) Optimize Recovery prefetch in Read Replica. Set the following Read Replica 1. recovery_prefetch=on (this is an excellent option if your Read Replica is PG v15 and higher) Whether to try to prefetch blocks that are referenced in the WALs that are not yet in the buffer pool, during recovery. Prefetching blocks that will soon be needed can reduce I/O wait times during recovery with some workloads. 2. wal_decode_buffer_size=16/32/64/128MB (again available from v15) A limit on how far ahead the server can look in the WAL, to find blocks to prefetch. 3. maintenance_io_concurrency=16 d) Monitor the load on Read Replica- option to add more CPU, memory or disk. Still worried about Physical Streaming Replication Lag? AlloyDB is a great option which has optimized the Streaming Replication. #PostgreSQL #Database #Replication #AlloyDB #GCP
-
Somdyuti Paul shared thisAre you facing the below challenges in AutoVacuum? Which tables shall we change the AutoVacuum default values? What will be the non-default values for these AutoVacuum parameters that need to be set in the identified tables or vacuuming frequency of the large and frequently updated tables be? How to adjust these parameters based on the current Database workload and Server Resource Utilization? If you are using AlloyDB then you do not need to worry. AlloyDB’s Adaptive AutoVacuum takes care of the above 3 challenges based on your workload. Autovacuum parameters get dynamically adjusted and set based on your workload, behind the scenes, which makes AutoVacuum optimal. Watch the video on AlloyDB Adaptive AutoVacuum https://lnkd.in/etaG2JCH #PostgreSQL #Performance #AlloyDB #GCP
-
Somdyuti Paul shared thisThe simple yet powerful gemini-cli extensions helps you to "talk" to your database in Natural Language from the command prompt. You do not need to know and write SQLs to explore and manage schema. You do not need to be a DBA to troubleshoot issues. Watch the video to understand how gemini cli Extension- Cloud SQL for PostgreSQL helps you to manage your databases, execute queries, explore schemas, and troubleshoot issues directly from the GEMINI CLI, using natural language prompts. #GCP #Gemini #CloudSQL #PostgreSQL #AI #Databases Bala Narasimhan Mosè Tronci Kiran Tailor https://lnkd.in/eU6T-sfE
-
Somdyuti Paul posted thisWant to use natural language in SQL queries for filtering and ranking criteria? Want to unlock deep semantic insights from enterprise data? With AlloyDB AI Query Engine you can integrate the power of AI and LLMs into your application using AI operators. It is Powerful: Bring the power of LLMs, especially real-world knowledge, into your SQL applications. It is Simple and easy-to-use: Use natural language-based descriptions seamlessly with columns & expressions in SQL Use Multimodal embeddings: Easily generate embeddings for text, images, and videos so you can search using any modality The following example shows the use of the ai.if operator, which is AI powered SQL function, as a filter to find restaurants with more than 500 positive reviews located in cities with a population greater than 100,000. The example uses restaurant_reviews and contains data like reviews and city location. The ai.if operator helps you understand the review sentiment and combine the locations from the database with Gemini general knowledge about the population in these locations. SELECT r.name, r.location_city FROM restaurant_reviews r WHERE AI.IF(r.location_city || ' has a population OF more than 100,000 AND the following is a positive review; Review: ' || r.review) GROUP BY r.name, r.location_city HAVING COUNT(*) > 500; #AlloyDB #AI #Gemini #PostgreSQL #GCP Sanjay Mishra Kiran Tailor
-
Somdyuti Paul shared thisMigrating Oracle to PostgreSQL? You are an Oracle DBA but not an expert in PostgreSQL? There are few challenges with the current Database Migration tools. The available Database Migration Assessment tools which are used to assess the source Oracle database and generate Oracle to PostgreSQL Migration Assessment reports based on the source schema, PL/SQL Code, initialization parameters, database options used do not recommend the target PostgreSQL configuration parameters or flags that need to be set. They also do not recommend the various extensions, functionalities that are available in PostgreSQL to implement the various Oracle database options. Many Oracle to PostgreSQL migrations happen without setting or modifying the default PostgreSQL parameters which cause performance and scalability issues post migrations. The default PostgresSQL parameters are conservative and they need to be changed based on the workload and also the source Oracle init parameters Please read on to configure your PostgreSQL parameters based on Oracle parameters. The document also gives you how you can use PostgreSQL to map Oracle functions. https://lnkd.in/ei2Ec5r3 #Oracle #PostgreSQL #Migration #Performance #CloudSQLOracle to CloudSQL for Postgres Migration- Parameters and Feature MappingOracle to CloudSQL for Postgres Migration- Parameters and Feature Mapping
-
Somdyuti Paul liked thisSomdyuti Paul liked thisEager to join the most dynamic database in the age of AI?
-
Somdyuti Paul reacted on thisSomdyuti Paul reacted on thisI'm thrilled to share that I’m starting a new chapter as VP of Software Engineering at Conga! The journey at PROS has been nothing short of incredible. It was a masterclass in the full 360 of enterprise software—from building products from the ground up to navigating the complexities of GTM delivery. I’ve learned that engineering excellence is unparalleled when we achieve that perfect alignment with how we create, sell, market, and deliver value. From a historic IPO, accelerating the SaaS transition, to now embarking on a new era with Thoma Bravo and Conga - it has been a deeply fulfilling and rewarding ride. Backed by mentors turned friends for life who challenged, shaped, and inspired me, to teams and colleagues who consistently redefined what possible means, these experiences have been the foundation of everything to come. Above all, this journey has taught me that with the right team and a powerful vision, anything is truly achievable. Which is exactly how I feel about this exciting next step at Conga. Brilliant teams and a unique opportunity to redefine the landscape of enterprise software by bringing contracting, quoting, document automation, and pricing together on one AI-driven powerhouse of a platform! I’m looking forward to leading our global engineering teams as we innovate, scale, and deliver exceptional value to our customers. To my PROS family: thank you for being an indelible part of my story. To my new Conga colleagues: let’s build something brilliant together! 🚀
-
Somdyuti Paul liked thisSomdyuti Paul liked thisHey! Did you know #GoogleCloud has a #Database that can do sub-millisecond response time? No not that one. No, jeez the other one. Yeah, I'm talking about #Bigtable this time! If you like #NoSQL keystore data, that scales up beyond your imagination? (Okay, maybe not BEYOND your imagination, I know some of you have really impressive imaginations) Look no further. I made a couple videos about it. One on what it is and why you might want to check it out, and another on a little under the covers peek about how it does what it does. Check out the comments below for the individual links to the videos.
-
Somdyuti Paul reacted on thisSomdyuti Paul reacted on thisI’m happy to share that I’ve started a new role as Enterprise Architect at TenneT Netherlands . After more than 12 years with APM Terminals and A.P. Moller - Maersk , I look back with gratitude on a period of fun, transformation and collaboration across global logistics, industrial technology and terminal operations. I’m excited to now contribute to TenneT ‘s mission in the energy transition and to take on this new challenge in a sector that is critical to everyday life.
-
Somdyuti Paul reacted on thisAt Google Cloud Next '26, AlloyDB demonstrated its presence as a critical piece of modern data strategy, built specifically for the agentic era! The "AlloyDB Everywhere" story has resonated deeply with our customers and partners. By leveraging AlloyDB Omni, organizations are gaining the flexibility to run a high-performance, 100% PostgreSQL-compatible database that has cutting edge vector capabilities across cloud & on-premises/multi-cloud in hybrid environments. The energy at our showcase booth and Birds of a Feather session were testaments to this demand, with numerous customer conversations about AlloyDB & Omni ! In our What’s new in AlloyDB session by Ravi Murthy and Sujatha Mandava, our esteemed customers Walmart and Barclays shared their journey with AlloyDB. If you missed it, you can catch the full session here: https://lnkd.in/g5yrJ8Ya - Riaan Postma - Managing Director, Group Head of Architecture and Platform Engineering @ Barclays joined us to discuss their strategic selection of AlloyDB Omni (their first GCP product) as their platform of choice for PostgreSQL, and their plans for using Omni for their AI use cases. https://lnkd.in/gDD48UZ2 - Bhaven Avalani - VP of Global Tech Platform @ Walmart showcased AlloyDB as the database of choice on GCP and how AlloyDB and Walmart partnered to deploy a complex retail catalog application on GCP. https://lnkd.in/gXtpHxnU Thank you to our incredible customers for sharing their innovation on the big stage! Whether it's scaling for agentic AI or optimizing price-performance for massive datasets, AlloyDB as a unified data store is proving to be the cornerstone of modern data strategy! If you want to explore AlloyDB in the cloud, I recommend starting a free trial: https://lnkd.in/gCHBejMs To try AlloyDB Omni, checkout the various deployment options:https://lnkd.in/gZBesb5s. (now with RPM option with orchestrators in Preview!) #GoogleCloudNext #AlloyDB #PostgreSQL #GenerativeAI #CloudTransformation #Database #Walmart #Barclays #AlloyDBomni #omni #GCP
-
Somdyuti Paul liked thisSomdyuti Paul liked thisMeet me at Cloud Next 2026 if you want to talk about databases. #googlecloudnext #cloudsql
-
Somdyuti Paul liked thisSomdyuti Paul liked thisI had a fantastic time at Google Cloud Next last week - presenting our recent database innovations, engaging with many customers and partners, and having a lot of fun ! Thanks to my co-presenters in the "What's new in AlloyDB for PostgreSQL" session: Sujatha Mandava, Bhaven Avalani (Walmart) and Riaan Postma (Barclays). You can watch the session recording at: https://lnkd.in/gTcbnXerWhat's new in AlloyDB: Scale PostgreSQL for agentic AI and hybrid cloudsWhat's new in AlloyDB: Scale PostgreSQL for agentic AI and hybrid clouds
-
Somdyuti Paul liked thisSomdyuti Paul liked thisA little delayed in sharing, but I am excited to share that I moved from SAP to PwC last month, joining as a Partner in Advisory focused on Cloud Engineering and Data Analytics. At PwC, I am partnering with clients to design and deliver large-scale cloud, data (and yes, ai) transformation programs helping organizations move beyond incremental improvements to fundamentally reimagine how they operate. A key part of my role is working with leadership teams to connect technology decisions to measurable business outcomes and to ensure transformation is not just delivered, but adopted, scaled, and sustained. My time at SAP has been deeply fulfilling - leading large-scale cloud and infrastructure transformations/operation, working with exceptional teams, and driving initiatives that connected technology with real business outcomes. I am grateful for the trust, the challenges, and the opportunity to collaborate. I am looking forward to what’s ahead at PwC bringing cloud, data/ai, and platform-led innovation together to deliver value clients can see and measure. Thank you to the colleagues, mentors, and teams at SAP who shaped this journey. Super excited for the next chapter. #Leadership #AI #Cloud #Transformation #PwC #SAP
Experience
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
Education
-
-
-
-
-
-
-
Licenses & Certifications
-
-
-
OCP 11G, Exadata Implementation Specialist, RAC Certified, IBM Senior Certified L2 Specialist
Oracle
IssuedCredential ID Or23756 -
-
-
-
Recommendations received
32 people have recommended Somdyuti
Join now to viewOther similar profiles
Explore more posts
-
Shanmuga Sunthar Muniandy
Ready to simplify and accelerate your #DenodoPlatform deployments? Modern #Kubernetes deployments can be complex — but Helm Charts make them fast, scalable, and repeatable. 🧠💡 Watch on-demand this #TechTalk session “Accelerate Your Denodo Journey with Helm Charts” to learn how to streamline your setup, reduce manual effort, and enable cloud-native agility with Denodo. 🎙️ Featuring: - Abbas Kessai, Senior Technical Consultant, Denodo - Guillermo Sandino Leira, Senior Technical Lead, Denodo 💻 What you’ll learn: ✅ The fundamentals of Helm and #HelmCharts for #cloud-native deployments ✅ How Denodo Helm Charts simplify the setup of a complete Denodo environment ✅ A live demo showing how to get your Denodo environment up and running in minutes 📅 Don’t miss this hands-on session to see Helm in action and take your Denodo journey to the next level! 👉 https://okt.to/C1AE3o #DataManagement #DevOps #Scalability #NoDataLeftBehind
-
Carlos Andres Marcelo
🚀 Big news! Oracle Database@AWS is now generally available. This means customers can now run Oracle's industry-leading database on AWS with the same performance, availability, and manageability as on Oracle Cloud Infrastructure. A major step forward for hybrid cloud flexibility and multicloud strategies. 🔗 Oracle Database@AWS Now Generally Available https://lnkd.in/eaKpNxSf #Oracle #AWS #Cloud #Multicloud #Database #Innovation #OCI
-
Jason Beattie
Day 11 of the Oracle Linux Basic Administration Series – Configuring and Securing SSH Access Welcome to Day 11 of the Oracle Linux Basic Administration Series! Today, we move into one of the most critical parts of system administration, securing remote access with SSH (Secure Shell). SSH is how administrators safely connect to remote servers, and securing it properly is essential to protecting your system. In this post, you’ll learn how to: Install and enable SSH on Oracle Linux Configure key-based authentication (passwordless login) Harden SSH settings for better security Change ports, disable root login, and limit access by user or IP Use the firewall to restrict SSH traffic These are the same steps used in real-world server environments to prevent unauthorized access and reduce attack risks. 👉 Read the full post: https://lnkd.in/eiDsj-9w Follow along daily! each topic builds on the last to help you grow from Linux beginner to confident system admin. #OracleLinux #LinuxAdministration #SysAdmin #SSH #ServerSecurity #LearningSeries #BlogSeries #LinuxForBeginners #TechForStarters #DevOps #SystemAdmin
-
Bolloju Yathindra
🧠 Kubernetes Architecture — Simplified Kubernetes may look complex at first, but at its core, it’s just a well-coordinated control system. Here’s how things actually work 👇 🔹 Control Plane This is the brain of the cluster. API Server → Entry point for all requests etcd → Stores the cluster state Scheduler → Decides where Pods should run Controller Manager → Keeps the desired state in sync 🔹 Worker Nodes This is where your applications run. Kubelet → Talks to the control plane Kube Proxy → Handles networking Pods → Run your containers using CRI 🔹 kubectl & Cloud Provider kubectl communicates with the API Server Cloud Controller integrates Kubernetes with cloud services 💡 Simple idea: You tell Kubernetes what you want. Kubernetes figures out how to make it happen and keeps it that way. That’s the power of Kubernetes architecture. #Kubernetes #DevOps #CloudNative #Containers #K8s #PlatformEngineering #CloudComputing
1 Comment -
StackBooster
𝗬𝗼𝘂𝗿 𝗞𝘂𝗯𝗲𝗿𝗻𝗲𝘁𝗲𝘀 𝗰𝗹𝘂𝘀𝘁𝗲𝗿 𝗶𝘀 𝗹𝗲𝗮𝗸𝗶𝗻𝗴 𝗺𝗼𝗻𝗲𝘆 — 𝗵𝗲𝗿𝗲’𝘀 𝗵𝗼𝘄 𝘁𝗼 𝗳𝗶𝗻𝗱 𝗼𝘂𝘁 𝘄𝗵𝗲𝗿𝗲. Most Kubernetes teams think they’re optimizing costs — but in reality, up to 65% of allocated resources go unused. Why? Because they measure allocations, not real usage. Here’s how to spot where your Kubernetes resources are leaking: - Track pod-level and cluster-level utilization over time. - Identify idle nodes and overprovisioned workloads. - Use intelligent automation to right-size resources continuously. When visibility meets automation, efficiency follows. That’s the philosophy behind StackBooster. We are an AI-agent for Kubernetes that optimizes performance with up to 80% cost reduction. Want to dive deeper? The full version of this article breaks down how to measure real usage, spot hidden leaks, and automate resource optimization step-by-step. 👉 Link in comments #Kubernetes #DevOps #CloudOptimization #PlatformEngineering #AIAutomation #StackBooster
1 Comment -
Mhammed HABIBALLAH
As part of my ongoing series on PostgreSQL database administration, I am sharing #Part 5, which focuses on a fundamental DBA responsibility: governing roles, privileges, and access control. This video provides a structured and practical overview of: Granting privileges using GRANT Safely revoking permissions with REVOKE Designing and managing PostgreSQL roles Understanding role inheritance and its impact on security Best practices for enterprise-grade PostgreSQL access management The objective is to offer a clear, professional perspective on how PostgreSQL security should be implemented in real-world production environments. This content is intended for DBAs, software engineers, data professionals, and students looking to strengthen their expertise in PostgreSQL administration and database security. https://lnkd.in/dUs6tMRw #PostgreSQL #DatabaseAdministration #DBA #DatabaseSecurity #AccessControl #DataEngineering #SoftwareEngineering #ProfessionalDevelopment
-
FOSSTechNix
Secure SFTP Setup Using AWS Transfer Family. AWS Transfer Family offers fully managed support for Secure File Transfer Protocol (SFTP), enabling easy and secure file transfers directly into and out of Amazon S3. This article walks through setting up an SFTP server using AWS Transfer Family, covering everything from IAM configuration to user isolation, file transfers, and log verification #aws https://lnkd.in/dUY9jT78
Explore collaborative articles
We’re unlocking community knowledge in a new way. Experts add insights directly into each article, started with the help of AI.
Explore More