1 views collected around this technical thread.
This article explains the problem of internal and external memory fragmentation in Linux systems, introduces eBPF as a powerful tracing tool, and provides step‑by‑step guidance for building, loading, and running eBPF programs to analyze and mitigate fragmentation on both Linux and Android platforms.
This article examines MySQL slow‑query troubleshooting, explaining how the optimizer estimates costs, why indexes may be ineffective even when present, the impact of memory fragmentation, pitfalls of prefix indexes, index merging techniques, and additional resource‑related factors that can cause seemingly healthy SQL statements to become slow.
The article explains C’s memory allocation, detailing how functions like malloc obtain heap space, the role of system calls such as brk, the distinction between stack and heap, address layout, fragmentation challenges, and practical implications for efficient dynamic memory management.
Linux’s memory compaction mitigates external fragmentation by moving movable pages, employing four strategies—direct, passive (kcompactd), proactive, and active—each invoking the compact_zone core with configurable compact_control parameters, migrate‑page and free‑page scanners, and distinct trigger and exit conditions.
This article investigates the performance degradation caused by DPDK memory fragmentation, explains the underlying heap and element management, presents detailed profiling and test results, and proposes two practical solutions that dramatically reduce allocation latency and improve overall system throughput.
This article explains why Redis may show high OS‑allocated memory despite low data usage, describes the causes of memory fragmentation, and provides practical steps—including configuration changes and automatic defragmentation settings—to diagnose and mitigate fragmentation for stable performance.
The article analyzes a recurring one‑second pause in a Redis 4.0 master‑slave setup, identifies the periodic BGSAVE fork operation as the root cause through metrics like latest_fork_usec, and presents mitigation strategies such as memory limits, activedefrag, and migration to a lower‑RSS instance.
This article explains why a low Redis memory fragmentation ratio is not necessarily caused by swap usage, demonstrates experiments with disabled swap and varying repl‑backlog‑size, and concludes that large replication buffers and small data sets can naturally produce ratios far below 1, which is normal.
This article examines typical C/C++ memory problems, classifying them into memory leaks and performance issues caused by fragmentation, explains root causes such as lost pointers, improper releases, unbounded allocations, and provides diagnostic examples and mitigation strategies using memory pools like tcmalloc.