Linux Operating System Reliability Technology Evaluation Test-general Linux technology-Linux programming and kernel information, the following is a detailed description. This article records the test results and analysis of the Linux kernel and other core OS components, from the library and device drivers to the file system and network, the test scope does not include, all tests were conducted under quite unfavorable conditions and took a long time. IBM Linux Technology Center has just concluded this three-month comprehensive Test and shared the results of their LTP (Linux Test Project) Test with readers on developerWorks.
The IBM Linux Technology Center (LTC) was established in August 1999 to allow the success of Linux to work directly with the Linux development group. Its more than 200 employees make it one of the largest open source code developers team. They provide code ranging from patching to structured kernel changes, from file systems and internationalization to GPL 'd drivers. They are also committed to tracking Linux-related development within IBM.
LTC is particularly interested in Linux scalability, applicability, reliability, and system management-all aimed at making Linux more suitable for enterprises. They have made many contributions to the Linux community, including enabling Linux to work on S/390 hosts and porting JFS log file systems to Linux.
Another core task of LTC is to perform professional testing on Linux in lab conditions by testing commercial projects. LTC, together with SGI, OSDL, Bull, and Wipro Technologies, contributed to the LTP Linux testing project (LTP ). The following is the result of a long and comprehensive test on the Linux kernel by using the LTP kit. As you may guess, Linux has withstood the continuous stress test.
Test Results
The following summary is based on the test and observation results during running:
* The Linux kernel and other core OS components-including libraries, device drivers, file systems, networks, IPC and memory management-run stably and complete all the expected runtime periods, no serious system faults.
* Each operation has a high success rate (over 95%) and only a small number of expected intermittent faults, which are the results of testing designed to overload resources at the same time.
* Linux system performance does not decrease during long running.
* In the SMP system, the Linux kernel is correctly extended to use hardware resources (CPU, memory, and hard disk ).
* The Linux system can well withstand the continuous full-load (over 99%) and extremely heavy memory pressure of the CPU.
* The Linux system correctly handles the overload.
The test proves that the Linux kernel and other core components are reliable and stable within 30 days, 60 days, and 90 days, it can provide users with a robust, enterprise-level environment that runs for a long time.
Linux reliability Measurement
Target
For IBM Linux Technology Center, the objective of Linux reliability is to use the LTP test suite to test the Linux operating system for a long time, the focus is on the workload related to the Linux User Environment (see references to learn more about LTP ). Instead of proving defects.
Test environment overview
This document describes the test results and analysis of Linux reliability measurement tests over the last 30 and 60 days using the LTP test suite. The SuSE Linux Enterprise Server v8 (SLES 8) is used as the testing kernel and the IBM pSeries Server is used as the testing hardware. A specially designed LTP stress test scenario is used to concurrently run a wide range of kernel components while using network and memory management, and generate high workload pressure on the test system. The Linux kernel, TCP, NFS, and I/O test components are designed for heavy workloads.
Test
30 days
PSeries 30-day LTP pressure execution result
* Machine: p650 LPAR
* CPU: (2) Power4-1.2 GHz
* Kernel: Linux 2.4.19-ull-ppc64-SMP (SLES 8 SP 1)
* LTP version: 20030514
* 99.00% average CPU usage (User: 48.65%, system: 50.35%)
* 80.09% average memory usage (8 GB)
Observed results:
* On the p650 LPAR, SLES 8 PPC64 runs successfully within 30 days.
* The test tool is LTPstress. Test Cases are executed in parallel and serial mode.
* The kernel, TCP, NFS, and I/O test components are all aimed at heavy workload.
* Successful rate: 97.88%.
* No serious system faults.
(400) {this. resized = true; this. width = 400; this. alt = 'click here to open new window';} "onmouseover =" if (this. resized) this. style. cursor = 'hand'; "onclick =" window. open ('HTTP: // linux.ccidnet.com/col/attachment/2006/12/971473.gif'); ">
60 days
60-day LTP pressure execution result: pSeries
* Machine: B80
* CPU: (2) Power3-375 MHz
* Kernel: Linux 2.4.19-ull-ppc64-SMP (SLES 8 SP 1)
* LTP version: 20030514
* 99.96% average CPU usage (User: 75.02%, system: 24.94%)
* 61.69% average memory usage (8 GB)
* 3.86% average swap partition utilization (1 GB)
Observed results:
* On pSeries B80, SLES 8 PPC64 runs successfully within 60 days.
* The test tool is LTPstress. Test Cases are executed in parallel and serial mode.
* The kernel, TCP, NFS, and I/O test components are designed for heavy workloads.
* Successful rate: 95.12%.
* No serious system faults.
(400) {this. resized = true; this. width = 400; this. alt = 'click here to open new window';} "onmouseover =" if (this. resized) this. style. cursor = 'hand'; "onclick =" window. open ('HTTP: // linux.ccidnet.com/col/attachment/2006/12/971475.gif'); ">
Testing infrastructure
Hardware and software environment
System processor memory hard disk swap partition Network
PSeries 650 (LPAR) Model 7038-6M2 2-POWER4 + (TM) 1.2 GHz 8 GB (8196 MB) 36 GB U320 IBM Ultrastar (other hard drives available but not used) 1 GB Ethernet controller: AMD PCnet32
PSeries 630 Model 7026-B80 2-POWER3 (TM) + 375 MHz 8 GB (7906 MB) 16 GB 1 GB Ethernet controller: AMD PCnet32
The software environment on the pSeries 630 Model 7026-B80 and pSeries 650 (LPAR) Model 7038-6M2 is the same.
Component version
Linux SuSE SLES 8 with Service Pack 1
Kernel 2.4.19-ul1-ppc64-SMP
LTP 20030514
Methodology
The stability and reliability of the system are usually measured by the continuous running time and the reliable running time of the system.
Initially, a group of 30-day baseline runs, and then increases to 60 and 90-day Linux testing run on xSeries and pSeries servers. The initial focus is on kernel, network, and I/O testing.
