1Faq - Frequently Asked Questions 2 3 The following may serve as a guide to the DTraceToolkit. 4 516-May-2005, ver 0.30 (first version of the FAQ) 6 7The DTraceToolkit is new, and as such there hasn't been many questions asked. 8This may be better called a "possibly asked questions" :) 9 10 11Questions 12 131. Intro 141.1. What is the DTraceToolkit? 151.2. Who wrote the DTraceToolkit? 161.3. Where do I get support? 171.4. Am I now a performance tuning expert? 181.5. Will this solve all my performance problems? 191.6. So the DTraceToolkit *is* DTrace? 20 212. Toolkit 222.1. What is in it? 232.2. What performance effect can the DTraceToolkit cause? 24 253. Contributing 263.1. Where do I send bugs? 27 28 29Answers 30 311. Intro 32 331.1. What is the DTraceToolkit? 34 35 The DTraceToolkit is a collection of tools written using DTrace for 36 the Solaris 10[tm] OS by Sun Microsystems[tm]. Many of these scripts 37 will also work on OpenSolaris. 38 391.2. Who wrote the DTraceToolkit? 40 41 Volunteers of the DTrace and OpenSolaris community. Check the scripts 42 themselves, Docs/Contrib, Docs/Who and Docs/History. 43 441.3. Where do I get support? 45 46 As the DTraceToolkit is a freeware product, there is no official company 47 offering support for this. Sun Microsystems does not support this. If you 48 post messages to the DTrace forums found in the Docs/Links file, a 49 volunteer may help you out. 50 511.4. Am I now a performance tuning expert? 52 53 The DTraceToolkit does not turn people into performance tuning experts in 54 the same way that owning a set of golf clubs won't make you a professional 55 golfer. Experience and understanding are necessary. The toolkit certainly 56 helps by fetching the data in an easy way, and also by providing some 57 documentation. So it is valuable, but not magical. 58 591.5. Will this solve all my performance problems? 60 61 This is similar to the previous point; the DTraceToolkit is valuable 62 for it's scripts and documentation, but it's no magical product. 63 Understanding and experience are necessary. 64 651.6. So the DTraceToolkit *is* DTrace? 66 67 The DTraceToolkit is one use of DTrace, but there is far more to DTrace 68 than just the toolkit. DTrace allows people to write their own customised 69 scripts to solve a wide number of problems. 70 71 Think of the DTraceToolkit as a starting point. Maybe your problem has 72 a solution in the kit. Maybe changing one of the toolkit programs slightly 73 is what you want. Finally you may need to write your script from scratch. 74 75 762. Toolkit 77 782.1. What is in it? 79 80 Read the Guide file for a table of contents, and Docs/Contents for a 81 list of commands. 82 832.2. What performance effect can the DTraceToolkit cause? 84 85 Enabling DTrace to monitor events has little effect on the system, 86 especially when compared to the disruptive behaviour of truss (See 87 http://www.brendangregg.com/DTrace/dtracevstruss.html for a comparison). 88 89 It really boils down to how often the events occur that you are monitoring. 90 The following numbers have been provided as an approximation: 91 92 1. Fixed rate scripts. For example, dispqlen.d samples at 1000 hz. 93 The impact will be negligible, close to 0% CPU. (in testing, 0.1% CPU). 94 95 2. Demand rated scripts. For example, iosnoop probes disk I/O events. 96 The impact depends on the rate of events, for many servers the disk 97 events would be slow enough for this to be less than 0.2% CPU. 98 Scripts such as execsnoop would expect even fewer events, their impact 99 would be close to 0.0% CPU. However scripts that monitor potentially 100 very rapid events will have a greater impact, for example running 101 dapptrace on Xorg (over 6000 lines of output per second) was consuming 102 around 10% of a CPU to do so. 103 104 3. Heavy voodoo scripts. A few scripts in the toolkit must probe either 105 a ton of different events, or very rapid events, or both. They are 106 going to hurt and there is no way around it. Scripts such as cputimes 107 and cpudists trace very frequent events, and can chew around 5% of 108 the CPUs; scripts such as dapptrace and dappprof trace extreamly 109 frequent events, and can chew over 20%. 110 111 There is an emphasis in the DTraceToolkit to write demand rated scripts 112 that measure the fewest events, such that their impact is close to 0.0% 113 CPU usage. Some scripts are fixed rate, which are safer as their impact 114 has a known upper bound, and are most suitable to run in production. 115 116 There are additional notes in Notes/ALLoverhead_notes.txt about the 117 overheads for running DTrace. 118 119 1203. Contributing 121 1223.1. Where do I send bugs? 123 124 The DTraceToolkit maintainer. See the Docs/Maintainer file. 125 126 127