| Summary: | [DOC] Obsolete information in the Guidelines for using large log support topic | ||
|---|---|---|---|
| Product: | z_Archived | Reporter: | Alex Nan <apnan> |
| Component: | TPTP | Assignee: | Alex Nan <apnan> |
| Status: | CLOSED FIXED | QA Contact: | |
| Severity: | normal | ||
| Priority: | P1 | CC: | ewchan, jkubasta, labadie, slavescu, smith |
| Version: | unspecified | Keywords: | plan |
| Target Milestone: | --- | ||
| Hardware: | PC | ||
| OS: | Windows 2000 | ||
| Whiteboard: | |||
|
Description
Alex Nan
This defect also belongs to the Platform doc actually but we should probably have a component Platform.Log.Doc part of the LogAnalyzer product for documentation topics related only to the LogAnalyzer. I have changed the doc to the following: XMI and large log support can be used together in a single session. However, mixing the two modes of resources may not perform as well as using a single mode operation. Using a combination of the modes may result in performance degradation and excess memory consumption. Follow these guidelines when running with mixed mode of log resources: - All types of correlations are supported when running in large log support mode. - If you have a list of resources and one is an XMI resource, Log actions can only be performed on resources of the same mode, you will be asked to select the mode of resource for the actions to apply and limited to the set of resources of the selected mode. Actions that can trigger a single-mode operation include opening a view, opening a log interaction, or a correlation of mixed resources. As well, opening a monitor, correlation, or a node with mixed resources will result in a single-mode operation. If you need to switch to a different mode of operation, rerun the action again. fixed. please reopen if there is problem with the updated content. Verified with TPTP-4.4.0-200705080100C. I don't think the updated information is clear. I know what it's trying to say but honestly I don't think it's sufficiently clear for a user. There is another thing which I must have overlooked the last time I have reviewed this chapter, I'm talking about the paragraph "A log resource is active when it is loaded into memory by any action that results in displaying the log resource in a view. The large log support mode should be used when working with active log files whose combined size on import is greater than 4% of the total Java heap size(default size is 512MB). This recommendation is based on tests performed on a Pentium III (1GB RAM, 1GHZ CPU) running the Log and Trace Analyzer.". I would like to hear your opinion about this paragraph, I think the information is wrong, because active log size according to the explanation is the size in memory of the imported log, but the 4% really applies to the size of the raw log file before it is imported into the workbench. So if you have logs with sizes (in raw format) summing up to 20MB then large resource should be turned on. What happens is in general transformed into CBE the logs expand with a factor of at least 10, so 20MB of raw log would generate at least 200MB of CBE, so if you have 512MB of heap it's safer to switch to large resource and import the log files into a database. Eugene, you can assign this defect to me, I'll try to rephrase, with the help of the people cc-ed on this defect, the information in this chapter. Alex. I am reassigning this to you as requested. Please target this defect as necessary. This is not a stop ship issue for 4.4 - deferring to future. This is nice to have from AC perspective We cannot do any documentation changes in 4.4.1 since translations are frozen, targetting to 4.5. Fixed. Verified with TPTP-4.5.0 GA. Closing. |