本文整理匯總了Golang中github.com/elastic/beats/filebeat/harvester.Harvester.Start方法的典型用法代碼示例。如果您正苦於以下問題:Golang Harvester.Start方法的具體用法?Golang Harvester.Start怎麽用?Golang Harvester.Start使用的例子?那麽, 這裏精選的方法代碼示例或許可以為您提供幫助。您也可以進一步了解該方法所在類github.com/elastic/beats/filebeat/harvester.Harvester
的用法示例。
在下文中一共展示了Harvester.Start方法的2個代碼示例,這些例子默認根據受歡迎程度排序。您可以為喜歡或者感覺有用的代碼點讚,您的評價將有助於係統推薦出更棒的Golang代碼示例。
示例1: checkExistingFile
// checkExistingFile checks if a harvester has to be started for a already known file
// For existing files the following options exist:
// * Last reading position is 0, no harvester has to be started as old harvester probably still busy
// * The old known modification time is older then the current one. Start at last known position
// * The new file is not the same as the old file, means file was renamed
// ** New file is actually really a new file, start a new harvester
// ** Renamed file has a state, continue there
func (p ProspectorLog) checkExistingFile(h *harvester.Harvester, newFile *input.File, oldFile *input.File) {
logp.Debug("prospector", "Update existing file for harvesting: %s", h.Path)
// We assume it is the same file, but it wasn't
if !oldFile.IsSameFile(newFile) {
logp.Debug("prospector", "File previously found: %s", h.Path)
if previousFile, err := p.getPreviousFile(h.Path, h.Stat.Fileinfo); err == nil {
p.continueExistingFile(h, previousFile)
} else {
// File is not the same file we saw previously, it must have rotated and is a new file
logp.Debug("prospector", "Launching harvester on rotated file: %s", h.Path)
// Forget about the previous harvester and let it continue on the old file - so start a new channel to use with the new harvester
h.Stat.Ignore()
// Start a new harvester on the path
h.Start()
}
// Keep the old file in missingFiles so we don't rescan it if it was renamed and we've not yet reached the new filename
// We only need to keep it for the remainder of this iteration then we can assume it was deleted and forget about it
p.missingFiles[h.Path] = oldFile.FileInfo
} else if h.Stat.Finished() && oldFile.FileInfo.ModTime() != h.Stat.Fileinfo.ModTime() {
// Resume harvesting of an old file we've stopped harvesting from
// Start a harvester on the path; a file was just modified and it doesn't have a harvester
// The offset to continue from will be stored in the harvester channel - so take that to use and also clear the channel
p.resumeHarvesting(h, <-h.Stat.Return)
} else {
logp.Debug("prospector", "Not harvesting, file didn't change: %s", h.Path)
}
}
示例2: resumeHarvesting
// Start / resume harvester with a predefined offset
func (p *ProspectorLog) resumeHarvesting(h *harvester.Harvester, offset int64) {
logp.Debug("prospector", "Start / resuming harvester of file: %s", h.Path)
h.SetOffset(offset)
h.Start()
}