process-ram.drawio.svg...

當程式把東西載入記憶體時，例如把機器學習模型存到變數中，或把大型檔案內容讀到變數中，這些都會「消耗一些伺服器的記憶體（RAM）」。

而多個行程通常「不共享記憶體」。這表示每個執行中的行程都有自己的東西、變數與記憶體。如果你的程式碼會用掉大量記憶體，「每個行程」都會消耗等量的記憶體。

### 伺服器記憶體 { #server-memory }

例如，如果你的程式碼載入一個「1 GB 大小」的機器學習模型，當你用一個行程執行你的 API，它就會至少吃掉 1 GB 的 RAM。若你啟動「4 個行程」（4 個 workers），每個會吃 1 GB RAM。總計你的 API 會吃掉「4 GB RAM」。

如果你的遠端伺服器或虛擬機只有 3 GB RAM，嘗試載入超過 4 GB 的 RAM 就會出問題。🚨

### 多個行程 - 範例 { #multiple-processes-an-example }

And if your remote server or virtual machine only has 3 GB of RAM, trying to load more than 4 GB of RAM will cause problems. 🚨

### Multiple Processes - An Example { #multiple-processes-an-example }

To summarize for any encrypted object there exists (at least) three different keys:

- [OEK](#oek): A secret and unique key used to encrypted the object, stored in an encrypted form as part of the object metadata and only loaded to RAM in plaintext during en/decrypting the object.
- [KEK](#kek): A secret and unique key used to en/decrypt the OEK and never stored anywhere. It is(re-)generated whenever en/decrypting an object using an external secret key and public parameters.

```

or

```sh
mc admin config set myminio/ api requests_max=1600
mc admin service restart myminio/
```

audio/basic                    au snd           # Audio Sound File
audio/x-aiff                   aif aiff aifc    # AIFF Sound File
audio/x-midi                   mid              # MIDI Sound File
audio/x-pn-realaudio           ra ram rm rpm    # REALAUDIO Sound File
audio/x-wav                    wav              # WAV Sound File
audio/x-mpegurl                mp3              # MP3 Sound File
audio/mpeg3                    mp3              # MP3 Sound File

}

var (
	// https://github.com/golang/lint/blob/master/lint.go#L770
	commonInitialisms         = []string{"API", "ASCII", "CPU", "CSS", "DNS", "EOF", "GUID", "HTML", "HTTP", "HTTPS", "ID", "IP", "JSON", "LHS", "QPS", "RAM", "RHS", "RPC", "SLA", "SMTP", "SSH", "TLS", "TTL", "UID", "UI", "UUID", "URI", "URL", "UTF8", "VM", "XML", "XSRF", "XSS"}
	commonInitialismsReplacer *strings.Replacer
)

func init() {

mapreduce.job.reduce.slowstart.completedmaps=0.99 # 99% map, then reduce
mapreduce.reduce.shuffle.input.buffer.percent=0.9 # Min % buffer in RAM
mapreduce.reduce.shuffle.merge.percent=0.9 # Minimum % merges in RAM
mapreduce.reduce.speculative=false # Disable speculation for reducing
mapreduce.task.io.sort.factor=999 # Threshold before writing to disk

* the mobile application
* the module
* the mounting
* the network
* the origin
* the override
* the payload
* the processor
* the property
* the proxy
* the pull request
* the query
* the RAM
* the remote machine
* the status code
* the string
* the tag
* the web framework
* the wildcard
* to return
* to validate

////

//// tab | Info

- **User Permissions**: Access control and label configurations

### 2. Infrastructure Requirements

Ensure your Fess environment meets these requirements:

- **Java**: JDK 17 or later
- **Memory**: Minimum 4GB RAM (8GB+ recommended for production)
- **Storage**: At least 2x your current index size
- **Network**: Access to crawl sources (web servers, file shares, databases)
- **Elasticsearch/OpenSearch**: Compatible version running