• However, when using globals, we still recommend using npm, especially if you use nvm (Node Version Manager). Use git for repository management and make regular commits, it is like taking notes on the process and lets you revert to a previous state in case you get stuck. Use Quasar boot files for any pre-mounting app routines.
• I'd say you should encapsulate as much of your code as possible into functions and classes, limiting the global state when possible. This serves two purposes - the first is that it improves debugging (by limiting the odds that something unintentionally alters global state) and readability (by making it easier to understand what everything does).

Overview

This guide explains how to migrate to safe Buffer constructor methods. The migration fixes the following deprecation warning:

In this guide you will learn how to import a node Node-RED flow. What is Node RED? Node-RED is a visual tool for wiring the Internet of Things developed by IBM Emerging Technology and the open source community. Using Node-RED, developers wire up input, output and processing nodes to create flows to process data, control things, or send alerts.

• Variant 1: Drop support for Node.js ≤ 4.4.x and 5.0.0 — 5.9.x (recommended)

Finding problematic bits of code using grep

Just run grep -nrE '[^a-zA-Z](Slow)?Buffers*(' --exclude-dir node_modules.

It will find all the potentially unsafe places in your own code (with some considerably unlikelyexceptions).

Finding problematic bits of code using Node.js 8

If you’re using Node.js ≥ 8.0.0 (which is recommended), Node.js exposes multiple options that help with finding the relevant pieces of code:

• --trace-warnings will make Node.js show a stack trace for this warning and other warnings that are printed by Node.js.
• --trace-deprecation does the same thing, but only for deprecation warnings.
• --pending-deprecation will show more types of deprecation warnings. In particular, it will show the Buffer() deprecation warning, even on Node.js 8.

You can set these flags using environment variables:

Finding problematic bits of code using linters

ESLint rules no-buffer-constructorornode/no-deprecated-apialso find calls to deprecated Buffer() API. Those rules are included in some presets.

There is a drawback, though, that it doesn't alwayswork correctly when Buffer isoverridden e.g. with a polyfill, so recommended is a combination of this and some other methoddescribed above.

Variant 1: Drop support for Node.js ≤ 4.4.x and 5.0.0 — 5.9.x

This is the recommended solution nowadays that would imply only minimal overhead.

The Node.js 5.x release line has been unsupported since July 2016, and the Node.js 4.x release line reaches its End of Life in April 2018 (→ Schedule). This means that these versions of Node.js will not receive any updates, even in case of security issues, so using these release lines should be avoided, if at all possible.

What you would do in this case is to convert all new Buffer() or Buffer() calls to use Buffer.alloc() or Buffer.from(), in the following way:

• For new Buffer(number), replace it with Buffer.alloc(number).
• For new Buffer(string) (or new Buffer(string, encoding)), replace it with Buffer.from(string) (or Buffer.from(string, encoding)).
• For all other combinations of arguments (these are much rarer), also replace new Buffer(...arguments) with Buffer.from(...arguments).

Note that Buffer.alloc() is also faster on the current Node.js versions thannew Buffer(size).fill(0), which is what you would otherwise need to ensure zero-filling.

Enabling ESLint rule no-buffer-constructorornode/no-deprecated-apiis recommended to avoid accidental unsafe Buffer API usage.

There is also a JSCodeshift codemodfor automatically migrating Buffer constructors to Buffer.alloc() or Buffer.from().Note that it currently only works with cases where the arguments are literals or where theconstructor is invoked with two arguments.

If you currently support those older Node.js versions and dropping support for them is not possible, or if you support older branches of your packages, consider using Variant 2or Variant 3 on older branches, so people using those older branches will also receivethe fix. That way, you will eradicate potential issues caused by unguarded Buffer API usage andyour users will not observe a runtime deprecation warning when running your code on Node.js 10.

Variant 2: Use a polyfill

There are three different polyfills available:

• safer-buffer is a drop-in replacement for theentire Buffer API, that will throw when using new Buffer().

  You would take exactly the same steps as in Variant 1, but with a polyfillconst Buffer = require('safer-buffer').Buffer in all files where you use the new Buffer API.

  Do not use the old new Buffer() API. In any files where the line above is added,using old new Buffer() API will throw.

• buffer-from and/orbuffer-alloc areponyfills for their respective part of the Buffer API. You only needto add the package(s) corresponding to the API you are using.

  You would import the module needed with an appropriate name, e.g.const bufferFrom = require('buffer-from') and then use that instead of the call tonew Buffer(), e.g. new Buffer('test') becomes bufferFrom('test').

  A downside with this approach is slightly more code changes to migrate off them (as you would beusing e.g. Buffer.from() under a different name).

• safe-buffer is also a drop-in replacement forthe entire Buffer API, but using new Buffer() will still work as before.

  A downside to this approach is that it will allow you to also use the older new Buffer() APIin your code, which is problematic since it can cause issues in your code, and will startemitting runtime deprecation warnings starting with Node.js 10(read more here).

Note that in either case, it is important that you also remove all calls to the old BufferAPI manually — just throwing in safe-buffer doesn't fix the problem by itself, it just providesa polyfill for the new API. I have seen people doing that mistake.

Enabling ESLint rule no-buffer-constructorornode/no-deprecated-apiis recommended.

Don't forget to drop the polyfill usage once you drop support for Node.js < 4.5.0.

Variant 3 — Manual detection, with safeguards

This is useful if you create Buffer instances in only a few places (e.g. one), or you have your ownwrapper around them.

Buffer(0)

This special case for creating empty buffers can be safely replaced with Buffer.concat([]), whichreturns the same result all the way down to Node.js 0.8.x.

Buffer(notNumber)

Before:

After:

encoding is optional.

Note that the typeof notNumber before new Buffer() is required (for cases when notNumber argument is nothard-coded) and is not caused by the deprecation of Buffer constructor — it's exactly why theBuffer constructor is deprecated. Ecosystem packages lacking this type-check caused numeroussecurity issues — situations when unsanitized user input could end up in the Buffer(arg) createproblems ranging from DoS to leaking sensitive information to the attacker from the process memory.

Warning Node Has Slots In Importing Staten Island

When notNumber argument is hardcoded (e.g. literal 'abc' or [0,1,2]), the typeof check canbe omitted.

Also, note that using TypeScript does not fix this problem for you — when libs written inTypeScript are used from JS, or when user input ends up there — it behaves exactly as pure JS, asall type checks are translation-time only and are not present in the actual JS code which TScompiles to.

Buffer(number)

For Node.js 0.10.x (and below) support:

Otherwise (Node.js ≥ 0.12.x):

Regarding Buffer.allocUnsafe()

Be extra cautious when using Buffer.allocUnsafe():

• Don't use it if you don't have a good reason to
  • e.g. you probably won't ever see a performance difference for small buffers, in fact, thosemight be even faster with Buffer.alloc(),
  • if your code is not in the hot code path — you also probably won't notice a difference,
  • keep in mind that zero-filling minimizes the potential risks.
• If you use it, make sure that you never return the buffer in a partially-filled state,
  • if you are writing to it sequentially — always truncate it to the actual written length

Errors in handling buffers allocated with Buffer.allocUnsafe() could result in various issues,ranged from undefined behavior of your code to sensitive data (user input, passwords, certs)leaking to the remote attacker.

Note that the same applies to new Buffer() usage without zero-filling, depending on the Node.jsversion (and lacking type checks also adds DoS to the list of potential problems).

FAQ

What is wrong with the Buffer constructor?

The Buffer constructor could be used to create a buffer in many different ways:

• new Buffer(42) creates a Buffer of 42 bytes. Before Node.js 8, this buffer containedarbitrary memory for performance reasons, which could include anything ranging fromprogram source code to passwords and encryption keys.
• new Buffer('abc') creates a Buffer that contains the UTF-8-encoded version ofthe string 'abc'. A second argument could specify another encoding: for example,new Buffer(string, 'base64') could be used to convert a Base64 string into the originalsequence of bytes that it represents.
• There are several other combinations of arguments.

This meant that in code like var buffer = new Buffer(foo);, it is not possible to tellwhat exactly the contents of the generated buffer are without knowing the type of foo.

Sometimes, the value of foo comes from an external source. For example, this functioncould be exposed as a service on a web server, converting a UTF-8 string into its Base64 form:

Note that this code does not validate the type of req.body.string:

• req.body.string is expected to be a string. If this is the case, all goes well.
• req.body.string is controlled by the client that sends the request.
• If req.body.string is the number50, the rawBytes would be 50 bytes:
  • Before Node.js 8, the content would be uninitialized
  • After Node.js 8, the content would be 50 bytes with the value 0

Because of the missing type check, an attacker could intentionally send a numberas part of the request. Using this, they can either:

• Read uninitialized memory. This will leak passwords, encryption keys and otherkinds of sensitive information. (Information leak)
• Force the program to allocate a large amount of memory. For example, when specifying500000000 as the input value, each request will allocate 500MB of memory.This can be used to either exhaust the memory available of a program completelyand make it crash, or slow it down significantly. (Denial of Service)

Both of these scenarios are considered serious security issues in a real-worldweb server context.

When using Buffer.from(req.body.string) instead, passing a number will alwaysthrow an exception instead, giving a controlled behavior that can always behandled by the program.

The Buffer() constructor has been deprecated for a while. Is this really an issue?

Surveys of code in the npm ecosystem have shown that the Buffer() constructor is stillwidely used. This includes new code, and overall usage of such code has actually beenincreasing.

Contents

Backup and Export Configuration

When you perform a backup through Cisco UCS Manager, you take a snapshot of all or part of the system configuration and export the file to a location on your network. You cannot use Cisco UCS Manager to back up data on the servers.

You can perform a backup while the system is up and running. The backup operation only saves information from the management plane. It does not have any impact on the server or network traffic.

Backup Types

You can perform one or more of the following types of backups through Cisco UCS Manager:

• Full state—A binary file that includes a snapshot of the entire system. You can use the file generated from this backup to restore the system during disaster recovery. This file can restore or rebuild the configuration on the original fabric interconnect, or recreate the configuration on a different fabric interconnect. You cannot use this file for an import.

• All configuration—An XML file that includes all system and logical configuration settings. You can use the file generated from this backup to import these configuration settings to the original fabric interconnect or to a different fabric interconnect. You cannot use this file for a system restore.

• System configuration—An XML file that includes all system configuration settings such as usernames, roles, and locales. You can use the file generated from this backup to import these configuration settings to the original fabric interconnect or to a different fabric interconnect. You cannot use this file for a system restore.

• Logical configuration—An XML file that includes all logical configuration settings such as service profiles, VLANs, VSANs, pools, and policies. You can use the file generated from this backup to import these configuration settings to the original fabric interconnect or to a different fabric interconnect. You cannot use this file for a system restore.

Considerations and Recommendations for Backup Operations

Before you create a backup operation, consider the following:

The backup location is the destination or folder on the network where you want Cisco UCS Manager to export the backup file. You can maintain only one backup operation for each location where you plan to save a backup file.

If you rerun a backup operation without changing the filename, Cisco UCS Manager overwrites the existing file on the server. To avoid overwriting existing backup files, change the filename in the backup operation or copy the existing file to another location.

You can run and export more than one type of backup to the same location. You need to change the backup type before you rerun the backup operation. We recommend that you change the filename for easier identification of the backup type and to avoid overwriting the existing backup file.

You cannot schedule a backup operation. You can, however, create a backup operation in advance and leave the admin state disabled until you are ready to run the backup. Cisco UCS Manager does not run the backup operation, save, or export the configuration file until you set the admin state of the backup operation to enabled.

You cannot perform incremental backups of the Cisco UCS Manager system configuration.

Starting with Release 1.1(1) of the Cisco UCS Manager, full state backups are encrypted so that passwords and other sensitive information are not exported as clear text. As a result, full state backups made from Release 1.1(1) or later cannot be restored to a Cisco UCS instance running an earlier software release.

Import Configuration

You can import any configuration file that was exported from Cisco UCS Manager. The file does not need to have been exported from the same Cisco UCS Manager.

Warning Node Has Slots In Importing State Park

The import function is available for all configuration, system configuration, and logical configuration files. You can perform an import while the system is up and running. An import operation modifies information on the management plane only. Some modifications caused by an import operation, such as a change to a vNIC assigned to a server, can cause a server reboot or other operations that disrupt traffic.

You cannot schedule an import operation. You can, however, create an import operation in advance and leave the admin state disabled until you are ready to run the import. Cisco UCS Manager will not run the import operation on the configuration file until you set the admin state to enabled.

You can maintain only one import operation for each location where you saved a configuration backup file.

Import Methods

You can use one of the following methods to import and update a system configuration through Cisco UCS Manager:

• Merge—The information in the imported configuration file is compared with the existing configuration information. If there are conflicts, the import operation overwrites the information on the Cisco UCS instance with the information in the import configuration file.

• Replace—The current configuration information is replaced with the information in the imported configuration file one object at a time.

System Restore

You can restore a system configuration from any full state backup file that was exported from Cisco UCS Manager. The file does not need to have been exported from the Cisco UCS Manager on the system that you are restoring.

The restore function is only available for a full state backup file. You cannot import a full state backup file. You perform a restore through the initial system setup.

You can use the restore function for disaster recovery.

Required User Role for Backup and Import Operations

You must have a user account that includes the admin role to create and run backup and import operations.

Backup Operations

Creating a Backup Operation

Obtain the backup server IP address and authentication credentials.

Running a Backup Operation

Modifying a Backup Operation

You can modify a backup operation to save a file of another backup type to that location or to change the filename and avoid overwriting previous backup files.

Deleting One or More Backup Operations

Import Operations

Creating an Import Operation

You cannot import a Full State configuration file. You can import any of the following configuration files:

• All configuration

• System configuration

• Logical configuration

Collect the following information that you will need to import a configuration file:

• Backup server IP address and authentication credentials

• Fully qualified name of a backup file

Warning Node Has Slots In Importing States

Running an Import Operation

You cannot import a Full State configuration file. You can import any of the following configuration files:

• All configuration

• System configuration

• Logical configuration

Warning Node Has Slots In Importing Statement

Modifying an Import Operation

Deleting One or More Import Operations

Restoring the Configuration for a Fabric Interconnect

Collect the following information that you will need to restore the system configuration:

• Fabric interconnect management port IP address and subnet mask

• Default gateway IP address

• Backup server IP address and authentication credentials

• Fully qualified name of a Full State backup file

  Note	You must have access to a Full State configuration file to perform a system restore. You cannot perform a system restore with any other type of configuration or backup file.