FreshTomato Wiki

User Tools

Site Tools

Trace:
advanced-misc

Differences

This shows you the differences between two versions of the page.

Link to this comparison view

Both sides previous revisionPrevious revision
Next revision		Previous revision
advanced-misc [2026/03/12 01:10] – -Formatting hogwildadvanced-misc [2026/03/12 01:42] (current) – hogwild
Line 5: Line 5:
  \\  \\
  
-**Boot Wait Time** * -  the period during which you can access the router via its CFE webserver.+**Boot Wait Time** *:  the period during which you can access the router via its CFE webserver.
  
 This refers to the bootstrapping portion of the process, or the loading of code before the OS loads. Change this setting if you need to delay the boot process (to stay in internal webserver CFE mode). This refers to the bootstrapping portion of the process, or the loading of code before the OS loads. Change this setting if you need to delay the boot process (to stay in internal webserver CFE mode).
Line 27: Line 27:
  \\  \\
  
-**Enabling CTF disables QoS / Bandwidth Limiter**. It can also cause issues for IPv6 tunneling protocols,** **like 6in4 Static. This is because the switching part of the packet bypasses parts of the standard Linux iptables chains.+**Enabling CTF disables QoS / Bandwidth Limiter**. It can also cause issues for IPv6 tunneling protocols,** **like 6in4 Static. 
 + 
 +This is because the switching part of the packet bypasses parts of the standard Linux iptables chains.
  
 With CTF disabled, the router stores the entire frame before sending it to its destination. With CTF disabled, the router stores the entire frame before sending it to its destination.
Line 48: Line 50:
 On ARM devices, you can expect performance in the range of 200 to 400 Mb/s with CTF enabled. On ARM devices, you can expect performance in the range of 200 to 400 Mb/s with CTF enabled.
  
- \\ + \\
  
-**Enable Jumbo Frames:** lets you increase the maximum frame size on the LAN. (Default: Disabled).+**Enable Jumbo Frames:** lets you increase the maximum frame size on the LAN. (Default: disabled).
  
 The default Jumbo Frames size is 2 KBytes. If enabled, Jumbo Frame capability will be enabled on all LAN/WAN ports. The default Jumbo Frames size is 2 KBytes. If enabled, Jumbo Frame capability will be enabled on all LAN/WAN ports.
  
-This feature is supported only in 1000 Mb/s mode. Also, gigabit switches use more memory when this feature is enabled. This can affect performance, depending on how many switch ports simultaneously use jumbo frames. Not all hardware models support this feature.+This feature is supported only in 1000 Mb/s mode. Also, gigabit switches use more memory with this feature enabled. This can affect performance, depending on how many switch ports simultaneously use jumbo frames. Not all hardware models support this feature.
  
  \\  \\
  
-**Jumbo Frame Size:** If enabled, this value overrides default frame size.+**Jumbo Frame Size:** if enabled, overrides default frame size.
  
-If disabled, default frame size is 2 KB. (Default: off). Generally, enable Jumbo Frames if you frequently transfer large files within your LAN. To function, all devices must support Jumbo Frames. You should also standardise the packet size across all LAN devices.+If disabled, default frame size is 2 KB. (Default: off). Generally, enable Jumbo Frames if you frequently transfer large files within your LAN. To function, all devices must support Jumbo Frames. You should also standardize the packet size across all LAN devices.
  
-It is best to leave this disabled unless you have specific requirements for it. Using Jumbo Frames means Internet traffic will be fragmented, unless the end device supports Path MTU Discovery. Often, the limited performance benefits of Jumbo Frames isn't worth the time and effort needed to tune/troubleshoot it.+It's best to leave this disabled unless you have specific requirements for it. Using Jumbo Frames means Internet traffic will be fragmented, unless the end device supports Path MTU Discovery. Often, the limited performance benefits of Jumbo Frames isn't worth the time and effort needed to tune/troubleshoot it.
  
 Not all hardware models support this feature. Not all hardware models support this feature.
Line 68: Line 70:
 ** **\\ ** **\\
  
-**Port Health:** Since release 2025.5 a "port health"  mechanism has been introduced. This is a watchdog counting errors on  physical connections (Ethernet) and taking a defined action when an  exception is spotted. This function was introduced to prevent a bad  Ethernet cable to take your router down (eating up CPU cycles) and  making the device unresponsive.+**Port Health:** Since r2025.5, this watchdog counts errors on Ethernet connections and takes a defined action when an exception is spotted.
  
- \\ \\ {{port_health.png}}+This function was introduced to prevent a bad  Ethernet cable to take your router down (eating up CPU cycles) and  making the device unresponsive.
  
- \\+ \\ \\ {{port_health.png?591}}
  
- \\+ \\  \\
  
-**Enable** - runs the porthealth process to monitor the defined ports scope, every minute.+**Enable:** runs the porthealth process to monitor the defined ports scope, every minute.
  
-The check is extremely lightweight and executes in about 0.02 ~ 0.04 seconds.+The check is extremely lightweight and executes in about 0.02 - 0.04 seconds.
  
  \\  \\
  
- \\+**Mode: **determines the script’s level of intervention when it detects network errors (a "flood").
  
-**Mode** - determines the script’s level of intervention when it detects network  errors (a "flood"). +Essentially, this sets the "remedy" strategy for a port currently exceeding your defined error threshold.
- +
-Essentially, this sets the "remedy" strategy for a port  that is currently exceeding your defined error threshold.+
  
  \\  \\
  
-  * Monitor (log only) - This is the passive mode. It acts as an early warning system without changing any hardware settings. It compares the current error count against the baseline. If the delta exceeds your threshold, it writes a warning to the system log (logger). Use this to gather data or troubleshoot a "noisy" line without risking a network disconnect. +  * Monitor (log only) - the passive mode. This acts as an early warning system without \\ changing hardware settings. It compares the current error count against the baseline. \\ If the delta exceeds your threshold, it writes a warning to the system log (logger). \\ Use this to gather data or troubleshoot a "noisy" line without risking a network disconnect.
-  * Recover - This is the active/adaptive mode. It attempts to "save" the connection by sacrificing speed for stability. It identifies the failing port and steps the speed down (e.g., from 1000FD to 100FD). It waits for the **Hold Time** to expire.  If errors persist, it steps down again (e.g., to 10HD). If it reaches the lowest setting and still fails, it disables the port entirely..The goal is to maintain a “limping” connection rather than losing connectivity completely. This is ideal for cables that are slightly damaged or picking up interference. +
-  * Disable port -This is the **aggressive** mode. It follows a "zero-tolerance" policy for port errors. As soon as a port exceeds the error threshold (''BASE_RATE''), the script immediately shuts the port down using robocfg. This approach protects the rest of the network or a background compilation process from being affected by a "chatty" or failing hardware interface. It forces manual intervention to re-enable the port.+
  
  \\  \\
 +
 +  * Recover - the active/adaptive mode. It attempts to save the connection by sacrificing \\ speed for stability. It identifies the failing port and steps down the speed (say, from 1000FD \\ to 100FD). It waits for the **Hold Time** to expire. If errors persist, it steps down again to \\ say, 10HD). If it reaches the lowest setting and still fails, it disables the port entirely. \\ The goal is to maintain a “limping” connection rather than lose all connectivity. This is \\ ideal for cables that are slightly damaged or picking up interference.
  
  \\  \\
  
-**Ports** - defines the type of logical ports to be monitored. For example, LAN only, WAN only or LAN + WAN.+  * Disable port - the **aggressive** mode. It follows a "zero-tolerance" policy for port errors. \\ As soon as a port exceeds the error threshold (''BASE_RATE''), the script immediately closes \\ the port using the robocfg tool. This approach protects the rest of the network or a\\ background compilation process from being affected by a "chatty" or failing hardware \\ interface. It forces manual intervention to re-enable the port.
  
  \\  \\
 +
 +**Ports: **defines the type of logical ports to be monitored. For example, LAN only, WAN only or LAN + WAN.
  
  \\  \\
  
-**Max Errors / Minute **- AKA: "threshold". The number of acceptable errors per minute before an action (defined under Mode) is triggered.+**Max Errors / Minute **AKA: "threshold". The number of acceptable errors per minute before an action (defined under Mode) is triggered.
  
-Under normal circumstances you'd expect the number of errors to be "0". However, in certain cases a little tolerance to deal with temporary situations like interference, temperature, etc might be advisable.+Under most circumstances, you'd expect the number of errors to be "0". However, in certain cases, it can be advisable to add a little tolerance to deal with temporary situations like interference, temperature, etcetera.
  
  \\  \\
  
- \\ +**Hold Time: **acts as "cooldown" and "efficiency" filter to prevent the script from making erratic or unnecessary changes to network ports.
- +
-**Hold Time** - acts as "cool-down" and "efficiency" filter to prevent the script from making erratic or unnecessary changes to network ports.+
  
 In this context, it is ideal for the router to have a "calm” decision-making process. When a port shows errors, the script steps down the link speed (say, from 1000FD to 100FD). The Hold Time ensures the system stays at the new speed long enough to see if the errors stop. If the script runs again before the 180 second (default) has passed, it will skip further down-stepping for that specific port, giving the hardware time to stabilize. In this context, it is ideal for the router to have a "calm” decision-making process. When a port shows errors, the script steps down the link speed (say, from 1000FD to 100FD). The Hold Time ensures the system stays at the new speed long enough to see if the errors stop. If the script runs again before the 180 second (default) has passed, it will skip further down-stepping for that specific port, giving the hardware time to stabilize.
Line 118: Line 118:
  \\  \\
  
- \\+**Cache TTL** - defines how often the script performs the "heavy lifting" of scanning router hardware to know which ports are active.
  
-**Cache TTL** - identifies what ports are active is a bigger job than scanning them individually for errors. In this sense the Cache TTL defines how often the script performs the "heavy lifting" of scanning router hardware to know which ports are active. The script saves the list of active ports to a file in "/tmp".  It will reuse that list (only) for 1800 seconds (30 minutes) by default. It only re-scans the hardware if the cache file is older than the TTL or if  the file is missing. This parameter doesn't affect the ultra-light scan for errors on active ports, happening every minute.\\ +Identifying which ports are active involves more than just scanning them individually for errors. The script saves the list of active ports to a file in: "/tmp".  By default, it will reuse that list (only) for 1800 seconds (30 minutes). It only rescans the hardware if the cache file is older than the TTL or if  the file is missing. This parameter doesn't affect the ultra-light scan for errors on active ports, happening every minute.
- +
- \\+
  
  
advanced-misc.1773277850.txt.gz · Last modified: by hogwild

Page Tools

Donate Powered by PHP Valid HTML5 Valid CSS Driven by DokuWiki