There are two standard RipEX units with identical configurations inside RipEX-HS. Both units are booted, however only one is active. Interfaces (COM1, COM2, Ethernet, Radio) of the second unit are disconnected.
Even if Rx on inactive RipEX unit LED panel is blinking when the active unit is transmitting, these packets are not received. i.e. they don’t take part in Statistic, Neighbours or Graphs.
When the active unit HW alarm output changes to “On” (when a controlled value exceeds the respective threshold), the controller automatically switches all interfaces (COM1, COM2, ETH, ANT–if applicable) to the second unit and it takes over all functions. Since both units are using the same MAC addresses (MAC cloning), there is a minimal drop-out while switching, less than 2 s.
When RipEX units inside RipEX-HS are in Bridge mode, the first switch-over takes approx. 30 sec. It is because hot-stand-by RipEX ARP table is clear and has to be filled with the first packets (forward delay). The next switch-over takes less than 2 sec. When Operating mode is Router, even the first switch-over takes less than 2 sec. When ARP proxy in Router mode is used, switch-over time(s) is equal to that of the Bridge mode.
Possible controlled values by HW alarm output (used for switching between “A” and “B” units) are: RSS, DQ, TXLost[%] – lost packets on Radio channel, Ucc – power voltage, Temp – inside temperature, PWR – RF power, VSWR – reflected RF power, ETH[Rx/Tx], COM1[Rx/Tx], COM2[Rx/Tx] – No of received/transmitted packets on respective interfaces. For recommended settings see chapt. Chapter 4, Configuration.
HW alarm output for Temperature is always On. Its thresholds can be Manually set in interval -50 to +90 OC (default -25 to +85 OC). There is also internal control independent on Temperature threshold settings – when temperature of the radio board is higher than 95 OC, HW alarm output is On and is switched to “B” (backup) unit.
SNMP trap with each switch-over can be sent (depending on configuration Settings/Device/Alarm management/HS active) to central SNMP management. When the inactive RipEX HW alarm output is On (controlled Watched value(s) exceeds the respective threshold), the controller never switches to it.
Auto – primary active is RipEX “A”, when it fails, controller automatically switches-over to RipEX “B”. When both units are with alarms, “A” unit remains active.
Auto toggle – the same as Auto mode, in addition after set time controller automatically switches-over to RipEX “B”, even if “A” doesn’t have any alarm and uses “B” for set period in order to confirm, that RipEX “B” is fully ready-to-operate.
A – only RipEX “A” is active and controller will never switch to RipEX “B”
B – only RipEX “B” is active and controller will never switch to RipEX “A”
Auto toggle is the most recommended mode if there are no specific requirements, because unit „B“ functionality is actively and periodically checked.
Auto mode has got minimum number of switch-overs, i.e. minimal possibility of eventual loss of packets in internal RipEX buffers.
A and B modes are supposed to be used only for the maintenance/testing and not for normal service.
Auto toggle mode is HW default, i.e. the RipEX-HS is in this mode after power cycle.
In order to achieve maximum reliability, the controller is software free. Switching-over is based only on HW alarm outputs of RipEX units and the HW signals from their power supplies.
The HW alarm of active RipEX or its power supply will cause the immediate and unconditional switch-over to hot-stand-by RipEX. If there are any packets waiting in queues of RipEX which becomes inactive, these packets are discarded after switch-over.
(=RipEX HW alarm output is “On”) becomes active based on Alarm management settings:
RSS, DQ, TXLost[%] – lost packets on Radio channel
Ucc – power voltage
Temp – inside temperature
PWR – RF power
VSWR – reflected RF power
ETH[Rx/Tx], COM1[Rx/Tx], COM2[Rx/Tx] – No of received/transmitted packets on respective interfaces
For recommended settings see chap. Chapter 4, Configuration.
At least 10 data values have to be averaged before checking for a possible alarm. Since different parametres are sampled over different periods, different times are required to obtain correct values:
Ucc, Temp – approx. 10 sec. after booting
(=power supply of respective unit is “Off”) when:
There is no voltage on the input terminals of power supply and battery voltage is lower than 20.5 V DC, where applicable
There is no voltage on the output terminals of power supply
Some borderline examples for Auto and Auto toggle modes (Primary unit is always “A” in Auto mode, “A” or “B” in Auto toggle mode. The second unit is Stand-by):
Power alarm of Primary unit is “On”
– Stand-by unit becomes active
Primary unit alarm is “On” and Power alarm of Stand-by is “On” or Stand-by unit alarm is “On”
– Nothing will be done. Primary unit remains active
Primary unit alarm is “On”, it was switched to Stand-by unit, Primary unit alarm disappeared
– it will be switched back to Primary unit immediately (the protective timeout of 30 s will be held for shorter HW alarm output changes)
When set (see chap. Section 4.4, “SNMP”), the SNMP trap with the information on the identity of the active unit will be sent by the unit which becomes active (the “Unit name” is inside the trap)
When testing, please wait between each switch-over for at least 30 seconds. A protective timeout of 30 s for the next switch-over will be kept after each change-over.
When Hot Standy mode in RipEX unit is set, there are some limitations:
CD pin on COM1 is used internally, it is not available to the user
HW alarm input is used internally, it is not available to the user
Save and Sleep modes are not available
Some alarms that cause switch-over from Primary to Stand-by unit remain active. The values are not measured after switch-over, because respective interfaces have been disconnected. These alarms can be manually cleared using a Save button in Diagnostic/Neighbours menu or they will be cleared when the log is automatically saved after its Log save period expired. This limitation is valid for following alarms:
SNMP – If both SNMP Alarm and HW alarm outputs are set for the same event the SNMP trap will not be sent out. See chap. Section 4.4, “SNMP”.