Home
>
Lucent Technologies
>
Communications System
>
Lucent Technologies DEFINITY Enterprise Communications Server Release 6 Instructions Manual
Lucent Technologies DEFINITY Enterprise Communications Server Release 6 Instructions Manual
Have a look at the manual Lucent Technologies DEFINITY Enterprise Communications Server Release 6 Instructions Manual online for free. It’s possible to download the document as PDF or print. UserManuals.tech offer 413 Lucent Technologies manuals and user’s guides for free. Share the user manual or guide on Facebook, Twitter or Google+.
DEFINITY Enterprise Communications Server Release 6 Maintenance for R6r Volumes 1 & 2 555-230-126 Issue 2 January 1998 Maintenance Architecture Page 1-15 SPE Duplication 1 nIf the two SPEs have different SOH, the SPE with the better SOH becomes or remains the active SPE. These descriptions apply to SOH levels on the standby SPE. Four possible levels of SPE SOH are supported and maintained by system hardware and software. Standby SPE Maintenance Architecture The maintenance strategy for the standby SPE is based on several independent components. nMaintenance of handshake communication so that software on the active SPE can control maintenance of the standby SPE and its components. nControlling memory shadowing and performing the standby memory refresh operation. nActivities, independent of handshake communication and memory shadowing, used to allow tracking of the standby SPE’s condition. This includes reading of hardware status to determine the actual state of standby SPE. functional The standby SPE is fully healthy with up-to-date memory content identical to active SPE memory content. An interchange into this SPE will cause minimal service disruption. not refreshed The standby SPE’s hardware and operational software are fully healthy but the standby memory content is not currently identical to active SPE memory content. Typically either memory shadowing is off or a memory refresh operation is in progress to bring the memories’ contents into agreement. Interchange into an SPE of this health level will lead to calls dropping and a service outage of several minutes. partially-functional One of the following conditions is in effect: — A failure of a critical standby SPE component has occurred. — The standby SPE has been busied out. — The SPE is in recent interchange mode (see ‘‘ STBY-SPE (Standby SPE Maintenance)’’ in Chapter 9, ‘‘ Maintenance Object Repair Procedures’’). non-functional This is the worst and most seriously disabled state of a standby SPE. The SPE has lost either power or basic sanity; the standby processor and its software are unable to cycle. Such an SPE cannot be made active by an interchange.
DEFINITY Enterprise Communications Server Release 6 Maintenance for R6r Volumes 1 & 2 555-230-126 Issue 2 January 1998 Maintenance Architecture Page 1-16 SPE Duplication 1 As shown in Figure 1-7, all maintenance capabilities for the standby SPE are built upon these three strategies. Figure 1-7. Components of Standby Maintenance Standby SPE maintenance software is designed to attempt to self-correct problems. If a problem occurs, this software automatically tries to address the problem, bring the standby SPE back to a state of availability and clear all alarms which might have been raised. Typically, if a standby SPE problem has not cleared, it is of a hardware nature and some type of hardware component maintenance or replacement action is indicated. Once such corrections have been made, the system software will automatically bring the standby SPE back to full availability. There is no management terminal command to stimulate refresh of standby SPE memory; system software automatically accomplishes this itself when conditions are appropriate. The same is true of efforts to turn on shadowing where no explicit user interface command to turn on/off shadowing is available (note that busyout/release, below, can be used to indirectly accomplish this). Standby Maintenance Monitor Software The Standby Maintenance Monitor (SMM) is a software package that is always running on key components of the standby SPE to verify its competence. SMM tests individual standby SPE components and reports back to the active SPE, by the handshake message, any failures of individual tests. Failure reports trigger enhanced maintenance attention to standby SPE component problems by active Standby SPE Hardware Status ReadingMaintain Handshake Communication Maintain Standby SOH, Shadowing, Lock Status Memory Shadowing Memory Refresh G3-MT Access to Standby SPE ComponentsError/Alarm Logging for Standby ComponentsHandshake Comm. UpTime of day clocks in synch Handshake Comm.Down Stby SPE down/lock G3-MT Access to SPE-Down Interface
DEFINITY Enterprise Communications Server Release 6 Maintenance for R6r Volumes 1 & 2 555-230-126 Issue 2 January 1998 Maintenance Architecture Page 1-17 SPE Duplication 1 SPE software. SMM also ensures that when handshake communication has been down for an extended period, the standby SPE will transition into the SPE-down state. Handshake Communication Every 30 seconds, the active SPE sends a handshake request message to SMM and waits for SMM to respond with a handshake response message. This message transmission occurs across the Duplication Interface circuit packs and their interconnecting cable. As long as SMM responds to these regular handshake request messages, handshake communication is considered up as reported on the status spe screen. The physical path of handshake communication is illustrated below. Hardware problems at any point in this route could interfere with handshake communication. If the standby SPE fails to respond to four successive handshake requests, handshake communication is considered down. A major alarm is logged against STBY-SPE with error type 1 logged. The status spe screen will indicate that handshake is down. It is then no longer possible to communicate with the standby SPE. Maintenance testing of the standby by the active SPE (or by command) is discontinued, and the error and alarm logs become outdated for standby components. Handshake communication failure is a severe and rare condition. It is due to either a failure of Duplication Interface hardware or a catastrophic failure of the standby SPE. As long as the active SPE is not locked by the switches, software attempts every 30 seconds, to re-establish handshake communication. When the SPEs are locked with the switches, handshake communication is physically impossible, but no alarm is raised. When the standby is busied out, handshake communication should remain up, but in any case, only the busyout WARNING alarm will be raised. Whenever the active SPE has undergone a restart (levels 1-5), handshake is technically considered down during and just after the restart. After a level 1 (hot) restart, if there are no standby SPE problems, handshake communication should be restored within 30 seconds. After active-SPE restarts of levels 2 and up, handshake should be restored within 3 minutes of G3-MT re-enabling. The active SPE keeps hardware configuration and vintage data about the components of the standby SPE. This data can be accessed with list configuration control. Whenever handshake is down, this data may be out of date. Whenever handshake has been down and is restored, the active SPE requests standby SPE software to transmit the current version of this data. The data is then stored in active SPE memory. Failure to use the lock-and-power- down method for standby circuit pack replacement can lead to incorrect standby component hardware configuration and vintage data.
DEFINITY Enterprise Communications Server Release 6 Maintenance for R6r Volumes 1 & 2 555-230-126 Issue 2 January 1998 Maintenance Architecture Page 1-18 SPE Duplication 1 Figure 1-8. Handshake Communication Path Maintenance of Standby Components When handshake communication is up, maintenance for individual components of the standby SPE is the same as that for the active (except in some details for PKT-INT). The same commands are used to test standby and active circuit packs, and the error and alarm logs maintained on the active side record data for both. If a major on-board alarm is raised against a standby SYSAM, Processor, Memory, MSSNET, or Packet Interface board, the standby SPE’s SOH is lowered to partially-functional. Once that board’s problem is fixed and the alarm cleared, system software automatically raises the standby SPE’s SOH to not-refreshed or functional, depending on whether its memory is up to date. Standby component faults can also affect memory shadowing. Certain faults can have negative effects on system operation if memory shadowing is left on. When these components get major alarms, memory shadowing is automatically kept off by system software. These are referred to as shadowing relevant components. Roughly, these include the hardware that provide shadowing or the hardware into which shadowed writes occur. Table 1-2 below shows the effect often major on-board alarms against standby components on standby SOH and on memory shadowing. Note that off-board . . . . . . . . . . . . . . . . . . . . . .P R O C RSMM software SMM: Standby Maintenance Monitor Request Response MAP: Maintenance Action Process Duplication Interface Cable D U P I N T Standby SPE Active SPE DUP Driver P R O C R MAP softwareD U P I N T
DEFINITY Enterprise Communications Server Release 6 Maintenance for R6r Volumes 1 & 2 555-230-126 Issue 2 January 1998 Maintenance Architecture Page 1-19 SPE Duplication 1 alarms, minor alarms and warning alarms have no effect on memory shadowing or on the SOH of the standby SPE. When handshake communication is down, but the standby SPE is not in SPE-down mode (SOH is not nonfunctional), autonomous testing of standby SPE components by the SMM occurs on the standby SPE. If a component fails a test while handshake is down, its red LED is lit and the standby SOH is lowered to partially-functional. A standby SPE component is considered to be testable if it can be tested with the usual maintenance commands from a management terminal connected to an ACTIVE connector on the SPE. In this condition, full maintenance software for it is running in the active SPE and the error/alarm data for it is up to date. Table 1-3 gives testability requirements for the various SPE components. Table 1-2. Effects of Major Alarms on Shadowing and Standby SOH Alarmed Component SOH Effect Shadowing Effect PROCR partially functional no effect MEM-BRD partially functional shadowing kept off SW-CTL partially functional shadowing kept off SYSAM partially functional no effect PKT-INT partially functional shadowing kept off DUPINT no effect shadowing kept off DUP-CHL no effect shadowing kept off HOST-ADAPTER no effect no effect DISK no effect no effect TAPE no effect no effect
DEFINITY Enterprise Communications Server Release 6 Maintenance for R6r Volumes 1 & 2 555-230-126 Issue 2 January 1998 Maintenance Architecture Page 1-20 SPE Duplication 1 Locking the Active SPE Duplication Interface hardware supports the ability to lock the active SPE in active mode by means of the SPE-Select switches. The procedure for safely doing this is described in Chapter 5, ‘‘ Responding to Alarms and Errors’’, and in ‘‘ STBY-SPE (Standby SPE Maintenance)’’ in Chapter 9, ‘‘Maintenance Object Repair Procedures’’. In locked mode, the system operates as if it is simplex: nThe standby SPE is inaccessible to the active SPE and active G3-MT login. nNo SPE-interchange is possible. nHandshake is down and memory shadowing is off. The locked state is intended for temporary use to prevent interchanges during maintenance sessions. No alarm is raised when the switches are locked. However, alarms against SPE-SELE are raised later if the switches are left out of the AUTO position for an extended length of time. Memory Shadowing Memory shadowing is used to keep the standby SPE’s memory content up-to-date relative to the active SPE’s memory. Memory shadowing is turned on automatically when the standby SPE has booted up and completed its own memory testing. Each write operation in active memory is replicated in the corresponding location in standby memory. Table 1-3. Testability Requirements for Standby Components Component Required Condition PROCR handshake up MEM-BD handshake up SW-CTL handshake up SYSAM handshake up PKT-INT handshake up and Stby Refreshed DUPINT handshake up DUP-CHL handshake up HOST-ADAPTER handshake up DISK handshake up and Stby Refreshed TAPE handshake up and Stby Refreshed
DEFINITY Enterprise Communications Server Release 6 Maintenance for R6r Volumes 1 & 2 555-230-126 Issue 2 January 1998 Maintenance Architecture Page 1-21 SPE Duplication 1 When shadowing into the standby SPE has been off (as when the system first comes up), system software checks to see if it is safe to restore shadowing. Handshake communication must be up. Then software verifies (with Test #920) that the SPEs have identical hardware configurations. If this passes and there are no shadow-relevant component failures, system software turns on shadowing again. Once shadowing is turned on, it is necessary to refresh the contents of standby memory to bring it into full agreement with the active’s by copying every word of active SPE memory to the standby. This takes approximately 5 minutes, though traffic load can increase the duration. When completed, the standby SPE is said to be refreshed. status spe or the Standby SPE Status Query Test (#855) in the STBY-SPE test sequence can be used to check the REFRESH status of the standby. Unless the standby SPE is refreshed, interchange into it can disrupt service for several minutes. Otherwise, interchanges are minimally disruptive. A standby SPE exiting lock mode or just released from busyout must undergo this full re-initialization. System software tracks the operation and raises a major alarm when refresh failure occurs. If shadowing stays on, system software automatically tries to refresh again 5 minutes later. Generally, memory shadowing should always remain on. But there are conditions when the system legitimately operates with shadowing off: nThe standby SPE is undergoing any restart. nThe active SPE is undergoing a restart level of 2 or greater. nThe active SPE is locked. nThe standby SPE is busied out. In any other situation, it is an error condition for shadowing to be off. The first 2 situations are transitory and shadowing should automatically be restored within 10 minutes. If shadowing has been on for several minutes, it is an error condition for the standby not to be refreshed. Initialization: Bringing the Standby SPE Up When the standby SPE has been out of service or is first coming up, SPE software executes the following steps: 1. Establishes handshake communication. 2. When SMM answers handshake, raises the standby SPE’s SOH to not refreshed if it has no critical component alarms, or partially functional there are critical component alarms. 3. Tests for component mismatch (test number 920). 4. If there is no mismatch, and no major alarms against shadow-relevant components, and if SMM permits, turns on memory shadowing
DEFINITY Enterprise Communications Server Release 6 Maintenance for R6r Volumes 1 & 2 555-230-126 Issue 2 January 1998 Maintenance Architecture Page 1-22 Power Interruptions 1 5. If memory shadowing is successfully turned on, initiates the process of overall memory refresh 6. When refresh completes, if there are no critical component major alarms, raises the standby SPE’s SOH to level functional Standby SPE initialization is a lower priority than initializing the active SPE and is therefore “paced” to lower CPU consumption. The above steps are carried out at 10 second intervals. During system initialization, the above sequence begins about 2 minutes after the terminal login prompt becomes available. Normally, the standby SPE should be fully initialized about 5 minutes after the availability of the login prompt. You can follow the execution of this sequence by repeatedly entering the command “status spe.” Should a step of this initialization sequence fail, system software retries that step at 30 second intervals until it succeeds. It does not proceed to the next step until the current one has succeeded. The failed condition is alarmed. A procedure for bringing up the standby SPE after being in the SPE-down or locked modes is described at the end of Chapter 4, ‘‘ Initialization and Recovery’’. Power Interruptions System cabinets and their associated power supplies can be powered by 110/208 volts AC either directly or from an Uninterruptible Power Supply (UPS) system. Alternatively, the cabinets and their power supplies may be powered by a -48 VDC battery power plant, which requires DC-to-DC conversion power units in the system. If power is interrupted to a DC-powered cabinet or an AC-powered cabinet without optional backup batteries, the effect depends upon the decay time of the power distribution unit. If the interruption period is shorter than the decay time, there is no effect on service, though some -48V circuits may experience some impact. If the decay time is exceeded for a PPN, all service is dropped, emergency transfer is invoked and the system must reboot when power is restored. If the decay time is exceeded for an EPN, all service to that Port network is dropped and the EPN must be reset when power is restored. If the EPN contains a Switch Node carrier, all service to Port Networks connected to that Switch Node is dropped. Single-carrier cabinets, which can be used for EPNs, also have no battery backup. If power is interrupted for more than 0.25 seconds, all service is dropped, and emergency transfer is invoked for the EPN. In the above cases, the cabinet losing power is unable to log any alarms. However, in the case of an EPN going down while the PPN remains up, alarms associated with the EPN will be reported by the system.
DEFINITY Enterprise Communications Server Release 6 Maintenance for R6r Volumes 1 & 2 555-230-126 Issue 2 January 1998 Maintenance Architecture Page 1-23 Power Interruptions 1 Nominal Power Holdover AC-powered multicarrier cabinets are equipped with an internal battery, powered by its own charger, that provides a short term holdover to protect the system against brief power interruptions. This feature, known as the Nominal Power Holdover, is optional on cabinets supplied by a UPS and required on all other AC-powered cabinets. The battery is controlled in such a manner that it automatically provides power to the cabinet if the AC service fails. The duration of the holdover varies according to the type of carrier and whether or not the system has a duplicated SPE. See Table 1-4 for duration times: Effects of Power Interruptions Power holdover is controlled by software in the above manner in order to allow the system to sustain multiple brief power interruptions without exhausting the batteries before they have time to recharge. After power is restored, the batteries are recharged by a circuit that monitors current and time. If the batteries take more than 30 hours to recharge, a minor alarm is raised, indicating that the batteries must be replaced or the charger replaced. The 397 Battery Charger Circuit immediately detects loss of AC power and raises a warning alarm against AC-POWER that is not reported to INADS. Certain maintenance objects such as external DS1 timing will report major alarms in this situation. When power is restored, the AC-POWER alarm is resolved. PPN Cabinet with Power Holdover When power is interrupted to a PPN cabinet, the effects depend upon the duration of the outage. Battery power is supplied to the whole cabinet for 10 seconds. If power is restored during that period, service is not affected. If the interruption exceeds the cabinet holdover period, but is restored before the control carrier holdover expires, all service is dropped and emergency transfer is invoked. The SPE is kept up allowing for a speedy restoration of service since a reboot is not required. All non-SPE circuit packs must be reinserted, taking about a minute, depending on the size of the system. If the interruption exceeds the control carrier holdover, all service is dropped and the system must reboot when Table 1-4. Nominal Power Holdover Cabinet Type Control Carrier Entire Cabinet PPN, duplicated SPE 5 minutes 10 seconds PPN, simplex SPE 10 minutes 10 seconds EPN 10 minutes 15 seconds
DEFINITY Enterprise Communications Server Release 6 Maintenance for R6r Volumes 1 & 2 555-230-126 Issue 2 January 1998 Maintenance Architecture Page 1-24 Protocols 1 power is restored, taking up to 15 minutes, depending on the size of the system. Human intervention may be required if central office equipment has been busied out. EPN Cabinet with Power Holdover When power is interrupted to an EPN MCC for less than 15 seconds, no service effect results. If the interruption exceeds 15 seconds, only the control carrier is kept up. Circuit packs on other carriers are powered down. Only calls and other services maintained by circuit packs on the control carrier are maintained. For this reason, critical services and those that require a long time to restore (for example, Announcement circuit packs) should be located on control carriers. All service to Port Networks connected to a Switch Node in the EPN is lost. When power is restored, all affected EPNs are reset by system software (see ‘‘ EXP-PN (Expansion Port Network)’’ in Chapter 9, ‘‘Maintenance Object Repair Procedures’’). As with the PPN, a warning alarm is raised against AC-POWER. External Alarm Leads Each cabinet provides two leads for one major and one minor alarm contact closure that can be connected to external equipment. These are located on the SYSAM and Maintenance circuit packs. If the switch is under warranty or a maintenance agreement, EXT-DEV alarms are generated by the equipment connected to these leads and reported to INADS. These may be used to report failures of UPSs or battery reserves powering the switch. They are also commonly used to monitor adjuncts such as AUDIX. Protocols This section describes the protocols handled by the system and the points where these protocols change. Figure 1-9 is a pictorial guide through data- transmission state changes. Figure 1-9 illustrates the flow of data from DTE equipment, like a terminal or host, through DCE equipment, like a modem or data module, into a communications port on the system. The data flow is shown by solid lines. Below these lines are the protocols used at particular points in the data stream. Not shown in the Figure 1-9 is the treatment of D-channels in ISDN-PRI and ISDN-BRI transmissions. PRI and BRI D-channels transport information elements that contain call-signaling and caller information. These elements conform to ISDN level-3 protocol. In the case of BRI, the elements are created by the terminal or data module; for the PRI, the elements are created by the system, which inserts them into the D-channel at the DS1 port. For ISDN transmissions, therefore, BRI terminals and data modules, and DS1 ports insert, interpret, and strip both layer-2 DCE information and layer-3 elements. Also, the DS1 port passes layer-3 elements to the system for processing.