SERVERS
Solution that provides consolidated
consistent way to monitor, configure, update and manage servers = OMSA Open
Manage Server Assistant
SAS = NOT Leverages SCSI command
set, supports multiple drives on channel, shares throughput bandwidth
View logs related to DRAM failure =
DSET, OpenManage Server Administrator
Drives that are complete and
consistent = Controller guide, Header guide, Time-stamp, Sequence number
Items detected and communicated
using DRAC = Temp, Voltage, Fan Speeds
Hot spare = used to replace failed
drive in redundant array
Clear non-BIOS errors from LCD = clear
SEL
LCD display during POST = NOT BIOS
version, date/time
Components in RAID that can contain
firmware = PERC, backplane, HD
32 bit diagnostics CD can create =
USB, Disk, HDD image
Valid color coded message displayed
on LCD = alert state – amber
Perform global re-scan of storage
drives = OpenManage Server Administrator
Perform HD self tests while server
running = NOT OpenManage, 32 bit diag
Dell diagnostics operating within
the OS = NOT DSET, 32 bit, MpMemory tests
If RAID drive goes offline, 1st
step = Attempt to rebuild the drive
Available in PERC BIOS = NOT modify
auto-rebuilding, modify background operation rates
Import and clear foreign disk
configs = PERC using ctrl R, OpenManage Server Admin
First digit of PE 2900 LCD message
code = E = Error
Virtual media support = remotely
boot managed systems from virtual media, install apps from vm, update drivers
from vm, install os from vm
Running 32 bit diag be helpful = computer
can’t access hard drive
OMSA PERC battery learning cycle
delayed = NOT 72 hours
Check firmware compatible with
drivers in use = OMSA
Test can be run from 32 bit = MpMemory
PE 2900 LCD Messages = written to
SEL, listed during POST, displayed with color coded background
Displays on LCD if multiple errors =
NOT all messages in sequence; only the most recent message
DRAC supports use of virtual media
Test run from 32 bit = MpMemory
Hot-plug drive replace =
automatically rebuild
Modify auto rebuild rate = OMSA
BMC’s available = NOT host in-band,
LAN in-band, LAN out-of-band
Management station perform when
using virtual media = provides physical media
Run 32 bit = f10
Watch diagnostics run = NOT test
selected tab
OMSA = solution for PE systems that
provides consolidated and consistent way to monitor, config, update, manage
servers
Clear NVRAM = power off server and
change NVRAM jumper
Baseboard Management Controller = BMC = proactively
forecasting the possible failure of hardware
-
Failovers: monitored and managed by PSU
controller and detect a failure and utilize backup. Controller will perform load balancing to
reduce the load on any single PSU which extends mean-time-to-failure (MTTF)
-
Fans are controlled by BIOS and hot swappable
-
Memory redundancy:
o
Spare bank
o
Mirroring
o
RAID
Types: Tower, Rack, SC, Blade
-
Tower = 6800 + 2900
-
Rack = 1-2U in size. 1U = 1.75 inches
-
SC = Cheaper, 2 CPU’s max, 1 PSU
-
BLADE = “server on a motherboard”; large server
farm
PE 2950 = Model 2, Gen 9, Rack optimized (5 for rack, 0 for
tower), server (0 for server, 5 for blade)
Racks and rails work together in server environment = Rack
systems ensure proper closure of servers, rail systems provide easy access
PERC = Poweredge Expandable RAID Controllers; Dell
proprietary RAID controller; E = EXPANDABLE and
not Edge
Components:
-
Side-plane connector = connecting SAS controller
to board
-
DRAC-5 = Dell Remote Access Card 5 = remote
manage; dedicated Ethernet port
-
PSPB/PDB = Power Supply Paralleling Board or Distribution
Board
PCI
-
PCI = 133mbs
-
PCIX = 1gbs; extended, parallel based
-
PCIE = +2gbs; express, serial based
-
Hot Plug PCI: not supported for Dell RAID and
SCSI controllers
o
Off - off
o
Green – On
o
Green blinking fast – slot being identified by
app, program, or driver
o
Amber blinking slow – card is faulty or
improperly installed
o
Amber blink twice, pause, then repeat – hot
plugged expansion card is slower than other cards on same bus
CPU’s
-
AMD Opt 8200, 2200
-
Intel Xeon 7100,5300,5100,5000,3000,Pentium
d915, Celeron d336
Power
-
LED:
o
Red or amber = failure
o
Green = ok
-
PSPB:
o
Distributes load across multiple PSUs
o
Larger systems with n+1
-
PDB:
o
Hot plug logic; power distribution
o
Smaller systems with 1+1
o
No load balancing
o
Provides redundancy and hot swapping
BIOS
-
BIOS Sequence:
o
Power supply board turns on and initializes;
chipset receives the signal from power supply board; BIOS POSTs; BIOS looks for
Video and runs it; memory count up; system inventory; summary screen displayed;
search for boot drive; start the process of booting OS
-
Factory default:
o
Alt + f = factory default for entire BIOS
o
Alt + d = single fields in BIOS
o
SCSI enabled will be default
Drives that come with 9th gen servers = SAS and
SATA
Networking
NIC Teaming = create a VLAN, load balancing, failover
-
Fault tolerance = Smart Load Balancing (SLB);
Adaptive Fault Tolerance (AFT)
o
If one fails, the other will be used
-
Load Balancing = SLB; Adaptive Load Balancing
(ALB)
o
Share large data loads preventing one from being
overwhelmed
-
Generic Trunking = FEC and GEC/802.3ad; FEC and
GEC/802.3ad
o
Link aggregation. Multiple cables/ports in parallel to increase
link speed beyond the limits of any single cable/port
-
Teamed using Adapter Config Utilities
o
Broadcom = BASP; Intel = PROSet
o
Broadcom utility is preferred whenever possible
TOE
-
Removes processor bottlenecks; reduce excessive
traffic across internal memory; traffic on internal PCI; Key required
-
Does not support iSCSI.
-
NIC is online before POST. Must install the TOE key before AC power is
applied
-
Layer 4 ethernet
-
BACS2 = Broadcom Advanced Control Suite v2
o
Basic diagnostics on Broadcom
o
Visibility in network stats
o
Teaming wizard
o
Netstat –nt to check for offloaded connections
Management
In-band = agents going through the OS
Out-of-band = through the DRAC and independent of the OS
-
Check for line voltage, verify BIOS, view logs,
power on/off, force boot from floppy
-
Once DRAC is plugged in, direct interface with
BMC is disabled and can’t use IMPI to talk to BMC
To talk to BMC, use OMSA or RAC.
DRAC 5 is hardware and software that talks to BMC and takes
over once installed on x9xx servers
-
Additional security, ipmi/bmc config
capabilities, enable.disable access points and features, SM-CLP (new CLI for
RAC), RAC users are same domain as BMC users, Virtual Media is USB not IDE, hardware chipset and faster
processor
-
RACADM = CLI for DRAC 5
o
Both hex and decimal numeric values, displays
numeric values as decimal, IPMI config groups
-
DRAC 5 GUI
o
Consistency with RACADM and manageability
features; target based rather than functional based navigation; consistency
with OMSA gui; enhance server sensor display; ssh, serial, telnet;
-
DRAC 5 Configuration
o
Web based interface; RACADM CLI
“cfgLanNetworking”; BIOS BMC Binary (BBB) (Ctrl E setup)
o
Network settings configured by web; RACADM CLI;
BIOS BMC Binary (BBB)
-
Virtual Media = USB, floppy, CD, DVD
-
Can’t configure a console redirection session on
the local system
System Event Logs (SEL)
-
Ways to view logs:
o
Baseboard Management Controller (BMC)
§
AKA Remote Access Config Utility
§
Ctrl E during POST to view logs
o
Dell System E-Support Tool (DSET)
§
Does not come standard
§
Extracts all event logs for offline viewing
(1MB)
o
Dell OpenManage Server Administrator
§
Nice pretty GUI that you can drill down and do
lots of things
-
RAC = Remote Access Control
-
DRAC 5 = Dell Remote Access Control
-
SEL = System Event Logs
OpenManage
-
Deployment tools = DSA (Dell Server Assist), DTK
(Deployment Tool Kit)
-
Monitoring Tools = OMSA (Open manage server
admin), ITA (IT Assitant), RAC, BMC (Baseboard Management Controller Management
Utility)
o
OMSA, RAC, and BMC = management solution that
provides a consolidated way to monitor, configure, update and manage servers
o
OMSA = does
not support SC servers!
-
Maintenance Tools = SUU (Server Update Utility),
PED (PowerEdge Diagnostics)
o
SUU = update BIOS, RAC, RAID firmware
-
IPMI = “discover” capabilities of manageable
devices
-
DRAC has virtual floppy, CD support and physical
drive or image support
-
Remote Management Hardware = DRAC; works out of
band, no OS interaction, can work when system is powered down
Storage
SATA
-
Point to point, no master/slave
SCSI
-
Timing errors at higher bus speeds; EMI (Electro
Magnetic Interference); Serial bytes converted to parallal data bits, parallel
move down bus system, re-serialized at receiving end
SAS
-
Uses proven SCSI functionality
-
SCSI command delivered via a serial topology;
16,387 SAS devices in SAS domain; enhanced SATA while adding support for a
second drive port; point to point, switched architecture
-
Point to point, full duplex, dual port, hot plug
-
PHY
o
Like a telephone conversation, source and
destination PHY; OPEN -> OPEN_ACCEPT… CLOSE
o
Two narrow links can be combined to become a
wide link where each PHY has a unique identifier but share a common SAS address
-
Expanders = port multiplier; similar to an
Ethernet switch in that it’s not an end point
-
Connection rate = +3gps; rate matching if
connection rate is slower than physical link rate
-
Protocols= SSP (Serial SCSI Protocol) supports
SCSI disk, tape, etc, STP (Serial ATA Tunneling Protocol) supports SATA drives, SMP (Serial Management
Protocol) supports SAS expanders
-
SAS domain = SAS devices + 1 or more expander
-
SAS controller online = goes out through each of
its phys, Registers every device it comes across, Notes the device WWN,
maintains the list of physical disks by WWN
-
If RAID supported, examine RAID info and build
map of virtual disks to physical disks
-
Backplanes, slots, enclosures obtained by
discovered SEPS
-
Basic SAS controller = SAS 5i = daughter card, I
denotes internal cables only, ctrl C BIOS setup
-
Added RAID hardware = PERC5i = daughter card,
ctrl R BIOS setup
PERC 5 =
PowerEdge Expandable RAID Controller = DDF is stored twice at the end of each
physical disk
-
Uses 512mb on each disk for DDF
-
Config data cached during normal operation and
NOT on the controller. Removal of member
disks = virtual disk deleted
-
Provides logic that interacts with CPU and
memory or CPU and storage devices to
read/write
RAID
-
Stripe order info stored on PERC and each disk,
adapters know where disks are
-
Move a disk or replace the PERC = mismatch
-
RAID types = 0, 1, 5, 10, 50
PERC = PowerEdge
Expandable RAID Controller
-
PERC 5/i
5 = generation, I = integrated vs controller card
-
Hot spare =
RAID configured with drive already in place; drive roaming = identify
drives that have been moved and re-associates the drives SCSI ID with the new
slot
-
BIOS does NOT control initial part of boot process
BIOS checks and configures memory in post process
AC Adapter = power cord + brick
Symptoms of no post = no OS messages, no response from
keyboard, no startup sounds, locks up at Dell splash screen
Successful post = keyboard status LED’s flash, OS messages
AC adapter LED changes from solid
green to blinking = board
NVRAM = ESCD
HD’s that are replacement will
contain diagnostic partition = FALSE
BIOS copied from ROM to RAM during boot = SHADOW MEMORY
POST related fault = Check state of
power LED, check diagnostic LEDs, Check AC adapter LED, Refer to online doc
Power up but no video on LCD = check
num lock, connect external display, wait for BIST to load
Ink spots on LCD is not covered by
support
Wireless unable to connect = NOT confirm card is on, confirm
enabled in BIOS, run PSA
LCD distorted = Dell diagnostics and
OBD, Update driver
BIOS = Basic Input/Output System
Chassis: Latitude D, E, Precision, Vostro
Portability, Design, Materials, Tech, Access to hardware
FRU = Field Replaceable Units
-
System board, Video, Audio, NIC, touchpad, LCD +
internal memory
CRU = Customer Replaceable Units
-
Keyboard, HD, memory modules, mini-OC cards,
Bluetooth cards, modem daughter card, coin-cell battery, main battery, modular
drives, external dries, AC Power
Components:
-
Front View
o
Ambient light sensor, keyboard LEDs, Status
LEDS, internal microphone, touchpad
-
Bottom View
o
HD, Memory, Battery, Dock connector
Can be removed:
-
AC Power, battery, coin cell battery, memory,
keyboard, Bluetooth daughter, expansion card, HD, Modular bays
Keyboard Bexel:
-
Diagnostic lights
-
“lock lights”
Chassis: XT,XT2
Tablet, integrated pen support, multiple gesture touch
screen, digital ink, handwriting recognition
Screen swivels 360
Size/weight, dual mode (pen and touch), structural
integrity, usability, performance
On the LCD:
-
Power, Diagnostic LED, Battery LED, Status
indicators
-
Wireless switch = FN F2
CRU
-
AC Power, Battery, Bluetooth, pen, express
cards, wireless antenna
FRU
-
Ambient light sensor, audio card, bio reader,
LCD, processor, board, touch pad, video
BIOS
SSU = System Setup Utility – F2 OR F12 -> system setup
Detected,
enabled/disabled, settings for: power, processor performance, video, etc
Stored in PROM on CMOS chip.
PROM BIOS
Code that allows a computer to boot
BIOS Security Policies
-
BIOS Password
BIOS Level Security
-
Passwords set by user: primary, admin, hard
drive
-
Primary = prevents booting any media
-
Admin = boot the OS but prevents BIOS changes
-
HD = HD specific password that will prompt on
any machine
POST process:
Power button -> signal through board to processor ->
cpu signals soft switched power supply to power up system -> power supply
sends current out
Initialize core chipsets, checks and configs memory, enables
L1 L2 cache and check cpu, copy into memory
Latitude LED
error codes for no post
Flash-off-off = no sodimms present (memory)
Flash-on-on = cpu error
Flash-on-flash = LCD
Off-flash-off = memory compatibility error
On-flash-on = memory detected but has errors
Off-flash-flash = modem
Flash-flash-flash = board
Flash-flash-off = option rom
Off-on-off = HD/storage
2 comments:
Thanks! This really helped!
Worked really well for me, too. Thanks loads!
Post a Comment