正文文本
DS8000 Power Briefing
Liu Qi (EC LSG)
2
Agenda
DS8300&DS8700 Power Distribution Diagram
DS8800 Power Distribution Diagram
Power Related Tips
erpowerparser tool
FA1802 and FA1803
DS8300&DS8700 A-Rack Power Distribution
3
DS8300&DS8700 A-Rack Power Distribution (continue)
4
DS8300&DS8700 B-Rack Power Distribution
5
DS8300&DS8700 B-Rack Power Distribution (Continue)
6
7
DS8800 A rack Power Distribution
8
DS8800 B Rack Power Distribution
9
DS8800 C Rack Power Distribution
10
DS8870 Power Distribution (Rack A)
11
DS8870 Power Distribution (Rack B)
12
DS8870 Power Distribution (Rack C)
Power difference among DS8000 series
13
| DS8300&DS8700 | DS8800 | DS8870
Power and Cool | Each PPS has 1 * 5V/12V DDM power module + 1 * 5V/12V DDM power module (optional)
Use bus bar to distribute power
2 fan sense card
DASD Fan Assemblies (3 fans) | Remove 5V/12V DDM power module . Thus each PPS has 2 * 208V Power module.
Use bus bar and PDU to distribute power
No fan sense card
No DASD Fan Assemblies | 2 * DC-UPS
Battery integrated in DC-UPS
1 or 2 Batteries in each DC-UPS
Each batteries contains 4 strings
New designed RPC
PLD: 4 seconds
ePLD: 50 seconds
PPS Tips
Removing 5/12V remote sense lines from J3A / J4A from each DDM power module in each frame with new bus bar design. Mandatory during CCL from R1, R2, R3 to R4. (Tip08842)
Replace uncrimped 5V/12V power cables connected between the 5V/12V bus bars and the PPS DDM power modules with old bus bar design. The replacement is disruptive, as machine need to be powered down. (Tip08512)
When replacing a rack PPS sub-component FRU (Sequence module, DDM power module, etc), you MUST use GUI exchange parts menu to perform a pseudo repair of the PPS, in which the PPS sub-component is actually replaced. (Tip08901)
14
BBU Tips
8 month charge date limit of BBU. Installing batteries that are more than 8 months old will hit this. Insert battery strings in different order when doing a Battery enclosure chassis replacement may also hit this. (Tip08482)
Rack Battery replacement when the "Born on date" between battery module sets exceeds 2 years. (Tip08398)
Battery Life is 5 years without battery check / smart charge function (R3) otherwise is estimated up to 7 years. If machine stayed on prior-R3 code, and then upgrade to R3 and above, then its battery life is shorter than 7 years.
Circuit Breakers on BBU can be tripped by debris. Action plan is to reset the circuit breaker and pseudo repair the BBU (Tip08908)
DS8000 do automatic battery test every 30 days since R3. batteryTest.log can be found in../hmca/ var_esshmc_log of pepkg.
15
Number of batteries in a DS8100/DS8300 (tip08028)
16
Number of batteries in a DS8700 (tip08028S)
17
Number of batteries in a DS8800 (tip08028S)
18
CEC/IO Tower PS Tips
Under some conditions, microcode calls out the wrong IO tower power supply to be replaced for SRC 10001511 and 10001521. When you see above 2 SRC called out, replace both tower power supplies by dummy replacing the tower RIO card. (Tip06461)
Repair the CEC or I/O Tower power supply failures before other failures in the CEC or I/O Tower. Some power supplies are not able to maintain full power if the other supply fails, I call it N+ 0.9 syndrome.
19
erpowerparser Tool
erpowerparser is used to interpret power error messages in PE.errpt_detail. All ESC=8xxx entries are power messages.
Usage (in pepkg)
erpowerparser -sgx <sequence number list of the PE.errpt_detail entries>
erpowerparser -m <sequence number list of the PE.errpt_detail entries>
For example, if want to interpret error message
05/08 15:21:51 75424 er_PPpsGenFau cpPPS10 ESC=83A2 Rack_A; PPS reporting batteries differ by more than 2 years.
Then use
erpowerparser –sgx 75424
erpowerparser –m 75424
Note: Power messages (ESC=8xxx) are only reported by LPAR0 when machine in dual cluster mode.
20
erpowerparser Tool (Continue)
Usage (Dial in)
1. From LPAR0 (ESS01), initiate a collect register call. This will put a power event in the AIX error log.
cat /dev/cpss0/er/rpc/collectregs
· 2. Create a detailed error log (ESS01)
cd /usr/lpp/diagnostics/bin/
errpt -a > PE.errpt_detail
· 3. Run Analyze logs form ESS01, then GREP on Power events
errpt -a | /usr/lpp/diagnostics/bin/analyzelogs v - > /tmp/analyze.out
cat /tmp/analyze.out |egrep -i 'esc=8' | egrep -v 'er_TAdditiona';date
4. Now run "Power Parser" against the detailed error log, and use event sequence number (from the analyze logs)
/usr/lpp/diagnostics/bin/erpowerparser -sgx <Rack_A Power System Health Check Event Seq #>
/usr/lpp/diagnostics/bin/erpowerparser -m <Rack_A Power System Health Check Event Seq #>
21
Fill FA 1802 and FA 1803 when open PMH for Power issue
22
FA1802
Fill FA 1802 and FA 1803 when open PMH for Power issue
23
FA1803
