ITEM: HB1883L
sysback:restoring from aix 4.2.1 server to 4.1.5 client
Question:
4.2.1
9076
sysback:restoring from aix 4.2.1 server to 4.1.5 client
..hangs
Response:
Customer Contact
Response:
jsh was still in the building - he will call customer
Response:
Calling customer
Response:
ENV: CWS = 4.2.1 (mca/up) node = 4.1.5 (high)
sysback 3.3.3.17
*CUSTOMER REP: Kevin McNamara
PROBLEM: Changed pp size
Restoring 4.1.5 image from CWS and then will migrate
restore failed during getrootfs after install
*ACTION TAKEN: Looks like version conflict without warning.
Will have cust boot from a 4.1.5 machine then restore
from the local cws images.
Still on call
*ACTION PLAN: If this fails, will type "DEBUG" at main screen and
analyze debug output in /tmp/debug.out
If this fails due to space, will increase ramfs in
/usr/sbin/mkboottape.4 from 16384 to 18432 and will
remake boot image.
Continuing call
Response:
*CUSTOMER REP: Kevin_McNamara@css.mot.com
*ACTION TAKEN: booting from other 4.1.5 node yields 888 crash:
888-102-700-0c8
Checked a bunch of items and spoke with JAD and BC
looks like the LED before the 888 indicates the problems
Cust looked up sequence:
888 = crash
102 = system halt
700 = software interrupt
0c8 = dump device not configured
Had cust reboot from 4.1.5 machine again and watch LED
bootp -> device config -> c31 -> please define the sys-
tem console... press "1" -> LED c33 -> "AIX SYSTE "
HANG 888-102-700-0c8
Trying local CWS again.
fail ? tty HW failure : tty software prob on 4.1.5 ;
tty HW ? call 1800 IBM serv : tty soft ? install sio : ;
*ACTION PLAN: Continuing...
Response:
*CUSTOMER REP: Kevin McNamara
*ACTION TAKEN: Booted fine off of 4.2.1 machine.
Sent /u/lmiller/standard.io mca device list
JAD recommends AP; I declined her suggestion.
*ACTION PLAN: Cust is installing sio filesets on 4.1.5 server node
cust will reboot 4.1.5 server node
Cust will rebuild boot image
cust will reboot client 4.1.5 node from server node
Response:
*CUSTOMER REP: Kevin
*ACTION TAKEN: Many of these i/o filesets are missing
Cust is downloading from fixdist
devices.mcs.dee6 and f6fe aren't on fixdist
Cust will check install media...
*ACTION PLAN: Continuing
Response:
*CUSTOMER REP: Kevin
*ACTION TAKEN: Customer disengaging
*ACTION PLAN: WOCB in the AM
Cust will pageout if necessary
If problems, will contact SP and AP
Response:
cust req the next avail. ts.
Response:
*CUSTOMER REP: Kevin McNamara
PROBLEM: Having problem booting up one of the node from the
workstation. It booted passed the C31 and C33. Hanged
on 888...
*ACTION TAKEN: Review item... customer stated that he believe
that the node need to have the tty to be able to
handle something? other than just terminal activity.
.. pssp require..
Conference in Bob D./SP for assistance.
*ACTION PLAN: Mark M./SP will assistance the customer.
TEST CASE: None.
Response:
Customer Rep: Kevin McNamara
Action Taken: Conferenced in AP and SP. SP group will
work with customer.
Action Plan: Waiting on FUP from SP.
Response:
*CUSTOMER REP:
Kevin
PROBLEM:
Sysback restore fails
*ACTION TAKEN:
Customer is using a 4.1 node as sysback boot install server.
Node is crashing when sysback is started
Customer feels he is missing device drivers
Conferenced in Mark Moody (B/E)
We can run spmon -open node\# on the node we made the sysback image Ok.
This shows the B/I server has all the required software for a tty .
Customer found tag in sysback manual about using newer versions
of sysback for restore. Customer appears to be doing everything correctly.
System hangs when /usr is mounted from ram.
Need to talk to Josh Davis in sysback. Josh will be in at 9AM.
WE feel this is a aix or sysback problem. The tty is opening OK.
Customer to go off-line, we\\ll conference with Josh when he arrives, then
call customer back
*ACTION PLAN:
Wait for Josh/Lee for conference.
TEST CASE:
N/A
Response:
Customer called back. JR is researching. Will call customer back
at 1-800-SKY-PAGE pin 5347287.
Response:
Response:
Xed Kevin to the hold que for nxt avail.
Response:
transferred customer to John's voice mail
Response:
SORRY- transferred to John, live NOT voice mail
Response:
*CUSTOMER REP: N/A
*ACTION TAKEN:
Spoke with John Rigler.
Turns out that in his version of sysback, during
the rc.boot, all of his local tape devices are
configured with cfgsctape, rather than checking for
the proper config method. Since customer has 3590's,
the proper method would be cfgatape.
John had customer modify rc.boot in sysback dir and
rebuild the boot image.
Aparently, all was well until some vg data was restored
to improper locations.
John will update further.
*ACTION PLAN: Leaving with John.
Response:
Here is the code that gets run where this is failing:
\# Ensure that all tape drives are
\# configured to block size 512.
lsdev -Cc tape | while read tapename dummy
do
ucfgdevice -l $tapename
chggen -l $tapename -a block_size=512
cfgsctape -l $tapename
done
This is later changed to work with Atape devices, 'cfgsctape' is the configuration method
for alot of standard tape drives, but none of the Atape drives. This is fixed at version
4, but the problem we have is that even though we aren't using the local drives, this
still fails.
Response:
Closing with Customer Approval
Support Line: sysback:restoring from aix 4.2.1 server to 4.1.5 client ITEM: HB1883L
Dated: August 1998 Category: N/A
This HTML file was generated 99/06/24~13:30:13
Comments or suggestions?
Contact us