Ce 6 Cookbook
Ce 6 Cookbook
Ce 6 Cookbook
HP-UX
October 2004
Version 6
http://eclipse.bel.hp.com:8888/users/qualification/
Thanks to the persons who made it possible to make this cookbook.
The maker of this cookbook
is not responsible
for the mistakes or the system problems
caused by the use of this cookbook.
Table of contents
Table of contents
Chapter 2: Storage.............................................................................27
1 Overview of most common commands for diskarrays ...............................................................27
ID drive/FRU –........................................................................................................................33
Include - auto ..........................................................................................................................33
Include - manual......................................................................................................................33
2 Special commands for VA 7XY0 ............................................................................................34
3 Argument for diskarray commands.........................................................................................37
4 Autoraid .............................................................................................................................38
5 DS2100 .............................................................................................................................39
6 DS2300 .............................................................................................................................40
7 DS2405 .............................................................................................................................42
8 FC60..................................................................................................................................44
9 SC10 .................................................................................................................................46
10 FC10 ..............................................................................................................................47
2
Table of contents
11 Nike array.......................................................................................................................49
12 VA7xy0...........................................................................................................................51
VA Troubleshooting Do’s and Don’ts .........................................................................................51
VA7100 ................................................................................................................................52
VA7110 – VA7400 ................................................................................................................54
VA7410 ................................................................................................................................55
Chapter 5: Procedures........................................................................91
1 SCSI bus error troubleshooting and log decoding ....................................................................91
2 Interpreting the FC hardware path .........................................................................................92
Types used for FC-addressing ...................................................................................................94
PDA: Peripheral Device Addressing........................................................................................94
LUA: Logical Unit Addressing ................................................................................................94
3
Table of contents
4
Table of contents
5
LVM and disk replacement
Verify that you have a defective disk that needs to be replaced. Use sysdiag or mstm to diagnose.
Depending on the disk failure, you may need to collect all or some of the following information using a
combination of the various data gathering commands listed here below.
Bit out the special device file to pin point the disk hardware path.
# lssf /dev/dsk/cXtXd0
The logical volumes residing on the volumegroup to which the disk belongs.
# vgdisplay -v /dev/vgXX
Get the map of the logical extends that correspond to the physical extends of each volume.
# pvdisplay -v /dev/dsk/cXtXd0
Get the characteristics and status of each logical volume in the faulty volume group.
# lvdisplay -v /dev/vgXX/lvolX
Get the BDRA information from the boot disk. (Boot Data Reserved Data)
# lvlnboot –v
If the boot disk is mirrored, record the hardware path and the software path of the mirror copy, from the
command above.
Use the dd command to identify a disk without crawling all over the floor. When you issue the command,
look for the solid blinking green LED. Ctrl+C will stop the read. Be sure disk is CLAIMED.
Note: If the disk is dead this command will not work. Also, dd is a good on line test. If the disk hangs or
if you get an I/O error, then the disk is defective.
# dd if=/dev/rdsk/cXtXd0 of=/dev/null bs=1024k
6
LVM and disk replacement
!!!! When you have to change a FC disk or EFI boot disk read section 3.
7
LVM and disk replacement
The same procedure as for a normal SCSI disk. See flowchart in section 2
The ‘World-wide name’ (unique identifier) for device at loop ID xx has changed when you replace a disk. If the device has been replaced intentionally, please use the fcmsutil
replace_dsk command to allow the new device to be used.
The main reason why this command was introduced is to avoid data corruption due to a user accidentally plugging in another device at the same nport_id.
When the system (initiator) establishes a login session (PLOGI) with the device (the target), it authenticates the device.
This authentication is to make sure that it is talking to the same device (at the nport_id).
Once the replace_dsk command is run, there will be no authentication on the device during the next login but the device will be authenticated on all subsequent logins.
To locate the N-port ID of the device you replaced use devstat all. Or you can use its Loop-ID.
Note: After running above command, you will get a message indicating the port will not be
authenticated
8
LVM and disk replacement
NOTE: A physical volume is considered to be attached, if the pvdisplay command is able to report a valid
status (unavailable/available) for it. Otherwise it’s called unattached.
In this LVM chapter we only consider the official procedure to be supported. The reason is that
there are potentially serious problems with replacing an attached device. Although the pvdisplay indicates
the device is unavailable, LVM could still be trying to recover it.
The consequences can be data corruption or obscure problems that can be difficult to track down, due to
the LVM metadata on the device being improperly written.
Follow these steps to replace a hot-swap disk module for attached Physical Volumes, which means that the
disk was not defective/unaccessible at the time the volume group was activated.
Hot-swapping a disk which was defective during activation (unattached) requires a different sequence of
commands. Skip to the alternative procedure, 5. Hot-Swap Procedure for Unattached Physical Volumes.
1. Reduce any logical volumes that have mirror copies on the faulty disk so that they no longer mirror
onto that disk.
NOTE: Be advised to check first, what LVs have mirror extents allocated on the faulty disk (to be
checked with pvdisplay –v /dev/dsk/cXtXdX). Then you should check for each found LV how it is
mirrored (use lvdisplay –v /dev/vgXX/lvolX). If the mirror extents span more than one PV then it is
highly recommended to specifiy all PVs with the lvreduce command that are in the “same mirror set
of disks” as the faulty one. Otherwise LVM may pick the “wrong” disks for reduction, leading to
undesired results (e.g. asymmetrical layouts). Take a note of this PV list, since you need this
information later when you re-establish the mirror using lvextend.
where list of PVs is the the list of devices determined according to the note above. We use the –A n
option to prevent the lvreduce command from performing an automaticic vgcfgbackup operation,
which is likely to get stuck on accessing a defective disk.
2. Replace the faulty disk. Please refer to the appropriate administration guide for
instructions on how to replace the disk.
Do an ioscan on the replaced disk to insure that it is accessible (CLAIMED) and also as a double
check that it is a proper replacement (see note above).
# ioscan -f /dev/dsk/cXtXdX
3. For fibre channel disks perform the replace_dsk steps described in the section 3.
4. Restore the LVM configuration/headers onto the replaced disk from your backup of the LVM
configuration.
# vgcfgrestore -n VG /dev/rdsk/cXtXdX
5. Attach the new disk to the active volume group with the vgchange command.
# vgchange -a y vgXX
or # vgchange -a e vgXX (for exclusively activated Cluster VGs)
9
LVM and disk replacement
6. Important: If the disk is the mirror of a root disk, then you must configure the LIF/BDRA
see LIF/BDRA Configuration Procedure
7. Lvextend the mirrors back onto the replaced disk. This may take several minutes as
it will have to copy all the data from the original copy of the data to the mirrored
extents. The logical volume(s) are still accessible to users' applications during this
command.
# lvextend -m 1 /dev/vgXX/lvolX /dev/dsk/cXtXdX & (for 1 way mirroring)
or
# lvextend -m 2 /dev/vgXX/lvolX /dev/dsk/cXtXdX & (for 3 way mirroring)
A shell loop like this could be used to extend a bunch of lvols automatically:
# for lvol in lvol1 lvol2 lvol3 ... (specify any LV you need to mirror)
> do
> lvextend –m 1 /dev/vgXX/$lvol /dev/dsk/cXtXdX
> done
10
LVM and disk replacement
Follow these steps to replace a hot-swap disk module for unattached physical volumes.
1. Replace the faulty disk. Please refer to the appropriate administration guide for instructions on how
to replace the disk.
Do an ioscan on the replaced disk to insure that it is accessible (CLAIMED) and also as a double
check that it is a proper replacement .
# ioscan -f /dev/dsk/cXtXdX
2. For fibre channel disks perform the replace_dsk steps described in the section 3.
3. Restore the LVM configuration/headers onto the replaced disk from your backup of the LVM
configuration.
# vgcfgrestore -n VG /dev/rdsk/cXtXdX
4. Attach the new disk to the active volume group with the vgchange command.
# vgchange -a y vgXX
or
# vgchange -a e vgXX (for exclusively activated Cluster VGs)
5. Important: If the disk is the mirror of a root disk, then you must configure the LIF/BDRA
see LIF/BDRA Configuration Procedure
6. Resynchronize the mirrors of the replaced disk. This may take several minutes as it will have to
copy all the data from the original copy of the data to the mirrored extents. The logical volume(s)
are still accessible to users' applications during this command.
# vgsync vgXX &
11
LVM and disk replacement
This subprocedure installs/updates information on disk that is mandatory for boot support.
Therefore it is referenced from several other parts of this chapter.
1. Write LIF header and LIF files (ISL, AUTO, HPUX, LABEL):
# mkboot -l /dev/rdsk/cXtXdX
# lifls –l /dev/rdsk/cXtXdX (to ckeck it)
NOTE: By default, LVM enforces a quorum of >50% of a VG’s PVs being available at activation
time. If e.g. the root VG contains 2 PVs, then the system rejects to boot unless you disable the
quorum check using the –lq option.
4. Write content of LABEL file, i.e set root, boot, swap and dump device:
NOTE: This step can be omitted if you replace a failed mirror disk. Then this information has
already been restored by vgcfgrestore. To be sure to have the latest information on the disk, just
do the following steps.
# lvlnboot -r /dev/<rootVG>/lvol3
# lvlnboot -b /dev/<rootVG>/lvol1
# lvlnboot -s /dev/<rootVG>/lvol2
# lvlnboot -d /dev/<rootVG>/lvol2
# lvlnboot –v (to ckeck it)
12
LVM and disk replacement
1. Reduce any logical volumes that have mirror copies on the faulty disk so that no longer mirror onto
that disk.
#lvreduce –m 0 –A n /dev/vg00/lvolX /dev/dsk/cXtYdZs2
3. Verify the disk partitions (cAtBdC is primary disk, cXtYdZ is mirror disk)
At a cold-installed UX 11.23 system the partition sizes are different compared to UX 11.22.
Use idisk to check the partition sizes.
# diskinfo /dev/rdsk/cAtBdCs1 | grep size
# diskinfo /dev/rdsk/cAtBdCs3 | grep size
6. Create the new device files for the new partitions (cXtYdsZ, s2, (s3))
# insf -e –Cdisk
# efi_fsinit -d /dev/rdsk/cXtYdZs1
NOTE: This step is not neccessary if it can be guaranteed that the mirror disk does not contain a
valid EFI filesystem. In this case efi_fsinit will be done automatically by the subsequent mkboot
command. But if you take e.g. an old UX 11.22 boot disk as mirror disk, mkboot will not
automatically run efi_fsinit. As a result only 100MB of the 500MB EFI partition (s1) can be used.
9. Use mkboot to format the EFI partition (s1) and populate it with the EFI files below /usr/lib/efi/
and to format the LIF volume (part of s2) and populate it with the LIF files (ISL, AUTO, HPUX, LABEL)
below /usr/lib/uxbootlf:
13
LVM and disk replacement
# mkboot -e -l /dev/rdsk/cXtYdZ
# efi_ls -d /dev/rdsk/cXtYdZs1 (to check EFI)
# lifls -l /dev/rdsk/cXtYdZs2 (to check LIF)
10. Check the content of AUTO file on EFI partition on primary disk:
NOTE: Specify the -lq option if prefer that your system boots up without interruption in case of a
disk failure:
# mkboot -a "boot vmunix -lq" /dev/rdsk/cXtYdZ
# mkboot -a "boot vmunix -lq" /dev/rdsk/cAtBdC
11. Check the content of AUTO file on EFI partition on new mirror disk:
12. Copy the HP service partition (UX 11.23 only): (skip this, if you don’t have a service partition)
14. Restore the LVM configuration/headers onto the replaced disk from your backup
#vgcfgrestore –n vg00 /dev/rdsk/cXtYdZs2
15. Attach the new disk to the active volume group with the vgchange command
#vgchange –a y vg00
14
LVM and disk replacement
2. Verify the disk partitions (cAtBdC is primary disk, cXtYdZ is mirror disk)
At a cold-installed UX 11.23 system the partition sizes are different compared to UX 11.22.
Use idisk to check the partition sizes.
# diskinfo /dev/rdsk/cAtBdCs1 | grep size
# diskinfo /dev/rdsk/cAtBdCs3 | grep size
6. # vi /tmp/partitionfile
5. # vi /tmp/partitionfile 3
2 EFI 500MB
EFI 100MB HPUX 100%
HPUX 100% HPSP 400MB
9. Create the new device files for the new partitions (cXtYdsZ, s2, (s3))
# insf -e –Cdisk
11. Use efi_fsinit to initialize the FAT filesystem on the EFI partition:
# efi_fsinit -d /dev/rdsk/cXtYdZs1
NOTE: This step is not neccessary if it can be guaranteed that the mirror disk does not contain a
valid EFI filesystem. In this case efi_fsinit will be done automatically by the subsequent mkboot
command. But if you take e.g. an old UX 11.22 boot disk as mirror disk, mkboot will not
automatically run efi_fsinit. As a result only 100MB of the 500MB EFI partition (s1) can be used.
12. Use mkboot to format the EFI partition (s1) and populate it with the EFI files below /usr/lib/efi/
and to format the LIF volume (part of s2) and populate it with the LIF files (ISL, AUTO, HPUX, LABEL)
below /usr/lib/uxbootlf:
# mkboot -e -l /dev/rdsk/cXtYdZ
15
LVM and disk replacement
13. Check the content of AUTO file on EFI partition on primary disk:
NOTE: Specify the -lq option if prefer that your system boots up without interruption in case of a
disk failure:
# mkboot -a "boot vmunix -lq" /dev/rdsk/cXtYdZ
# mkboot -a "boot vmunix -lq" /dev/rdsk/cAtBdC
14. Check the content of AUTO file on EFI partition on new mirror disk:
15. Copy the HP service partition (UX 11.23 only): (skip this, if you don’t have a service partition)
17. Restore the LVM configuration/headers onto the replaced disk from your backup
vgcfgrestore –n vg00 /dev/rdsk/cXtYdZs2
18. Attach the new disk to the active volume group with the vgchange command
#vgchange –a y vg00
16
LVM and disk replacement
You may come into a situation where you have to remove a PV from a VG that has failed or not even
physically connected but still recorded in the lvmtab. Such a PV is sometimes called a “ghost disk” or
“phantom disk”. You can get a ghost disk if the disk has failed before VG activation, maybe because the
system has been rebooted after the failure.
If you cannot use vgcfgrestore to write the original LVM header back to the new disk because a valid LVM
configuration backup file (/etc/lvmconf/vgXX.conf[.old]) is missing or corrupted you have to remove that
PV from the VG (vgreduce) to get a clean configuration.
NOTE: In such situations the vgcfgrestore command may fail to restore the LVM header, complaining
about a ‘Mismatch between the backup file and the running kernel’. If you are 100% sure that your backup
is valid you may override this check using the –R option.
In order to remove a PV from a VG you have to free it first, i.e. remove all logical extents from it. If the LVs
on such a disk is not mirrored data is lost anyway. If it is mirrored you need to reduce the mirror before
removing the PV.
A ghost disk is usually indicated by vgdisplay reporting more current PVs than active ones.
Additionally LVM commands may complain about the missing PVs:
# vgdisplay vg01
vgdisplay: Warning: couldn't query physical volume "/dev/dsk/cXtXdX":
The specified path does not correspond to physical volume attached to this volume group
vgdisplay: Couldn't query the list of physical volumes.
--- Volume groups ---
VG Name /dev/vg01
…….
Cur PV 2 (number of PVs recorded in the lvmtab)
Act PV 1 (number of PVs recorded in the kernel)
…….
17
LVM and disk replacement
The PV key of a disk indicates its order in the VG. The first PV has the key 0, the second has
the key 1, etc. This does not necessarily have to be the order of appearance in lvmtab altough
it is usually like that, at least when a VG is initially created.
The PV key can be used to address a PV that is not attached to the VG. This usually happens
if it was not accessible during activation, e.g. due to a hardware or configuration problem.
NOTE: The PV may be unattached due to some temporary problem during VG activation which is no
longer present. In this case you should try to re-activate the VG to force LVM to re-scan the
devices listed in lvmtab:
# vgchange -a y vgXX
or
# vgchange -a e vgXX (for exclusively activated Cluster VGs)
If the problem persists follow these steps to clear the situation:
Compared to the output above the ??? of the lvdisplay without –k have been replaced with the PV
key (0, 1).
If the LV was not mirrored, re-create the LV (lvcreate), create a FS on it (newfs) and recover
your data from backup.
18
LVM and disk replacement
Perform normal disk installation. You do need the mirroring software installed.
NOTE:
The -lq switch could be added to overwrite quorum on HIGH AVAILABILITY systems.
# mkboot -a "hpux -lq (;0)/stand/vmunix" /dev/rdsk/cXtXd0
If vPar use # mkboot -a “/stand/vpmon -a” /dev/rdsk/cXtYd0
19
LVM and disk replacement
Import the information form the disk that was removed by accident.
Note: If the system has not been rebooted since the accident, you should do it now.
If you have naming conflicts, remember that logical volumes can be renamed, but volume groups have to
be exported and imported for renaming.
20
LVM and disk replacement
We want to move the the root disk /dev/dsk/c0t6d0 to disk /dev/dsk/cYtXd0 staying within the same
volume group.
4. Move all physical extends from the source disk to the destination disk.
# pvmove /dev/dsk/cXtXd0 /dev/dsk/cYtXd0
6. Shutdown and reboot from the new root disk in maintenance mode.
ISL> hpux -lm (;0)/stand/vmunix
8. (Always use caution using these commands. Failure to do so can create data
loss!)
# lvlnboot -b /dev/vg00/lvol1
# lvlnboot -s /dev/vg00/lvol2
# lvlnboot -r /dev/vg00/lvol3
# lvlnboot -Rv ( On a 10.20 system verify lvol1 is defined as boot.)
# vgchange -a n /dev/vg00
9. Reboot the system and update the stable storage form the ISL prompt.
21
LVM and disk replacement
This procedure shows the basic use of vgexport and vgimport for moving a volume group/disks from one
system to an other. The following pages shows examples of using these two commands for recovery of
data disks which belong to volume groups.
1. Unmount all mounted lvols within the volume group you are exporting. Then deactive the volume
group.
# umount /dev/vgXX/lvolX /mount_point_name (repeat for all lvols)
5. Create the volume group and the group file on the other system.
# mkdir /dev/vgXX
# cd /dev/vgXX
# mknod /dev/vgXX/group c 64 0xXX0000 (XX = volume group number, info from step
3)
6. Import the new disk or disks into the other system. Note the possible new special device file which
might be different.
# vgimport -v -m /tmp/vgXX_map /dev/vgXX /dev/dsk/cXtXd0
Add the mount point directories and mount the lvols within that volume group.
Edit the /etc/checklist or /etc/fstab file.
22
LVM and disk replacement
If you can, make sure the database is not active before the split.
For the example we will use /dev/vg02/lvol1.
Use lvdisplay -v /dev/vg02/lvol1 to verify /dev/vg02/lvol1 is mirrored and current.
3. Perform the backup of the file system, "/dev/vg02/lvol1backup". (favorite backup utility)
4. Unmount the file system (if you had mounted it) in the split-off logical volume when you are ready
to merge the split pairs. When backup is done.
5. Merge the split-off logical volume and the original logical volume back into one mirrored logical
volume.
# lvmerge /dev/vg02/lvol1backup /dev/vg02/lvol1
23
LVM and disk replacement
*****************************************************************************
Use the "dd" command to identify a disk with out crawling all over the floor.
When you issue the command, look for the solid blinking green LED. "Ctrl+c" will stop
the read. DD can also be used as an online disk read test. Files in should equal files out.
# dd if =/dev/rdsk/cXtXd0 of =/dev/null bs=1024k
*****************************************************************************
24
LVM and disk replacement
25
LVM and disk replacement
Before you shutdown, document /etc/lvmtab, note which disks are in vg00.
Shutdown the system and connect the existing root drive to the new path.
Boot up the system and escape boot sequence, the boot off the root drive at the new hardware path to ISL.
3. Install the special file for the root disk at the new hardware path.
# insf -e -H <hardware_path>
7. Create the volume group and the group file to be used for the vgimport command.
# mkdir /dev/vg00
# mknod /dev/vg00/group c 64 0x000000
8. Import the root disk and any other disks that might be in the vg00 volume group.
# vgimport -v -m /tmp/vg00_map /dev/vg00 /dev/dsk/cXtXd0
<and any other disk in vg00>
10. Update the LIF's Lable file with the information contained in the BDRA.
# lvlnboot -Rv
DO NOT forget to update the Stable Storage new boot path from the ISL prompt!
26
Storage
Chapter 2: Storage
27
Storage
/var/opt/sanmgr/commandview/
server/logs/device/wwn/date
wwn is the 50060B0xxxxxx of controller
Log Directory /var/opt/hparray/log /var/opt/hparray/log date is year-month 2002-10
/var/opt/sanmgr/commandview/
server/logs/performance
xx-usage.log = daily armdsp -a
xx-cntrl.log
Logs files LXXXXXXX AMMXXXXX
where xx is a calander date
example 19-usage.log
logprn -a $id -v or –b
for valt
logprn -t All -v -a s/n > logprn.txt
armlog -c s/n > armlog.txt 1.07
armlog –e s/n > armlog.txt pre 1.07
for performance logs
armperf -c ARRAY -x COMMA s/n >
armperf_array.txt
armperf -c OPAQUE -x COMMA s/n >
armperf_opaque.txt
arraylog -u -e $id amlog -a $id -s -e opt avail
Logs Display $ id must be s/n
armperf -c LUN -x COMMA s/n >
logprint $id -s -e opt avail armperf_lun.txt
28
Storage
Remove logs kill monitor, rm files, touch kill monitor, rm files- lock, touch
/var/opt/hparray/log/LOGCATG catalo g logdel –t all –a s/n
restart monitor rm /var/opt/hparray/log/AMM* option stop time –e MMDDHHMM
arraylog -C -c $id rm /var/opt/hparray/AM60Srvr.lock
touch Note: only logs older than two months can
/var/opt/hparray/log/AMMCATLG be removed
restart monitor
amutil -p remove oldest log file
FIRMWARE AutoRaid 12 & 12H FC60 VA 7XX0
Directory /opt/hparray/bin /opt/hparray/bin /opt/samngr/commandview/client/fwdownload
armdownload -C –O cnt_# fwRev $id
armdownload –B fwRev $id
amdload -C all file_name $id
Download Controller download -C file_name $id armdownload -L JAx/Cx fwRev $id
amdload -S file_name $id bcc
clear advisory lock
armmgr –b FwDownload $id
clear advisory lock
armmgr –b FwDownload $id
armdownload –I -D drive_# fwRev $id
or clear advisory lock
Download Drive download -D # file_name $id amdload -D file_name $id
armmgr –b FwDownload $id
armdownload –D –P drv model_# fwRev $id
Failure to clear advisory locks could result
in a rebuild taking place
Migrate Controller armdownload -M # $id
download -M $id from primary
firmware where # is the source FRU M/C1or M/C2
29
Storage
31
Storage
32
Storage
33
Storage
Disabling Security –
Setting up security – New with 1.07
Allow normal access to commandview
/opt/sanmgr/commandview/client/sbin/
vi /etc/opt/sanmgr/commandview/server/
secadmin -host my_host
config/PanConfigParams.txt
username Administrator
SECURITY_ENABLED=true change true to false
password administrator
Resetting security on Windows ->Administrator
Commandview 1.07 add –user root –group Administrator –host my_host
remove old
secadmin Note :
C:\Program Files\Hewlett-Packard\sanmgr\
groups and users must be in /etc/passwd & /etc/group
commandview\server\data\UserDBFile
enter password
Resetting security on UNIX ->Adminsistrator
re-enter password
remove old
exit
/var/opt/sanmgr/commandview/server/data/
web GUI will require a valid user and password
UserDBFile
Obtaining license – New with 1.07
60 day grace period – free use Installing License – New with 1.07
License Documentation CMD 1.07 licApp - easy to use GUI – very slow
License web site http://www.webware.hp.com must run from a X window – like reflection X
1. Generate product password ; next requires file already be created
2. use CVSDM107UPGRADE as order # ; next # licApp –i HP_CommandView_SDM
Commandview 1.07 3. check product box for T1086A ; next 1. Browse to /var/opt/OV/HPOvLIC
licApp or licUtil 4. enter system name and select OS ; next 2. select cmdview_lic.txt ; Choose
5. enter customers email address & password ; next 3. import
6. enter license requestor & owner info ; next licUtil - requires more input
7. print and save license requires file already be created in
cut the license info and place in a file on the host /var/opt/OV/HPOvLIC/cmdview_lic.txt
/var/opt/OV/HPOvLIC/cmdview_lic.txt # licUtil –i cmdview_lic.txt HP_CommandView_SDM
save as one line in a file with no embedded cr’s.
7100/ 7400 7110/ 7410
Queue depth entered is assigned to both controllers Queue depth entered is assigned to a controller host
Queue Depth Formula # of disks in RG x Multiplier =que# bus
RG1 odd disks Formula # of disks in RG x Multiplier / 2=que#
RG2 even disks RG1 odd disks
34
Storage
35
Storage
36
Storage
Note: These are not all the possible arguments - check the man pages for more
options.
37
Storage
4 Autoraid
Commands
arraydsp -i gives the serial number of all the arrays
export <ID>=<serial number> gives a an arrayID (a name you assign so that you do not to use
the long serial number with the commands
arraydsp -a <arrayID> all (status)
arraydsp -c <arrayID> controller status
arraydsp -d <arrayID> disk status
arraydsp -h <arrayID> hardware status
arraylog -e <arrayID> displays the contents of disk array controller event log
arraymgr -s shut<arrayID> Shutdown array
drivetest -D AX -r 100<arrayID> performs read/verify test on 100% of disk AX on
<arrayID> non-destructive
dteststat -D AX -? <arrayID> displays results of drivetest on AX
logprint -d {-s starttime} {-e stoptime} <arrayID> logfile is put in default location for the
specified time (mmddhhmm)
/var/opt/hparray/log
Partial Menu Selection
Push "Menu", then use the "+" or "-" buttons to scroll through the menus, "enter " to activate a selection
Shutdown → Confirm (or Cancel)
View Settings → Disk States gives the state of each disk in the array
Rebuild in progress or complete
System State
Warnings
Number of disks
Primary Cntrl
SCSI ID x
SCSI ID y
Firmware
Array S/N
ActivSpare
L-Drv States
Cntrl Changes → Reset Cntrl
[Copy Firmware]
[Recover] Restores map of drives
[Switch Cntrl]
Reset Cntrl
38
Storage
5 DS2100
Notes
• Only 1 enclosure in supported in the SE mode.
Up to 3 enclosures for a maximum of 12 drives otherwise.
• Maximum cable length is 12 meter for LVD
• Disk system requires a terminator at the end of the bus.
If you connect DS2100 to TWO HBA (SCSI-id 6 and 7), switch can NOT be set to A
39
Storage
6 DS2300
40
Storage
SCSI ID
Physical Disk Slot
Full Buss Split Bus (Right BCC) Split Bus (Left BCC)
1 0 0
2 1 1
3 2 2
4 3 3
5 4 4
6 5 5
7 6 6
8 8 8
9 9 9
10 10 10
11 11 11
12 12 12
13 13 13
14 14 14
Right BCC 15 15
Left BCC 15
41
Storage
7 DS2405
42
Storage
Encl ID 5 6
Loop ALPA Loop ALPA
Disk slot Bus Tgt Bus Tgt
ID (hex) ID (hex)
1 80 5 0 55 96 6 0 3A
2 81 5 1 54 97 6 1 39
3 82 5 2 53 98 6 2 36
4 83 5 3 52 99 6 3 35
5 84 5 4 51 100 6 4 34
6 85 5 5 4E 101 6 5 33
7 86 5 6 4D 102 6 6 32
8 87 5 7 4C 103 6 7 31
9 88 5 8 4B 104 6 8 2E
10 89 5 9 4A 105 6 9 2D
11 90 5 10 49 106 6 10 2C
12 91 5 11 47 107 6 11 2B
13 92 5 12 46 108 6 12 2A
14 93 5 13 45 109 6 13 29
15 94 5 14 43 110 6 14 27
SES 95 5 15 3C 111 6 15 26
43
Storage
8 FC60
Commands
amdsp -i Gives the arrayID number of the machine
amdsp -a <arrayID> Gives the status of all the drives, bcc, & power
supplies. Also gives the firmware levels
amdsp -d -S <arrayID> Status information of all disk
swlist -l product | grep -e "HP Array" Identifies patch level & patches on
HPUX 10.20 & 11.0
ammgr -b <arrayID> resets battery cache, must be done after replacing battery
and after battery is charged, will not work if disk rebuild is in progress .
Identifying Split and Full Bus
• Split bus will have a will have 2 SCSI cables coming from the FC60 Controller (channels 1-6) to
each SC10. Each FC60 can handle up to 3 SC10's in a split bus configuration.
Ex: Channel 1 on the FC60 controller connected to BCC A on the first SC10 and channel 2
connected to BCC B on the same SC10.
• Full bus will have one cable coming from the FC60 Controller (channels 1-6) to each SC10. Each
FC60 Controller can handle up to 6 SC10's in full bus
Notes:
• Controller upper A, lower B, Each has a unique FC address assigned through dip switches.
Controllers are hot swappable but you must follow the proper procedures for replacing them.
Dimms and Simms may need to be transferred to new controller
• Cache Dimm must be the same on both controllers
• Unused LVD ports must be terminated
• Power supplies, MIA, fan controller module, BBU, disk, disk fan modules and power supply fan
modules are hot swappable
• JBOD - address 4, 5, 6 & 7 reserved for host controller. ARRAY - 5,6, & 7 reserved for host
controller
44
Storage
•Reading the codes ex: 0x2A = Ignore the 0x ," 2 " represents the leftmost set of four status LED's in binary ( Ο Ο Ο ) , "A" represents the right set of four status LED's in
binary ( Ο Ο ) A = 10
45
Storage
9 SC10
46
Storage
10 FC10
Enclosure Ids (0 through 9) are set with a dial on the LCC. This settings determines the FC-Al Ids and
corresponding HW-paths for each slot and LCC based on the Enclosure ID.
The Enclosure Ids on both LCCs in the same disk system must be identical
0 0 0 0 12 0 12 24 1 8 36 2 4 48 3 0
1 1 0 1 13 0 13 25 1 9 37 2 5 49 3 1
2 2 0 2 14 0 14 26 1 10 38 2 6 50 3 2
3 3 0 3 15 0 15 27 1 11 39 2 7 51 3 3
4 4 0 4 16 1 0 28 1 12 40 2 8 52 3 4
5 5 0 5 17 1 1 29 1 13 41 2 9 53 3 5
6 6 0 6 18 1 2 30 1 14 42 2 10 54 3 6
7 7 0 7 19 1 3 31 1 15 43 2 11 55 3 7
8 8 0 8 20 1 4 32 2 0 44 2 12 56 3 8
9 9 0 9 21 1 5 33 2 1 45 2 13 57 3 9
LCC 10 0 10 22 1 6 34 2 2 46 2 14 58 3 10
RSVD 11 0 11 23 1 7 35 2 3 47 2 15 59 3 11
47
Storage
Encl ID 5 6 7 8 9
0 60 3 12 72 4 8 84 5 4 96 6 0 108 6 12
1 61 3 13 73 4 9 85 5 5 97 6 1 109 6 13
2 62 3 14 74 4 10 86 5 6 98 6 2 110 6 14
3 63 3 15 75 4 11 87 5 7 99 6 3 111 6 15
4 64 4 0 76 4 12 88 5 8 100 6 4 112 7 0
5 65 4 1 77 4 13 89 5 9 101 6 5 113 7 1
6 66 4 2 78 4 14 90 5 10 102 6 6 114 7 2
7 67 4 3 79 4 15 91 5 11 103 6 7 115 7 3
8 68 4 4 80 5 0 92 5 12 104 6 8 116 7 4
9 69 4 5 81 5 1 93 5 13 105 6 9 117 7 5
48
Storage
11 Nike array
Get into FE mode, go to the Presentation Utility and type lower case i or the # key.
The i key gives you a hardware inventory of the boards and hard drive sizes.
The # key gives you more detail about the hard drives themselves.
CTRL P at the same time then holding down the SHIFT key type USER.
1. Boot from the dfdutil utility tape or CD that has Nikearry file.
2. Go into ODE
3. Disable one of the Nike processors. (Phoenix controllers are done in grid manager by going to
Change SP parameters menu, then go into Select soft disable).
4. Type nikearry to run the firmware utility.
5. Enter the password.
6. Put in the path or paths of the sp controllers to update.
7. Type download
8. Get address of sp controllers still enabled (if updating more than 1 controller
9. Type SP_925_00 for Suana or older controller or SP_955_01 for phoenix or newer controller.
49
Storage
10. Once download is complete enable the sp controller that you disabled earlier. If firmware is not
moved over from updated controller to the controller that was disabled repeat from step 3 disabling the
controller that was just updated and update the other one.
11. After the sp controllers are updated you can now update the nike drives is need be.
12. Type download
13. Put in firmware file NIKE2GHH03 for ST15550N drives or NIKE2GLP08 for ST32550N drives.
14. Can do a reset command when done to check if the updates took.
15. Check grid manager for drive updates.
50
Storage
12 VA7xy0
When troubleshooting a VA disk array (VA7100, VA7400, or VA7410), it is very important to keep the following
points in mind to avoid data loss and minimize downtime to the customer.
1) DO NOT power off the array. Disk arrays are designed to be serviced while powered up, so powering
down the array to replace parts can cause problems and even data loss under rare circumstances. When at
all possible, leave it powered up.
2) DO obtain and analyze all log data before replacing or even removing any parts. This should be done
with logprn on the host to get historical data, armlog to pull the current logs from the array, and vfplog from
the serial port VFP in the event that the array is not accessible from the host.
3) DO find out if the host can communicate with the array with armdsp and armlog. If the array is not
accessible from the host, disconnect the fibre cables to the host and connect a laptop to the serial port of each
controller one at a time. Collect logs and look for any error indications displayed while logging the output to
a file. If the array appears to be continually resetting itself, the initialization steps will be displayed along
with any messages like “Abterm.” Only the controller experiencing an error or abterm will display that error
message.
4) DO NOT hesitate to get help if you see “No Map,” “No Quorum,” or “Abterm,” and are not sure exactly
what to do. These conditions are very serious, and data loss can result if the wrong decisions and steps are
taken. When in doubt, escalate.
5) DO know what will be required when you escalate. An armdsp –a, logprn, and armlog or vfplog will be
asked for to proceed with an escalation. DO everything you can to get that log data.
6) DO NOT pull batteries off of any controller removed from the array unless you are sure the array is in a
Ready or Single Controller Warning state with one functioning controller installed in the array; or, you are
sure there is no usable map in memory; or, you want to purposely clear the maps in memory. This step is
used to safely move the memory DIMM(s) from a suspect controller to a new replacement controller.
7) DO NOT troubleshoot a controller problem without parts. If it means you have to wait several hours to get
parts, and then wait. If you can get the array working in the Single Controller Warning state by pulling one
of the controllers’ part way out, then leave the array in this configuration until parts arrive.
8) DO NOT swap memory DIMM(s) from one controller to another if you suspect the problem is memory. You
must keep at least one controller working in the array at all times. If you have to wait for parts, then wait.
Replace the memory on the one controller removed, reinstall and make sure the array comes up to the Ready
state.
9) DO NOT use the vfprecover command unless you are absolutely sure the maps cannot be recovered from
battery backed memory on any other controller. This command along with the online version, armrecover, is
very dangerous. The VA array does not work the same way as the Autoraid Model 12 & 12H. If any portion
of the maps are recoverable but ultimately the command fails, then a corrupt, partial version of the maps will
become the most current. This means the maps are no good and the data is effectively gone anytime the
command is not completely successful. A MULT_FAIL_EH event will likely be found in the logs.
10) DO update firmware and software as soon as possible and keep current.
51
Storage
VA7100
52
Storage
A B
Host FC LED's Indication
Green Amber
off off unit not powered or link down or GBIC not installed
on off normal operation
off on controller fault
on on controller fault
immediate intervention required, GBIC requires attn. or operator
off flashing
induced fault
A B
Controller LED's (on Host) Indication
Green Amber
CONTROLLER
off off unit not powered or link down or GBIC not installed
on off normal operation
off on array controller fault
on on array controller fault
off flashing immediate intervention required. check logs or operator induced fault
flashing off array controller self-test in progress or controller I/O activity
BATTERY
off off unit not powered battery not installed
on off normal operation
off on module fault
on on module fault
off flashing operator induced manually
flashing off battery self-test in process or battery is powering NVRAM contents
DIMM
A B
A-DIMM 1 B-DIMM 1 Indication
C-DIMM 2 D-DIMM 2
off off unit not powered, battery not installed
on off normal operation
off on DIMM fault
on on DIMM fault
off flashing operator induced fault manually
flashing off DIMM self-test in process
53
Storage
VA7110 – VA7400
54
Storage
VA7410
55
Patches
Chapter 3: Patches
1 Install
3 Copy a depot
4 Registration of a depot
Retrieve the patches for instance from a patch server or web site:
mkdir /tmp/patches
cd /tmp/patches
sh PHxx_A
sh PHxx_B
sh PHxx_C
# cd /tmp
# for i in PH*
> do
> sh $i
> done
#
56
Patches
mkdir /tmp/patch_depot
swcopy -v -s /tmp/patches/PHxx_A.depot PHxx_A @ /tmp/patch_depot
swcopy -v -s /tmp/patches/PHxx_B.depot PHxx_B @ /tmp/patch_depot
swcopy -v -s /tmp/patches/PHxx_C.depot PHxx_C @ /tmp/patch_depot
This can also be done with the following commands in one operation:
# mkdir /tmp/patch_depot
# for i in PH*.depot
> do
> swcopy -s ${PWD}/$i \* @ /tmp/patch_depot
> done
#
If you need to copy the depot to a tape use the following command :
6 Patch Clean-up
57
Servers
Chapter 4: Servers
58
Servers
59
Servers
60
Servers
2 A L N general
For revision A GSP, the console path must be assigned to path 0/0/4/0.
For revision B/C GSP, the console path must be assigned to path 0/0/4/1.
Type CTRL+B
At login and password just hit enter or ask login of CU
At the ' GSP>' prompt type one of the following commands
SL logs troubleshooting
Alert Level: How the problem has affected the system operation Alert Levels
0 No failure detected
Source: What major part of the system the alert is referring to (i.e, 1 Service processor alert, no failure detected. Used by the
platform, memory, processor, etc.) SP to perform a specific action, such as changing the
system state from boot start to boot complete
Source Detail: What subpart of the system the alert is referring to 2 Non-urgent operator attention required (i.e, Notification that a
(i.e, cabinet fan, Dimm, high voltage DC power, etc.) previous failed fan is now functioning)
3 System blocked waiting for operator input.
Source ID: Specific FRU referred to in Source & Source Detail (i.e, 4 Unexpected configuration change detected
cabinet fan #4) 5 Boot possible, environmental problem exist
6 Boot possible, pending failure or environmental problem with
Problem Detail: Specific problem info (i.e, power off, functional action required (i.e, power failure, overtemp, etc.)
failure, etc. 7 HPMC detected
8 Boot possible, performance impaired
Timestamp: When the problem occurred. 9 Reserved
10 Boot possible, functionality lost
11 Reserved
12 Software failure (i,e, HPUX panic)
13 System hang detected via timer popping
14 Fatal power or environmental problem prevents operation
15 Fatal hardware or configuration problem prevents operation
SOURCE SOURCE DETAIL SOURCE ID PROBLEM DETAIL FRU ACTION
1 - Processor Processor check processor status,
"in pr" in BCH or "ss"
in GSP
2 - ProcessorCache Processor check proc status
3 - PDH System Board replace sytem board or
run ss_update if new
system board
4 - Power 1 - AC Mains 9 - Power off AC Power check AC supply
4 - Power 3 - Low voltage DC CPU Support CPU Support replace CPU support
power Module # Module module ref. in Source ID
4 - Power 4 - High voltage DC Power supply # Power Supply replace power supply
power ref. in Source ID
61
Servers
62
Servers
3 A 400/500 (RP24xx)
A400/500 internals
A400/A500 IO-paths
63
Servers
IO Paths /Locations
Physical
Description Device
Path
10/100/1000BT Onboard Lan 0/1/2/0 ACPI(HWP0002,100)/PCI(2|0)
10/100 Onboard Lan 0/0/3/0 ACPI(HWP0002,0)/PCI(3|0)
External Ultra320 LVD SCSI VHDCI 0/1/1/1 ACPI(HWP0002,100)/PCI(1|1)
MP External RS232 0/6/1/1 ACPI(HWP0002,700)/PCI(1|1)
MP External RS232 Port UPS 0/6/1/0 ACPI(HWP0002,700)/PCI(1|0)
MP VGA 0/6/2/0 ACPI(HWP0002,700)/PCI(2|0)
IDE Controller for rx2600 0/0/2/0 ACPI(HWP0002,0)/PCI(2|0)/Ata
Onboard Serial A 250/1
Onboard Serial B 250/2
64
Servers
65
Servers
66
Servers
67
Servers
L1000 – L2000
68
Servers
L1500 - L3000
69
Servers
70
Servers
71
Servers
Fru Location
72
Servers
73
Servers
74
Servers
MP
Login: Admin
Passwd: Admin
75
Servers
BCH
Main menu:
Boot Continues booting from the specified path
Path Displays or sets the PRI, HAA or ALT boot paths
SEArch Searches for boot devices
ScRoll Displays or sets the scroll mode- Pauses every 24 lines. Turn OFF before
using Error Log or PIM commands.
RECONFIGRESET Reset to Reconfig Complex Profile. Same as MP> RR. Sets “BIB”. Use
after CPU/Cell config!
Config menu:
CEllConfig Config/ Deconfig cell <cell> [ON] [OFF]
CPUconfig Config/ Deconfig cpu [<cell>[< cpu>]] [ON] [OFF]
PathFlags Display/ set Boot
PF <path> = [PRI | HAA | ALT] <value> = [0| 1| 2| 3] 0 = BCH, 1 = boot this
path -> BCH, 2 = boot this path -> next path, 3 = skip -> next path
Information menu:
ALL Display all of the information. [< cell>] Displays processor, PDC, Chip revision, Boot
information, I/O information and Memory.
FwrVersion Displays the firmware version of the various firmware modules in the partition.
Service menu:
Battery Displays the status of the battery on the cell board <cell> = [0 - 3]
DimmDealloc Display or deallocate a dimm on a cell - displays the configuration status of the memory
DIMMs on the cell or schedules them for configuration, or deconfiguration on the next
reboot. <cell> <[ 0A, 0B, 0C, 0D,..., 3A, 3B, 3C, 3D]> <[ ON | OFF]>
PDT Used to display and clear the contents of the page deallocation table. <cell>
CLEAR
PIM Displays the PIM logged on the first, most severe occurrence HPMC, LPMC or a
Transfer of Control (TOC) for any processor present in the partition. <cell>
<cpu> <[ HPMC | LPMC | TOC]>
SCSI Displays/ sets the SCSI controller parameters. SCSI [< path> [INIT| RATE|
TERM| WIDTH|DEFAULT [< id>]]] All parameters are kept in Cell NVRAM –
not complex profile
76
Servers
77
Servers
or DDS 1
Cell
1/0/0/3/0.6.0 Internal disk Slot 1A Acpi(HWP0002,100)/Pci(3|0)/Scsi(Pun6,Lun0)
1
built in LAN LAN Cell
1/0/1/1/0/4/0 Acpi(HWP0002,108)/Pci(1|0)/Pci(4|0)
1000bt SCSI 1
External SCSI
LAN Cell
1/0/1/1/0/1/0 Fast Wide Acpi(HWP0002,108)/Pci(1|0)
SCSI 1
LVD
Cell
1/0/1 PCI card Slot 8 Acpi(HWP0002,108)
1
Cell
1/0/2 PCI card Slot 7 Acpi(HWP0002,110)
1
Cell
1/0/4 PCI card Slot 6 Acpi(HWP0002,120)
1
Cell
1/0/6 PCI card Slot 5 Acpi(HWP0002,130)
1
Cell
1/0/8 PCI card Slot 1 Acpi(HWP0002,140)
1
Cell
1/0/10 PCI card Slot 2 Acpi(HWP0002,150)
1
Cell
1/0/12 PCI card Slot 3 Acpi(HWP0002,160)
1
Cell
1/0/14 PCI card Slot 4 Acpi(HWP0002,170)
1
78
Servers
79
Servers
80
Servers
81
Servers
Note 1: The I/O slot HW pathes change with the new PCI-X I/O chassis from cell#/0/LBA#/0 to
cell#/0/LBA#/1.
Olympia, Rainier and SEU have always PCI-X chassis. Keystone may have PCI or PCI-X.
Note 2: These slots are available only if the optional A6434A Server Expansion Unit (SEU) is
installed. The SEU provides an additional 16 PCI-X I/O slots for hp9000 rp84xx servers or hp
Integrity rx8620 servers. Each cell in the hp9000 rp84xx server or hp Integrity rx8620 server
enables eight IO slots. The server must contain four cells to enable all 16 of the SEU PCI-X IO
slots.
82
Servers
83
Servers
84
Servers
11 V-CLASS TROUBLESHOOTING
Flashing Light Bar - indicates the presence of an error code and an environmental error detected at power
on. Start with "dcm 0” or "pce_util".
Using dcm 0
At one of the bottom (KSH, lower 2) windows at the test station type:
dcm 0
For bad memory, look for a bank that has been deconfigured, to find dimm ex:
EMB0
======
B0S2 - 64H00 / 64MB the "H" signifies it being "hardware disabled" because of a bad dimm
in the bank
B1S2 - 64S00 / 64MB look for dimm that has been software (64S00) deconfigured, this is
the bad dimm, 128 mb.
B2S6 - 64H00 / 64MB
B3S6 - 64H00 / 64MB
Using pce_util
pce_util generates an error code "LED" hex code. 00-67 shuts down the system, 68-73 leaves the system
up.
At one of the lower 2 windows type:
pce_util for 2200's
pce_util -n node_number for 2500/2600
Look under the LEDS heading to find the hex error code. ex:
85
Servers
/spp/data$ ll | more hit the space bar to scroll down to" hard_hist" file to check for last
date and time of HPMC. No current entry could indicate a
software problem.
more or tail -xxx hard_hist file, call backline to translate if nothing obvious is found
V2500/2600 (Multinode)
cd /spp/data directory
/spp/data$ ll | more to find the complex_name (node name)
cd /spp/data/complex_name
more or tail -xxx hard_hist hit the space bar to scroll down to" hard_hist" file to check for last date
and time of HPMC. No current entry could indicate a software problem.
Call backline to translate if nothing obvious is found.
86
Servers
12 Superdome
Blower
Cell with
CPU and
memory
BPS
IO bay
87
Servers
88
Servers
x/0/1/x/x Core IO 0 0
x/0/1/x/x 2x PCI card 1 1
x/0/2/x/x 2x PCI card 2 2
x/0/3/x/x 2x PCI card 3 3
x/0/4/x/x 4x PCI card 4 4
x/0/6/x/x 4x PCI card 6 5
x/0/14/x/x 4x PCI card 14 6
x/0/12/x/x 4x PCI card 12 7
x/0/11/x/x 2x PCI card 11 8
x/0/10/x/x 2x PCI card 10 9
x/0/9/x/x 2x PCI card 9 10
x/0/8/x/x 2x PCI card 8 11
89
Servers
90
Procedures
Chapter 5: Procedures
0 x 1 f 0 1 4 0 0 0
major# bus# target lun flags
c1 t4 d0
Record the bus# and the hardware path of the suspect bus:
Then go to EXERCISE A DISK section run diskinfo & dd command
91
Procedures
Loop ID +0 +1 +2 +3 +4 +5 +6 +7 +8 +9 +A +B +C +D +E +F
(hex)
00 EF E8 E4 E2 E1 E0 DC DA D9 D6 D5 D4 D3 D2 D1 CE
10 CD CC CB CA C9 C7 C6 C5 C3 BC BA B9 B6 B5 B4 B3
20 B2 B1 AE AD AC AB AA A9 A7 A6 A5 A3 9F 9E 9D 9B
30 98 97 90 8F 88 84 82 81 80 7C 7A 79 76 75 74 73
40 72 71 6E 6D 6C 6B 6A 69 67 66 65 63 5C 5A 59 56
50 55 54 53 52 51 4E 4D 4C 4B 4A 49 47 46 45 43 3C
60 3A 39 36 35 34 33 32 31 2E 2D 2C 2B 2A 29 27 26
70 25 23 1F 1E 1D 1B 18 17 10 0F 08 04 02 01 00 --
92
Procedures
93
Procedures
Sw#: Switch domain number of switch device (not the host) is connected to.
Port#: port number, with some brocade switches port#=port+16
94
Procedures
2. fcmsutil /dev/fcmsX Shows the status of the of the fibre channel card,
/tdX topology, drive state etc Look for:
Topology = IN_LOOP
Driver state = READY
3. fcmsutil /dev/fcmsX stat gives error statistics on the fibre channel card
/tdX Scroll down and look for these headings. Run this
command multiple times to see if the Counts are
increasing:
Bad Tx Char Count: 100 to 300 is normal
Link Fail Count:
Loss of signal Count:
Loss of Sync Count:
4. fcmsutil /dev/fcmsX reset Resets the fibre channel card. Don’t do a reset on a
/tdX running system.
95
Procedures
4 Exercise a disk
dd if =/dev/rdsk/cxtxdx of=/dev/null bs=1024k
or
dd if =/dev/rdsk/cxtxdx of=/dev/null bs=1024k & & causes this command to run in background
tail –f syslog.log to receive ongoing errors from the disk
Ctrl + C to quit this process
Find Error info on a disk using MSTM if no errors are seen the problem could be a patch issue,
controller problem, disk header problem etc.
type mstm
highlight the disk you want to run the test on
tools, expert, run
or
tools, information, run, log tab log, then view.
96
Procedures
ISL>hpux –is For single user mode. Only root and /stand directories
are mounted. Only vg00 is activated. To run commands
such as (ioscan, diskinfo, etc.)
type: mount /usr
ISL> hpux –is (;0)/stand/vmunix.prev To boot to previous kernel in single user mode
ISL> hpux –lm Boots to maintenance mode. Root directory is the only
one mounted. No volume group is activated
97
Procedures
6 EMS errors
Summary:
I/O power supply has failed
Desciption of Error:
Probable Cause/Recommended Action:
If problem is not obvious call backline to have Event # translated or for event code lookup go to:
http://wojo.rose.hp.com/hp ux/onlinedocs/diag/ems/eme_summ.htm
98
Procedures
Using MSTM
Type ‘mstm’ at the root prompt.
Move the cursor to the processor and highlight it by pushing the spacebar.
Tab over to ‘Tool’ on the toolbar, push Enter (or Return)
Move the cursor to ‘Information’ and Enter, highlight ‘Run’ and Enter.
Record the chassis codes. Do this for each processor.
You can also find HPMC’s in ‘var/tombstone/ts99’ file if the machine has diagnostics loaded.
99
Procedures
4. To exit Logtool Utility press F8 then F2 for OK to exit. This puts you back to the main menu in
mstm.
5. To view any information off of a specific scsi address or memory or cpu information; Arrow down
to the desired path, press the spacebar to highlight that address, press Tab then T for Tools then
I for Information and R for Run. This is a great help with a memory problem, cause you can view
the memory size in each pair bank of memory. You will know what size memory we will need to
bring out with us. Be aware that when looking at the memory in each pair bank that is the total
memory and not the actual memory simm size. Example – Memory information says that in 1A/B
the total physical size in (MB) is 16. That means the actual memory simms in each bank are 8mb
simms. Both banks total are 16 but you have two memory simms making up that total. So a rule
of thumb is take half the total and that is the size of the actual memory simm in each bank. Press
F8 for Done to get back to main menu screen in mstm
100
Procedures
To find the hostname, type the following before shutting down the system:
just ask the customer
or #/usr/bin/hostname
or #/usr/bin/more /etc/hosts (usually first entry)
Try before starting mstm if clear command works, if not you have the wrong terminal emulation.
101
Procedures
PDC upgrade
102
Procedures
103
Procedures
1. Before you start with the GSP upgrade make sure that you have your laptop available via ftp. IP-addressed in the same
SUBNET
And the right directory were the files are located
example : /tmp/GSP/APPLICATION.ROM
/tmp/GSP/BOOTSTRAP.ROM
/tmp/GSP/upgrade.cnf
3. Ping the LAN where the GSP files are located via the XD command item #3 (LAN access (PING)
GSP> XD
Non destructive tests:
1. Parameters checksum
2. I2C access (get Power Monitor status)
3. LAN access (PING)
4. Modem selftests
4. Reset the GSP via the XD command (use the R option) This will free up any unallocated memory for the GSP update.
GSP> XD
Non destructive tests:
1. Parameters checksum
2. I2C access (get Power Monitor status)
3. LAN access (PING)
4. Modem selftests
GSP> XD
GSP>he
……
GSP>xu
XU
This command activates the upgrade mode. All connections will be closed, the network will be stopped, the session will be
aborted and the modem connection will be dropped.
Please, confirm your intention to activate the upgrade mode (Y/[N]) : y
Enter source system IP address : xxx.xxx.xxx.xxx
104
Procedures
GSP>he
If the firmware is B.02.14 (C.02.09) or below, you will have to update to B.02.15 (C.02.10) prior to go
to the latest.
1. Just prior to update, insure that the following operating system patches are installed to change the device drivers. In
addition to the firmware update, these patches are required to resolve the UPS HPMC issue.
11.00
Console Driver Patch PHNE_27788
INSF Patch PHCO_27370
11.11
Console Driver Patch PHNE_27760
INSF Patch PHCO_28169
2. If B.02.14 (C.02.09) or below, you will first need to update to B.02.15 (C.02.10).
The B.02.15 (C.02.10) version requires two firmware update steps. You should reset the GSP with GSP>xd command
after the first updateonly image download. Not completing this step may cause the update in the next step to hang.
Make sure the GSP files are in a known directory, check file sizes too, make sure the names are correct. Also check the
name and extension. You need to specify the file path like as follows.
4. Reset GSP
105
Procedures
If you want to extract files with a fixed path to a path of your choice use the following command:
pax -r -s![path on tape]!.[wanted path]!gp -t -f [tape device]
106
Procedures
15 Service guard
1. /var/adm/cmcluster
default directory for the cluster files
3. Start cluster
# cmruncl
To run a cluster on a subset of nodes you need to use the –n option of cmruncl. In this case you
have to explicitely confirm that you intent to override Serviceguard’s internal integrity protection.
4. Halt cluster
To halt a high availability cluster use the command cmhaltcl. This causes all nodes in a configured
cluster to stop their cluster daemons, optionally halting all packages or applications in the process.
This command will halt all the daemons on all currently running systems. The -f option causes all
packages to be halted first automatically.
# cmhaltcl –f
5. Start node
Please note that cmrunnode only joins to an existing cluster!
# cmrunnode –n node
6. Halt node
# cmhaltnode –n node
This causes a node to halt its cluster daemon and remove itself from the existing cluster. When
cmhaltnode is run on a node, the cluster daemon is halted and, optionally, all packages that were
running on that node are moved to other nodes if possible.
7. Start package
To run a high availability package use the command cmrunpkg.
# cmrunpkg pkg1
8. Halt package
To halt a high availability package use the command cmhaltpkg.
# cmhaltpkg pkg1
9. Modify a package
# cmmodpkg –e pkg1 Globally enable switching for a package.
# cmmodpkg –d pkg1 Globally disable switching for a package.
# cmmodpkg –n node1 –e pkg1 Enable a node to run a package.
# cmmodpkg –n node1 –d pkg1 Disable a node to run a package.
10. The cmviewconf comand can be used to retrieve the current cluster lock configuration:
# cmviewconf | grep –e "Node name" –e lock
The cluster lock information is backed up and restored with LVM’s vgcfgbackup and vgcfgrestore
commands. If no vgcfgbackup was done after cluster lock initialization then a new cmapplyconf
needs to be done to get this fixed.
107
Procedures
16 Mount a cd
1. ioscan –fn to find device file
2. ls from root to check for SD_CDROM dir
3. mkdir SD_CDROM will create directory if not present
4. mount –r /dev/dsk/cXtXdX /SD_CDROM
5. cd SD_CDROM
6. ls will check files on CDROM
7. umount /dev/dsk/cXtXdX
8. bdf to check to that CD is not mounted
# mv infile /stand/infile
# mv /stand/ioconfig /stand/ioconfig.sav
# mv /etc/ioconfig /etc/ioconfig.sav
# shutdown -ry 0
Recreate ioconfig:
# cd /stand
# /sbin/ioinit –c
9. Once the system reboots, verify that all the instance numbers are correct. It may be necessary to re-import all volume
groups to ensure that /etc/lvmtab contains the correct entries.
108
Procedures
18 Cisco routers
After hooking OMNI book to router then choose icon VT100 SERIAL
User name: cisco
You are now in user mode.
To access privilege mode at the router prompt:
Router> en
Password: san-fran or customer supplied password
Useful commands
109
Procedures
19 DTC16
(underline = flashing)
Normal startup
8. Display being tested
0 Self-test in progress
0 20 sec. To press any key for diagnostics screen via diag. port
0. Download requested
1 Download in progress
1 Download completed OK. Server operational
1. Download completed OK. Server operational and active
Errors
2 Download interrupted. No response from network manager. Auto retry in 30 sec.
2. Download interrupted. Error in data being download. Auto retry in 30 sec.
2. Download interrupted. Error in server. Auto retry in 30 sec.
3 Upload requested, but interrupted. No response from netw. manager. Auto retry in 30 sec
3 Upload in progress
3. Upload requested. Error in protocol. No response from netw manager. Auto retry in 30 s
3. Upload requested. Error in server. No response from netw. manager. Auto retry in 30 sec
6 Reprogramming in progress
6 Fatal error in reprogramming. Server is dead
6. Non-fatal error in programming. You must retry reprogramming from start.
NOTE: Once the machine type has been confirmed, you cannot change it.
If you made an error and selected a wrong machine type, you will have to use a second
exchange DTC16.
110
Procedures
Information
# irdisplay
Configuration
WARNING:
#irdiag –v is used to get configuration information but in case of hard problem on disks if irdiag –v is
executed a 2nd time, configuration can be lost and must be recreated.
KNOW ISSUE
After a complete system shutdown and power OFF the physical drives connected to the RAID CONTROLLER
could be in status 'FAIL' or 'OFFLINE'.
Select the disk who is offline and press enter. And make it online.
111
Procedures
200T, 120T
After replacing the AC controller (on power up it will give a "fail 1") go into, Options→ Config 16→
enter→ when it says "save", hit next until it says "clear"→ enter, power machine off, then back on, then
run test 10 (initial element status). Reenter scsi ID's.
Test Menu
2 Wellness - one loaded cartridge, drive and mailslot empty, one empty slot
5 Init Mech
10 Initialize Element status
11 Mechanism Exercise - One loaded cartridge, drive and mailslot empty, one empty slot
14 Flip Test
15 Storage Slot - one loaded cartridge, one empty slot
16 Drive I/O – One loaded cartridge, drive must be empty
17 Mailslot I/O – One loaded cartridge, mailslot must be empty
25 Empty picker - ** Caution CU will have to perform a reinventory **
43 Mailslot sensor
44 Clear path
112
Procedures
22 ICOD
iCOD Commands
iCOD Install ( you must use the DEC. 2000 or later Support CD )
It's the CU's responsibility to turn the processors on, CE's to turn them off.
113
Procedures
23 Partitioning
vPar Commands
The daemon vpard is started with the script /sbin/init.d/vpard. It will synchronize the master database and
the local disk database for all “up” state vPars,
The daemon vphbd is started with the script /sbin/init.d/vparhb. It provides a heartbeat status, written to
the local disk.
Once the monitor is booted, the database selected (/stand/vpdb is default) is copied into memory and
becomes the master copy.
114
Procedures
On a non-vPars computer, the setboot command allows you to read from and write
to the system-wide stable storage of non-volatile memory.
However, on a vPars computer, the setboot command does not affect the stable storage. Instead, it reads
from and writes to only the partition database.
The mkboot -a command will no longer need to be used to set boot options for the HP-UX kernel. These will
now need to be specified as part of the configuration of the virtual partition. Instead, mkboot -a needs to
be used to change the AUTO LIF file to get /stand/vpmon loaded instead of the regular HP-UX kernel.
The system’s boot command needs to be changed to boot “/stand/vpmon -a”, rather than the normal HP-
UX kernel.
The “mkboot -a” command is used to change the boot command in the LIF AUTO file.
mkboot -a “/stand/vpmon -a” /dev/rdsk/cXtYdZ (the device file of your boot disk)
The “-a” option to vpmon tells it to start all the partitions with the auto boot flag set, normally all virtual
partitions.
115
Procedures
parstatus displays information about the nPartitions or hardware within a server complex
-s Indicate whether the system is an HP server that supports nPartitions. Not supported with the
-g... -h... options.
-w Display the nPartition number for the local nPartition. Not supported with the -g... -h...
options.
-X Display the server complex attributes.
-A Only display the available resources in the complex.
-V Increase the amount of information displayed.
-M Produce output suitable for machine parsing.
-C Show information for all the cells in the complex.
-I Show information for all I/O chassis in the complex.
-B Show information for all cabinets in the complex.
-P Show information for all nPartitions in the complex.
-c cell Show information about the specified cell.
-i IOchassis Show information about the specified I/O chassis.
-b cabinet Show information about the specified cabinet.
-p partition Show information about the specified nPartition.
parcreate This command assigns the specified cells (and any attached I/O chassis) to an nPartition,
assigns a number to the new nPartition, and returns the partition number of the newly
created nPartition.
parmodify You can use the parmodify command to modify the following attributes of an existing
nPartition:
- Partition name
- Cell assignments (add cells or remove cells)
- Attributes of existing cells: Cell type, Use-on-next-boot value, Failure usage, HP-UX B.11.23
and Windows SMS only: Cell local memory (CLM) value
- Core cell choices
- Boot paths (the primary, alternate, and HA alternate paths)
- IP address (if set, this value should be consistent with the IP address assigned to the
nPartition when HP-UX is booted)
parremove This command removes an existing nPartition. This removes all cells from the nPartition and
destroys the nPartition definition.
parunlock The parunlock command unlocks the Stable Complex Configuration Data or Partition
Configuration Data.
fruled The fruled command blinks hardware attention indicators (LEDs) or turns them off.
frupower This command turns on, turns off, or displays the current status of power for cells and I/O
chassis in nPartition servers.
cplxmodify This command modifies attributes of an nPartition-capable server complex, such as the
complex name. Only HP-UX B.11.23 and Windows SMS.
116
Procedures
The following is a summary of the most useful HPUX commands and file locations for the typical HP CE.
117
HPUX
/usr/sbin/diag/contrib/pdcinfo –nobanner - Reads the pim dump info on the system bd.
Normally runs at system startup and sends output to ts99 file.
Important files:
Now change the 700/96 or the webconsole settings to match the emulation you just set.
/usr/bin/X11/xterm –fn 6x13 –sb –sl 10000 –ls –display %IP#% -name %T% &
- Set the DISPLAY variable when running Reflection-X
118
Procedures
2 LVM Commands
119
HPUX
/etc/fstab – Filesystems are mounted at boot time from entries in this file.
120
Procedures
4 Networking Commands
Important files:
5 Software/Patches
Important files:
/var/adm/sw/swagent.log - software installation log file. Check after swinstall for errors!
/var/adm/sw/swremove.log - command log file for swremove
/var/adm/sw/swinstall.log - command log file for swinstall
121
HPUX
6 FTP commands
7 Vi Commands
Adding text in vi
a append new text after the cursor
i insert new text before the cursor
o open a line below the current line and insert
esc Get back to command mode
Deleting text in vi
x Delete the character at the cursor
dd Delete the entire line
Changing Text in vi
rchar Replace the current character with char
Copying in vi
yy Yank the current line
Undo in vi
u Undo the last change
U Undo all changes to the current line
122