horcm not working

[Erik_Smith]

Setting up a new system with BCV, and having some troubles getting everything to work properly. Probably forgetting something simple.

tail /etc/services
horcm0 5000/udp #XP1024 Disk Array
horcm1 5001/udp #Xp1024 Disk Array

# cat /etc/horcm0.conf
HORCM_MON
#ip_address service poll(x10ms) timeout(x10ms)
dcuxphr1 horcm0 1000 3000

HORCM_CMD
#dev_name dev_name dev_name dev_name
/dev/rdsk/c1t0d5 /dev/rdsk/c3t0d5
#two paths to the same device

#HORCM_DEV
#dcuxphr1 manages P-Vols via CL1-E and S-Vols via CL2-E
#dev_group dev_name port# TargetID LU# MU#
# Use following cmd to scan for port#, TargetID, and LU#
# vgdisplay -v /dev/vgsan|grep dsk|sed 's/\/*\/dsk\//\/rdsk\//g'|raidscan -find

#HORCM_INST
#dev_group ip_address service
hrmsdb dcuxphr1 horcm1

# cat /etc/horcm1.conf
HORCM_MON
#ip_address service poll(x10ms) timeout(x10ms)
dcuxphr1 horcm1 1000 3000

HORCM_CMD
#dev_name dev_name dev_name dev_name
/dev/rdsk/c1t0d5 /dev/rdsk/c3t0d5
#two paths to the same device

#HORCM_DEV
#HPUX57 manages P-Vols via CL1-E and S-Vols via CL2-E
#dev_group dev_name port# TargetID LU# MU#
# Use following cmd to scan for port#, TargetID, and LU#
# vgdisplay -v /dev/vgsan|grep dsk|sed 's/\/*\/dsk\//\/rdsk\//g'|raidscan -find

#HORCM_INST
#dev_group ip_address service
hrmsdb dcuxphr1 horcm1

disk 11 0/0/14/1/0/4/0.5.15.0.0.0.5 sdisk CLAIMED DEVICE HP OPEN-3-CVS-CM
/dev/dsk/c1t0d5 /dev/rdsk/c1t0d5

disk 14 1/0/14/1/0/4/0.6.15.0.0.0.5 sdisk CLAIMED DEVICE HP OPEN-3-CVS-CM
/dev/dsk/c3t0d5 /dev/rdsk/c3t0d5


# /sbin/init.d/rmxp start
starting HORCM inst 0
HORCM inst 0 has failed to start.
EXIT CODE: 3

[Nigel Poulton]

Exit code 3 is not much to go on. Do you have anything from the HORCM 0 log directory?

Also, I notice that in your HORCM 0 config file you define a group (hrmsdb) in the HORCM_INST section but there is no associated group (hrmsdb) defined in the HORCM_DEV section.

With no groups defined in the HORCM_DEV section you not be able to perform any HORCM operations, you will only be able to start the instance.

If you just want to start the instance you could try commenting out the last line in HORCM_INST

[Erik_Smith]

Thanks for the quick response. Indeed, I forgot to comment out the last line, as I don't have the group info yet. So with the last line commented out and the process started. I should be able to query the manager for the raidscan.

# /sbin/vgdisplay -v /dev/vg57san|grep dsk|sed 's/\/*\/dsk\//\/rdsk\//g'|raidscan -find
raidscan: [EX_ATTHOR] Can't be attached to HORC manager
Refer to the command log(/HORCM/log/horcc_dcuxphr1.log) for details.

09:53:56-d4970-02089- 2089:HORCM death detected
09:54:00-0d592-02089- ERROR:cm_open[scmclcon() timeout_err]
09:54:00-0e078-02089- ERROR:horcm_lep_create
09:54:00-104f2-02089- [raidscan][exit(251)]
[EX_ATTHOR] Can't be attached to HORC manager
[Cause ]:Couldn't connect with the HORC manager.
[Action]:Please check if HORC manager is running or if HORCMINST is set correctly.

[Nigel Poulton]

Are you certain that your HORCM instance is running?

If so, have you configured the correct HORCM environment variables? these are -

HORCMINST=0 (if you are issuing commands to HORCM isntance 0)

HORCC_MRCF=1 (if you want to issue commands to the BCV command environment rather than Continuous Access).

These variables must be available to the shell you run your HORCM commands from.

[Erik_Smith]

That was it! I had those set in my .profile, but when I check the ENV, they weren't set, so I reexported, and now get a response.


Thanks, I knew I was overlooking something simple.

[Erik_Smith]

Trying another server now. same exit code 3

tail /etc/services
horcm0 5000/udp #XP1024 Disk Array
horcm1 5001/udp #XP1024 Disk Array

disk 24 0/0/12/1/0/4/0.5.135.0.0.0.2 sdisk CLAIMED DEVICE HP OPEN-3-CVS-CM
/dev/dsk/c9t0d2 /dev/rdsk/c9t0d2



disk 30 1/0/12/1/0/4/0.6.135.0.0.0.2 sdisk CLAIMED DEVICE HP OPEN-3-CVS-CM
/dev/dsk/c13t0d2 /dev/rdsk/c13t0d2


/etc/horcm1.conf
HORCM_MON
#ip_address service poll(x10ms) timeout(x10ms)
dcuxphr2 horcm0 1000 3000

HORCM_CMD
#dev_name dev_name dev_name dev_name
/dev/dsk/c9t0d2 /dev/dsk/c13t0d2
#two paths to the same device

#HORCM_DEV
#dcuxphr1 manages P-Vols via CL1-E and S-Vols via CL2-E
#dev_group dev_name port# TargetID LU# MU#
# Use following cmd to scan for port#, TargetID, and LU#
# vgdisplay -v /dev/vgsan|grep dsk|sed 's/\/*\/dsk\//\/rdsk\//g'|raidscan -find

#HORCM_INST
#dev_group ip_address service
#hrmsdb dcuxphr1 horcm1


/etc/horcm1.conf
HORCM_MON
#ip_address service poll(x10ms) timeout(x10ms)
dcuxphr2 horcm1 1000 3000

HORCM_CMD
#dev_name dev_name dev_name dev_name
/dev/dsk/c9t0d2 /dev/dsk/c13t0d2
#two paths to the same device

#HORCM_DEV
#HPUX57 manages P-Vols via CL1-E and S-Vols via CL2-E
#dev_group dev_name port# TargetID LU# MU#
# Use following cmd to scan for port#, TargetID, and LU#
# vgdisplay -v /dev/vgsan|grep dsk|sed 's/\/*\/dsk\//\/rdsk\//g'|raidscan -find

#HORCM_INST
#dev_group ip_address service
#hrmsdb dcuxphr1 horcm1



# HORCC_MRCF=1
# HORCMINST=0
# export HORCC_MRCF HORCMINST

# /sbin/init.d/rmxp start
starting HORCM inst 0
HORCM inst 0 has failed to start.
EXIT CODE: 3
#

Thanks again for any help

[Nigel Poulton]

Hi Erik,

Since you have nothing defined in HORCM_DEV and HORCM_INST there isnt much that can be wrong. Are you certain that your CMD devices are correctly defined and are visible on the device addresses specified in the HORCM_CMD section (I see your output but its worth checking again)?

Im unsure for UNIX but for Windows you can find the CMD device with the following -

raidscan -x findcmddev 0,100

(might be different for UNIX but Im not in the office now so cant check the doc for you).

Also are you certain that there are no other services already bound to and listening in the UDP ports defined in the HORCM_MON section. Ive seen on Windows where a bug in the MS DNS server hogged hundreds of UDP ports in the higher ranges. May be something else is squatting on that UDP port?

Also are the hostnames defined in HORCM_MON and HORCM_INST resolvable?

Oh and last but not least. There must be a carriage return after the last line of the HORCM_INST section. The parser will fail to read the HORCM file if there is not at least one blank line at the end of the file.

Hope this helps