tech.gate.io blog

this might be an issue for clusters, or if you have more systems with the same hostname (which should not be the case, but unfortunately sometimes it is)

 
in your install dir, /tivoli/ITM in my case, most often might be /opt/IBM/tivoli

 
go to the config directory, and change the lz.ini file for linux agents, or the ux.ini file for unix agents, change the value of

CTIRA_HOSTNAME=$RUNNINGHOSTNAME$
CTIRA_SYSTEM_NAME=$RUNNINGHOSTNAME$

 
TO

CTIRA_HOSTNAME=yourhostname
CTIRA_SYSTEM_NAME=yourhostname

 
(maybe it's enough to just change $RUNNINGHOSTNAME$, but I haven't tried this)

stop the agent, remove all offline entries from the teps portal, and restart it

there are several other files with this two variables, but changing these has no effect!

I had to do a ssh display forwarding for a web browser, from a CentOS 5.5 to a solaris 10 system, but I guess this is a general issue on linux systems, or at least on red hat / centos /fedora

I red dozens of forum entries, but no one could solve my issue

ssh_config was ok on client system, sshd_config was configured correctly on server system,

tried with ssh -Y... tried several combinations of ~/.xauthority

but the server never set the DISPLAY variable to localhost

 
after one hour of searching an trying, I figured it out my self: there was a damn package missing, this package only consists of 2 binaries, and one of them is the
/usr/bin/xauth file

from the package

xorg-x11-xauth

yum install xorg-x11-xauth

 

and everything worked fine, thanks to clear error messages, that point you directly to the problem!

I was experiencing the same problem with htc legend and htc wildfire, both on same android version 2.2.1

after getting and accepting the call, it took me about 10sec to actually hear the caller, in this time, the callers just heard it ringing normally as if I was simply not accepting the call

 

the reason was a corrupted fat filesystem on the sd card, probably I just plugged out usb cable without ejecting the device or something like that

I couldn't fix the fs on linux, needed to use windows7

after short filesystem check, I don't have any problems anymore

lspv and other high level lvm commands are reading the pvid from the ODM database, not from the disk itself
this is a great performance improvement, but may lead to an inconsistence of pvids

when the storage admin for example, deletes a hdisk and maps a new one, before you delete it on the system, after running "cfgmgr/cfgdev" the new hdisk appears as the old one with it's old pvid

mapping such an unlucky disk to a lpar, results in different pv ids on lpar and vio server, which may be an absolute disaster

to check the real pvid, read the disk-header directly with the command

od -A n -j 128 -N 8 -x /dev/hdisk120

 
compare this value with the one from lspv
if it's different, use

chdev -l hdiskx -a pv=yes

 
this command will even work, if the disk is already mapped, since the disk already has a pvid, so chdev just reads the pvid from the disk-header, and writes it to the ODM

other than on the hmc, it's not possible to add additional virtual scsi adapters to lpars from the Integrated Virtualization Manger (IVM) GUI

when setting up a Power HA (formerly HACMP) Cluster, it's wise to use different vhosts for different storage subsystems

therefore you need to log in to the vio-server, and add the adapter manually from command line, with the command (as padmin):

$ chhwres -r virtualio --rsubtype scsi -p lparname -o a  -s 6

 
when adding more then one adapter, choose different slots

$ chhwres -r virtualio --rsubtype scsi -p lparname -o a  -s 7

$ chhwres -r virtualio --rsubtype scsi -p lparname -o a  -s 8

 
these changes are persistent over reboots