| T.R | Title | User | Personal
Name | Date | Lines |
|---|
| 1963.1 | Use 'file'UNIX command to DRD devices | AMCUCS::KKDAS | Let's Train The Trainers | Mon Mar 24 1997 13:30 | 7 |
| If you execute 'file' UNIX command to all relevant DRD devices from Oracle
account, it will detect availability of all devices (A DRD device may not be
available or accessible by Oracle due to one of the following reason: DRD
service is down, underlying disk is not physically accessible, or DRD device is
not owned by Oracle).
/KK/
|
| 1963.2 | | USCTR1::ASCHER | Dave Ascher | Mon Mar 24 1997 14:41 | 14 |
|
I have no idea why this should be necessary but at another
customer site I implemented a script in /sbin/rc3.d that basically
did 'asemgr -d' to check for any 'unassigned' services. It
ALWAYS finds some (I don't understand why) and it simply issues
a 'asemgr -x servicename; asemgr -s servicename' for each service
that was 'unassigned'. When it is all done it invokes the oracle
startup stuff.
It is not at all elegant but it seems to have made them happy
for now.
d
|
| 1963.3 | start/stop scripts up & running | TRN02::ALMONDO | Quid ut UNIX ? | Thu Mar 27 1997 03:20 | 19 |
|
Thanks for the hints,
The solution in .2 is similar to what i have done, the only
difference is that in the start script, if any of the required
service is not available, the script (running in batch mode)
sleeps sometime and then reloop to check the services, the stop
script, checks for the start script sill running and if it is there
then means that the db never started and so it just kill the start
action.
Mario.
PS: i had some problem of unassigned drd services due to network
problems between the cluster members. note that this happen
if (by mistake) durimg the setup of ASE on the TCR the
hostname are specified as members instead of the memory
channel ip network names.
|
| 1963.4 | | USCTR1::ASCHER | Dave Ascher | Thu Mar 27 1997 09:56 | 19 |
| re: <<< Note 1963.3 by TRN02::ALMONDO "Quid ut UNIX ?" >>>
-< start/stop scripts up & running >-
PS: i had some problem of unassigned drd services due to network
problems between the cluster members. note that this happen
if (by mistake) durimg the setup of ASE on the TCR the
hostname are specified as members instead of the memory
channel ip network names.
Hmmmm... in your case was it always the same drd services
came up unassigned? in this case it was random... How do you
fix this problem? do you have to re-install TCR? or can you
modify the members (add mc_host names ? replace host names
with mc_host names?)
ciao,
d
|