Re: /LGICMD suggestion
- From: John Santos <john@xxxxxxx>
- Date: Fri, 15 Dec 2006 06:41:58 GMT
Stephen Hoffman wrote:
Some suggestions...
If the disk is offline and is disrupting the login process, then there are second-level problems likely to arise all over the place. Applications and tools and images and most anything that references the (failed) device will itself fail. I'd tend to look at why the particular disk dropped offline, and at what can be done to reduce the exposure to that -- the failed and inaccessable login is probably the tip of a very large iceberg of problems that can arise here, and dealing with those errors is going to be interesting.
If it's the SYSTEM username, then that's the system disk and there are other problems with that disk being offline.
If it's another username that's involved with this (failed) login, then /NOCOMMAND (or the console, for access to the SYSTEM username or the SYSUAF is locked up or otherwise corrupt) is the way in.
If you need remote emergency access, then a terminal server (or as they are often now known, as a "console management processor") or a management processor is the way in. If the box has the option, as do various of the Integrity servers, the manage ment processor can be very useful for these cases.
I really hate it when I get an error message says "ask your system manager
for assistance" and I am the system manager and have no clue :-(
I think the OP is trying to log in to fix the myriad of problems or
at least determine which drive has broken so he can tell the on site
personnel what drive to have replaced and which backup tape to load
on the tape drive while he drives in to work (at 3 in the morning.)
Not some random Joe User trying to log in to a broken system, he's
the system manager trying to get a head start on fixing things.
Certainly he could get system access via the console (opa0:) remotely
via a null-modem to a terminal server or a console management processor,
but if he can't log in, he's still stuck (short of ctrl/P; HALT; reboot,
which he might *not* want to do until he understands exactly what is
wrong. For example, the disk with the user accounts, development
sources, etc. might have crashed, but the server application itself
is still fine, so there may be no need to reboot, just replace the
drive and restore the backup.
I think the "search-list" logical name for the login command file
might do the trick. This would probably not work for a captive
account, but might work for a restricted account. (Some companies'
security policies say all users must be in restricted accounts,
which prevents bypassing login.com.) But maybe it would be
happy if it finds *any* login.com, not just the one in the first
device/directory in the search list.
Also, be sure that you don't have an "on error then exit" (which
is the default) in your login.com (and maybe in sys$sylogin as well)
so you don't get kicked out prematurely. Either "set noon" or
appropriate "on error" commands to deal with whatever happens.
--
John Santos
Evans Griffiths & Hart, Inc.
781-861-0670 ext 539
.
- Follow-Ups:
- Re: /LGICMD suggestion
- From: Bob Koehler
- Re: /LGICMD suggestion
- References:
- /LGICMD suggestion
- From: Tom Linden
- Re: /LGICMD suggestion
- From: Richard B. Gilbert
- Re: /LGICMD suggestion
- From: Tom Linden
- Re: /LGICMD suggestion
- From: David B Sneddon
- Re: /LGICMD suggestion
- From: Tom Linden
- Re: /LGICMD suggestion
- From: David B Sneddon
- Re: /LGICMD suggestion
- From: Tom Linden
- Re: /LGICMD suggestion
- From: Stephen Hoffman
- /LGICMD suggestion
- Prev by Date: Re: Mentec US is gone!
- Next by Date: Re: Mentec US is gone!
- Previous by thread: Re: /LGICMD suggestion
- Next by thread: Re: /LGICMD suggestion
- Index(es):
Relevant Pages
|
