Another disaster recovery exercise (but not yet)...
- From: Duncan McC <hard@xxxxxxx>
- Date: Thu, 26 Jan 2006 20:18:10 +1300
Hi all,
Some pics taken last night remotely posted at:
http://olympus.servepics.com/sbsprobs/
Yesterday afternoon after 5pm, I updated a client's Server. SBS2000,
all patched etc, regularly rebooted (as many monthly updates require -
indeed, I reboot before updating, and after (anyway)).
ipconfig /all is normal.
The update sounds pretty inocuous - I updates Sophos PureMessage to the
nu version - 2.51e - per Sophos recomendation.
Quite a nightmare follows...
Here's a log of the procedure to-date:
Update PureMessage, OK. Reboot req'd - seemed to shutdown waay too
fast, but no BSOD was noted, nor is any shown in the Event log.
Reboot gone from 10min to 25min. Lots of Services not loading anymore.
Chkdsk c:, d:, e: - it's a hardware RAID1 (2x16Gb=16Gb) + RAID10 (4x32Gb
= about 70Gb). So c: on the mirror, and a fast d: for data (exchange,
SQL Server). Acutally there's a wee e: drive on the fast array for
"me" - so 10Gb out of that total for d:. I store updates and "my
stuff" etc on the e: drive.
Chkdsk is all good on all drives. Still very slow boot time and many
Services failing to start (none are set to disabled, they are set to
Auto or Man - but just fail to start (sometimes see Start Pending) and
go to Stopped.
Reboots to Safe Mode and Dir Serv Rest. Mode - both take two reboots to
make it happen (it reboots 1st time round by itself - no reason noted).
I removed PureMessage - but it failed at 52%.
I phoned Sophos, we stepped through removing all Sophos everything,
including registry (CurrentControlSet to stop Services) - and all files
on the PC w' Sophos in 'em (folders etc).
Enabling boot logging option at boot - well, doesn't seem to write one.
Not found one yet.
I feel like just a file or two is/are corrupted. And if I could suss
them, I would be up and running again. The alternative is to rebuild
the server - I guess from a format (the Server is here at home with me
now - and my client (20 user law firm) have been down for the day.
(it's going to be the start of a long night :)
Event ID 2114 was one event I looked up on eventid.net and I added
LanmanServer to HKLM/System/CurrentControlSet/Services/Netlogon
(DependOnService).
Other things I've tried:
logging on from a wk.station - amazing (though it takes 5 min to logon
instead of seconds) - the Iternet is available for me (and everyone
else).
Services that aren't starting:
Exchange Services - all of them I think.
(Veritas) Backup Exec Services (most of them)
Netlogon Service (shouldn't need it on the Server ???)
No shares are available from any wk.station (they're there in the Server
though).
ICW doesn't run.
In Network and Dialup Connections: the NIC's - each of them, if I click
into the TCP/IP settings, the Properties is greyed out (double-clicking
doesn't work either).
I have disabled the NICs and rebooted. (They won't remove). I started
in Safe Mode, removed them, and reinstalled them. They look good in
Safe Mode, but on a normal boot up (still as slow, Services still
failing etc) - they show as disabled. Properties of them show drivers
failed to load (ok, I've done that (it *was* working)).
I've put SP4 (partially installed then stopped) and SP1 Rollup
(installed OK) ontop. No help.
I'm about to clean install I think. I'm going to ensure (probalby are
already) that Exch. and SQL S are stopped and will copy the latest data
to the e: drive (databases and store). And get the latest docs that are
"linked" via the 3rd party SQL Server app and copy those accross.
I can get V. Backup Exec to load it's services and stick - in normal
mode only. My attempt to restore the registry/active directory -
failed, due being a domain controller, it needs booting up in Directory
Services Restore Mode. I can get there (2nd boot) - but Backup Exec
services *won't* load and stick in that mode. Dependent services won't
start.
Lucky I have short hair already.
PS: no MS free support for us - non-USA. :( I have an MSDN, but - it's
now past 5pm local time and can't get a tech support phone number to
phone (local number I have says to call back during work hours).
Perhaps this is a silly RPC corruption - and I just need to fix a
corrupt file or two or three. It seems to me to be not much more than
that - and then everything will work. However, if I can't get anywhere
soon, I'll have to begin a clean install I fear.
Any help and advice from anyone that's been here before, much
appreciated.
Some pics taken last night remotely posted at:
http://olympus.servepics.com/sbsprobs/
--
Duncan
.
- Follow-Ups:
- Re: Another disaster recovery exercise (but not yet)...
- From: Jeff Middleton [SBS-MVP]
- Re: Another disaster recovery exercise (but not yet)...
- Prev by Date: Re: Allowing 'Port' access on Client Machines
- Next by Date: Re: Allowing 'Port' access on Client Machines
- Previous by thread: SBS2000 overnight lockup, poss Trend related
- Next by thread: Re: Another disaster recovery exercise (but not yet)...
- Index(es):
Relevant Pages
|