Update Popup Goes Rogue

[Cycloid Torus]

Update popup for 701 failed on machine 1 (WinXP home SP2) after I said "ok"
to install. Install process could not complete. gcaSWUpdater.exe hogged 100%
of CPU and stalled. I had to 3-finger and "shutdown".

Machine 2 jumped up with same popup. I answered "NO" and it jumped ahead and
started to install (disregarding my response). gcaSWUpdater.exe hogged 100%
of CPU and stalled. This time I made printscreens in case any one wants
them. Had to "shutdown" to escape.

Downloaded MSAS 701 and it installed with not hitch on machine 1.

Bad install routine?

Bad installer?

 

[Cycloid Torus]

Download and update to machine 2 was flawless.

Bad installer?

 

[Bill Sanderson]

I don't know. I don't think it is TOO bad, 'cause I don't see hundreds of
such reports. Neither the installer nor the bits is much changed, so there
shouldn't be anything new or different popping up at this point.

--

 

[Cycloid Torus]

No, not too bad, but still "rogue". Is there any way of tesing InstallShield
for integrity and completeness? Having 2 machines go into "adrenal shock"
makes me wonder.

The downloads both installed properly - so it isn't the native program, but
something in the packaging - and that is NEW.

 

[Bill Sanderson]

I did have one first-hand experience with a machine where a user reported
seeing a dialog box, applying an update, and then having the machine crash
on restart. I ASSumed that the update with the Microsoft Antispyware
version change, but the user was told by another staff member to use System
Restore to a restore point before the incident (after having restarted with
last known good)--and did so. When I checked the Microsoft Antispyware
version in place it was .701.

So--I can't pin that blue-screen crash on Microsoft Antispyware. I think I
didn't check the event logs--I'll do that tonight or tomorrow and see if the
crash was recorded.

The installer routines have never been the strongest part of the product,
judging by the number of issues we've had reported. But crashes would
definitely be something new.

--

 

[Cycloid Torus]

It wasn't actually a bluescreen. gcaSWUpdater.exe hogged 100% of CPU, was
actually not doing anything except rapidly twiddling thumbs (CPU temp hit
44C an all time high) and was unresponsive.

I had similar fugue state with Intellishield in another install a couple of
weeks ago. I thought I had caught it by disabling shadow BIOS/Video caching
and trimming frame size. I had thought it was a memory poke which entered
previously allocated memory. The 2 machines are similar. AMD Athlons, XP
home SP2 (from same disk provided from MSFT), IGP motherboards. I have a
similar problem with MailWasherPro when the heuristic analysis file gets
large.

What gets me is that these machines are sufficiently different except for
operating system and a shared UPS that the fugue state should be appearing
in other machines elsewhere, but apparently not.

The machines have not been reformatted in 2 years and maybe a fresh start is
called for at this time.

Oddity in errors.log "7::ln 10:Out of
memory::gcasDtServ:ScheduleScans:Update::12/1/2005 10:50:29 PM:XP:1.0.701" -
yet I saw it run, I saw it finish, I clicked it closed.

I may have several active accounts open (some Admin some LUA), but only the
one Admin account last night.

I think I am going to do a Safe Mode run of the anti-parisiticals.

 

[Cycloid Torus]

More Oddities
The subsequent downloads directly from Microsoft had a file creation date of
12/1/2005 and time based on machine clock at time of download as you would
expect. Size seemed exactly the same.

The creation dates for the file automatically downloaded were back in July
AND WERE DIFFERENT!!

Friday, July 22, 2005, 4:38:56 PM
and
Wednesday, July 20, 2005, 10:42:20 AM

Check your Microsoft Antispyware folder for the TempUpdates folder (if it
didn't vaporize after doing its job - mine crashed so it was still there).

Anyone have a good binary "file comp" utility? (I wrote my last one in 1987
and don't think I want to again.)

 

[Bill Sanderson]

As far as I can tell, trying to make sense out of the log entries is like
reading tea leaves......
I visited my own blue-screen crash machine yesterday, and I suspect its a
hardware issue--crash in an IDE-related driver at boot. It isn't solid,
though--happens at boot, repeat the boot and the machine starts and runs
fine all day.

In 18,000,000 machines, there are going to be a certain number of crashes on
reboot.

It hasn't been uncommon to see complaints about 100% cpu usage during
definition updates. One time that a definition update may occur is on
startup of a new version--there's apparently an automatic no UI update
check.

--

 

[Bill Sanderson]

I'm not sure what files you are talking about, but drop to a command prompt,
put both files in the same folder and rename them with short (8.3)
filenames, and do

fc a b <enter> to compare them. I can never figure out how to interpret the
result when files are different, but it is easy to tell when they are the
same!

--