Ground Hog Day - failed system image restoring yet again

[nudone]

Not really sure how many times this has happened, maybe it's just the once and I'm reliving it Ground Hog Day style, over and over - until I get it right.

Here's the scenario:

1) New, one month, old machine.

2) Bluescreens a few times a week - only pattern I can see is that Gmail or Google Reader are open in Firefox 8 sometimes.

3) Made several system backups during the month using Paragon 2011.

4) Not tested any of the backups - didn't think I needed to, it's a NEW system, why would it not work?

5) Today, restarted after bluescreen and firefox has lost all the settings I've carefully done. Putting the setttings back will take longer than I want... so, I decide it's time to use the most recent system backup and quickly get everything working again.

6) Paragon completely fails and decides to destroy the drives boot manager files instead.

7) I waste an hour trying to get Paragon to work, then have a brainwave and realise that the Windows boot disk will fix the boot manager problem - assuming Paragon has been so pathetic it hasn't even removed my system files.

8 ) Windows does fix the boot manager and the machine boots into the last known good system. Phew.

9) I type this post whilst waiting for Windows built-in backup to make a system image...




So, here we are again. Another thread about why you should check your backup/recovery process. I feel like I've been very lucky as Paragon could have really mucked things up for me. Now, I've got to assume that Windows backup will work when I need it to - the problem being that I can't afford to test it right now; if it fails then I've got a week's worth of installing and setting up to do.

It seems that a golden rule ought to be that backup/recovery should be tested at the earliest point possible, i.e. immediately after installing windows. If the process fails, then you've only wasted an hour (or less) rather than an entire week, or month, of reinstalling everything and configuring to your usual requirements.

Does this mean Ground Hog Backup Day is finally over for me? I really hope so. I'm sick of my own incompetence with all this - it's almost like a compulsion to fail. It's just sick. If I ever write a post like this again, I want Carol Haynes to come here and confiscate all my computer equipment (Carol, I'll gladly pay you for the trouble).

[Ath]

So, despite the 24 hour (or was it 48, forgot that detail) burn-in test done, the system is not stable.

the problem being that I can't afford to test it right now; if it fails then I've got a week's worth of installing and setting up to do.

-nudone (December 13, 2011, 02:13 PM)

You just have to set up that machine right. And it might involve a full re-install of Windows.
I know you said you updated all and every driver in the system, but that doesn't mean Windows isn't corrupted in some way. I've seen that happen from Windows 3.11 MS-DOS 2.x to Windows 7, and all versions in between. Just starting over with a format c: /s and restarting the tedious job of re-installing Windows and all your required software will eventually fix the issue. I understand you are not (yet) ready to go there.

About 15 % chance the memory (probably just 1 module) in the system is the cause of trouble (even after a burn-in test!)
About 10 % chance the video adapter/drivers is the cause of the trouble
About 5 % chance the hard-drive (boot/swap) is the cause of the trouble
About 0.01 % chance the PS/2 keyboard is the cause of trouble (saw that about twice in about 2000+ PC's in all my years building of PC's)
The rest (~70 %!) is caused by Windows not installed/decompressed right.

(Percentages are estimates)

Just my  

[nudone]

I admire your determination to keep telling me to check the machine out. I think I will, but not until February at the earliest.

I used to find it all a challenge and I would be determined to resolve the problem, even if it took a few days. Now, I just find it very tiresome and annoying and would rather just pay for a computer that worked - but that's exactly what I've meant to have done with this new machine, I paid with the (unrealistic) assumption that it would work.

It's even more annoying as I know I could plug my six year old machine back in and everything would just work - slower, I admit, but I wonder if that really matters now. What's the point of speed if it's not stable.

Anyway, I could moan about it all day and think of plenty of reasons why someone should just wave a magic wand for me and make everything work. I'll just have to prepare for the inevitable week(s) of reinstalling and testing.

[IainB]

4) Not tested any of the backups - didn't think I needed to, it's a NEW system, why would it not work?
+
+
Now, I've got to assume that Windows backup will work when I need it to - the problem being that I can't afford to test it right now; if it fails then I've got a week's worth of installing and setting up to do.
+
+
It seems that a golden rule ought to be that backup/recovery should be tested at the earliest point possible, i.e. immediately after installing windows.
+
+
I'm sick of my own incompetence with all this - it's almost like a compulsion to fail. It's just sick.

-nudone (December 13, 2011, 02:13 PM)

I almost could not believe that I was reading this.
For many years, IT Service Management SOPs (Standard Operating Procedures) have included these mandatory steps as part of prudent risk management and mitigation:

• Avoid risks that you can anticipate.
• Prior to installation, take a backup. Test the backup.
• After installation, take a backup. Test the backup.

This was drummed into our heads in training in mainframe days, and it is as valid today as it was then.
If this was not drummed into your head before, it presumably could have been as a result of this hard-earned lesson from trial-and-error.

And I would suggest that you are quite wrong where you say:

I'm sick of my own incompetence with all this

This does not necessarily meet the criteria for incompetence. Incompetence is knowing how to do a job, but failing to do it anyway.
This case looks more like either ignorance or a lack of adequate training (or both) is what got you here.

If that is the case, then it is a management failure (and grossly unfair) to put someone into a a job - a position of responsibility for something - without ensuring that they are adequately equipped to properly execute that job/responsibility to the best of their ability. - paraphrased from W.Edwards Deming.

If you had known better, but still took the risk anyway, then that would probably be a different matter.

Good luck with the recovery.

[Ath]

I admire your determination to keep telling me to check the machine out. I think I will, but not until February at the earliest.

-nudone (December 13, 2011, 03:14 PM)

It's just that I know how you feel. I've been there too, several times, with several types of equipment, and the only way to get things back on track...

Being an experienced optimist , this aparatus started out wrong, and it'll be really hard to get it right.

I'm not trying to push you anywhere, just expressing my sympathy.

[nudone]

This does not necessarily meet the criteria for incompetence. Incompetence is knowing how to do a job, but failing to do it anyway.
This case looks more like either ignorance or a lack of adequate training (or both) is what got you here.

If that is the case, then it is a management failure (and grossly unfair) to put someone into a a job - a position of responsibility for something - without ensuring that they are adequately equipped to properly execute that job/responsibility to the best of their ability. - paraphrased from W.Edwards Deming.

If you had known better, but still took the risk anyway, then that would probably be a different matter.

Good luck with the recovery.

-IainB (December 13, 2011, 03:20 PM)

Oh, I do know better. I've been using system backups since Ghost was made available to me (before Norton picked it up) when it took away the pain of reinstalling Windows off of floppy disks. So, I've simply become lazy and happy to pretend that I don't have time to test backups - because they've GOT to work, they can't fail, I haven't time for them to fail, that's just not an option.

So, like I say, I'm incompetent. I ought to hand in my resignation and start basket weaving for a job or something. But I'm self-employed so I'm stuck with my own IT Dept. run by a total arrogant sod who believes testing is for losers.

edit:
Just thought I'd better add. Now I'm not even sure if did or did not test the first backup I made for this new system. I'm beginning to think I did test it as I've had issues with Paragon before. So, if I DID test it, that means I can't even trust it when it works. Which then implies that one method of backup isn't enough. Looks like I've just got to start taking matters seriously if I want things to remain reliable. It's all a bit of a game at the moment; gambling on whether the computer will work or not. And I can't afford that - so I'm being more stupid than I first thought.

[Stoic Joker]

It might help if you mentioned what code the BSOD contained.

The behavior does sound vaguely (I emphasize vague due to a lack of detail ) familiar. I had a batch of new Dells that exhibited much the same behavior, until I uninstalled the ATI Catalyst Control Center. No more BSODS after 2 month of 24/7 running after removal of the CCC.

Just a thought - Shit drove me (and the client) nutz for 3 months.

[nudone]

Adding a bit more, for what it's worth...

...looking around the web, people are reporting that the OCZ ssd likes to bluescreen and freeze - which is exactly what this new machine does, and yep, I have a OCZ ssd. This is what I suspected but didn't want to believe (I had to return the previous "new" machine because the ssd in that one died after 5 days).

If I do anything with all this testing rigmarole, I'll first just rip out the ssd and stick a proper drive in its place. In fact, I think I'll just order a small drive tomorrow and try putting the windows backup onto it.

[nudone]

It might help if you mentioned what code the BSOD contained.

-Stoic Joker (December 13, 2011, 04:36 PM)

I'll start paying more attention to what it is and write it down - but, as I've just posted before I saw your post above, the machine also likes to freeze so no bluescreen codes there.

One pattern (I think) to these problems is that they tend to happen in the afternoon - perhaps when the ssd is getting too hot???

[nudone]

Okay, having read about even more people with OCZ vertex 3 ssd problems it now seems very, very likely that this is the problem with my machine. And it also sounds like OCZ haven't resolved the problem regardless of the firmware updates they released.

Looks like I can resolve this problem very quickly by hammering a 6" nail through the ssd.

[Stoic Joker]

Looks like I can resolve this problem very quickly by hammering a 6" nail through the ssd.

-nudone (December 13, 2011, 04:59 PM)

Ah...the value of catharsis...

[tomos]

^

re testing an image - do you (people) mean to do a test restore to the actual windows partition that was backed up?

[nudone]

^

re testing an image - do you (people) mean to do a test restore to the actual windows partition that was backed up?

-tomos (December 14, 2011, 02:01 AM)

Yes. You need to know that your backup/restore routine is going to work 100%. Otherwise it isn't a backup method at all.

The only real way of knowing if it works is to go through the whole restore process. So, it would make a lot of sense to do that at the earliest point possible; no point installing software and setting your computer up and THEN making a backup - if the backup process doesn't even work.

Now, if you know your backup/restore process works AND then find one day that it doesn't, it's all a bit too late. In which case it would have been a good idea to have a secondary backup/restore process to fall back on. Maybe that scenario hardly ever happens - I honestly can't remember if I tested my system at the beginning or not.

[4wd]

About 15 % chance the memory (probably just 1 module) in the system is the cause of trouble (even after a burn-in test!)
About 10 % chance the video adapter/drivers is the cause of the trouble
About 5 % chance the hard-drive (boot/swap) is the cause of the trouble
About 0.01 % chance the PS/2 keyboard is the cause of trouble (saw that about twice in about 2000+ PC's in all my years building of PC's)
The rest (~70 %!) is caused by Windows not installed/decompressed right.

(Percentages are estimates)

-Ath (December 13, 2011, 02:43 PM)

You forgot that which is allocated to PEBKAC.

[Ath]

You forgot that which is allocated to PEBKAC.

-4wd (December 14, 2011, 03:59 AM)

I hid that in the 70% "not installed right" part