TeraByte Unlimited

You need to go through that article and do all the checks.

Different memory is used an different io patters are used. It just confirms
the file is okay so most likely a hardware problem.


"jeffw_00" wrote in message news:[email protected]...

Please understand -this is on a recent model whitebox PC with intel i5
motherboard and Windows 7. It's used to run a financial business, and other
this this issue has no problems and runs flawlessly. Including (generally)
weekly full backups and daily incrementals. It looks like IFW generated a
valid backup it can't process?

On Wed, 23 Dec 2015 11:54:23 PST, just as I was about to take a herb,
jeffw_00 disturbed my reverie and wrote:

>It looks like IFW generated a valid backup it can't process?
Create a new full backup and try to create a differential based on it.
May work.
--
Cheers,

DrT

"If you want to find out what is wrong
with democracy, spend five minutes with
the average voter." - Winston Churchill

Indeed - a week went by, a new full backup was created, and so far, the first differential created off of it worked fine. I understand that with all the chaff out there it's not unreasonable for TU to have a first response of "it's your computer", but in this case where

a) the failure was on a -different- subfile of the full backup every time and
b) the full backup passed multiple validations and
b) replacing the supposed "known good (validated) full backup) eliminated the problem

It's hard not to suspect some subtle IFW issue (even if the problem with IFW is not flagging a subtly corrupted full backup).

anyway - hopefully this information will lead to more improvements in an exemplary application.

best
/j

What still bothers me is that there are so many different numerical error #s that lead to "image Stream corrupt". The different numbers must indicate some sort of information and I hope someday TU will share it with us. It might help us track down subtle problems.

what you saw is completely consistent with memory or hardware issues. The
file and location will typically change. I would suggest you run memtest86+
from memtest.org on your system at least overnight.


"jeffw_00" wrote in message news:[email protected]...

Indeed - a week went by, a new full backup was created, and so far, the
first differential created off of it worked fine. I understand that with
all the chaff out there it's not unreasonable for TU to have a first
response of "it's your computer", but in this case where

a) the failure was on a -different- subfile of the full backup every time
and
b) the full backup passed multiple validations and
b) replacing the supposed "known good (validated) full backup) eliminated
the problem

It's hard not to suspect some subtle IFW issue (even if the problem with IFW
is not flagging a subtly corrupted full backup).

anyway - hopefully this information will lead to more improvements in an
exemplary application.

best
/j

What still bothers me is that there are so many different numerical error #s
that lead to "image Stream corrupt". The different numbers must indicate
some sort of information and I hope someday TU will share it with us. It
might help us track down subtle problems.

12 hours of memest 86+ and no errors -

The three most common things will be:

1) Memory 2) Cable or Port
3) If on RAID drive (the files), the need to run a RAID consistency check
(or verify and fix).

cold power cycle (unplug cord) can also be helpful.

intermittent hardware issues are hard to find.

memory testers can usually find them, but not always, especially depending
on how intermittent it is, when you run a tester it's trying to scan all
memory and blows through an area quickly, when a memory area is being used
for reading or decompressing the files, the same area is used over and over
for the entire time. An hour went by with without any errors, processing
120G, or over 1,800,000 passes. That's why you can consider it a diagnostic
tool as well, when you start seeing validation issues like that, epically
different locations, you have an intermittent hardware problem. If it were
me, I'd change out the RAM, or if I wanted to find a module to ensure it's
RAM and not cable/port, I'd get to where I know the problem shows up, change
each module until it is okay, but if I had RAID on the target, I'd ensure a
RAM test pass 1 level then run the RAID consistency check. I've come
across RAID consistency issues (mirrored) a few times over the years.





"jeffw_00" wrote in message news:[email protected]...

12 hours of memest 86+ and no errors -

No RAID - just a simple desktop PC. Still - it is odd. Backups never were a problem before last weeks' full backup (which still Validated), and this weeks full backup seems to be working fine too. It -could- be intermittent hardware that will fail again in the future, or it could have been something subtly wrong with last week's backup. There's no cable or port involved except for the internal SATA cable, which I guess I could replace if it recurs. One also could suspect the disk controller and thus replace the MOBO, but feels a little early for that, given this has been the only issue....

But you're right - given all the pieces that make up a PC, it's never possible to be sure that IFW (or any SW component) is the cause. I'll keep that in mind before I post a problem in the future :-}

Best!
/j

I'm pretty sure you'll see problems come back... just a matter of time
before it fails more often. The worse is when you start getting system lock
up and random bsod, or issues with files saved. I wish all memory had at
least ecc in todays system, it would save people a lot of issues.

"jeffw_00" wrote in message news:[email protected]...

No RAID - just a simple desktop PC. Still - it is odd. Backups never were
a problem before last weeks' full backup (which still Validated), and this
weeks full backup seems to be working fine too. It -could- be intermittent
hardware that will fail again in the future, or it could have been something
subtly wrong with last week's backup. There's no cable or port involved
except for the internal SATA cable, which I guess I could replace if it
recurs. One also could suspect the disk controller and thus replace the
MOBO, but feels a little early for that, given this has been the only
issue....

But you're right - given all the pieces that make up a PC, it's never
possible to be sure that IFW (or any SW component) is the cause. I'll keep
that in mind before I post a problem in the future :-}

Best!
/j

Well - after 3 days of success, I did get an error creating a differential backup (using a full that still validates), but this time it was error 94 (but still stream corrupt). I really wish I understood what these error codes (80, 94) meant and the difference between them.

Also - you seem to feel strongly that it's memory (as opposed to disk controller, OS, etc). Do you have some foundation for that? If there was some, I could try things like

a) remove one of the 2 memory modules and see if there are any problems (would have to run a couple of weeks, I guess - but the machine has way more memory than it needs), or even

b) swap the memory modules - again the machine has 8GB, and there's typically nothing else running when IFW runs so I would assume it lives in the first 4GB...
but is this a worthwhile experiment?

Thanks!
/j

94 is generic corrupt error, all that matters is what it says in the log.

You can try one at a time, or replace both.



"jeffw_00" wrote in message news:[email protected]...

Well - after 3 days of success, I did get an error creating a differential
backup (using a full that still validates), but this time it was error 94
(but still stream corrupt). I really wish I understood what these error
codes (80, 94) meant and the difference between them.

Also - you seem to feel strongly that it's memory (as opposed to disk
controller, OS, etc). Do you have some foundation for that? If there was
some, I could try things like

a) remove one of the 2 memory modules and see if there are any problems
(would have to run a couple of weeks, I guess - but the machine has way more
memory than it needs), or even

b) swap the memory modules - again the machine has 8GB, and there's
typically nothing else running when IFW runs so I would assume it lives in
the first 4GB...
but is this a worthwhile experiment?

Thanks!
/j