Page 1 of 2

IfW - Gathering information

Posted: Tue Apr 18, 2017 1:54 am
by DrTeeth
Hi,

When the v3 line first came out, we noticed that the backup started
much more quickly than the v2 line using PHYLock.

Since 3.06a was released, it takes a long time for the backup to
start. Even the Image for DOS in BIBM took longer to start than the
blink of an eye it used to take. I had to resort to using IfDOS as I
could not get the backup to start using IfW earlier today.

I think I know what the issue is.

I have set both the hash file options to be used. The problem is that
the backup does not start as PHYLock is "gathering information". Image
for DOS took over a minute to start instead of a second or so in the
past. I could not get IfW to start in spite of waiting for over 20
minutes.

This "gathering information" stage used to be quite fast. I do have a
few more tests to run, but something is definitely not right. Just to
clarify, it is not IfW waiting for the disks to stop.

Sorry if this is not clear as it is early morning (02:53 hrs).
--
Cheers,

DrT

** My best moments are when I am asleep.
** The nightmare starts when I wake up.

Re: IfW - Gathering information

Posted: Tue Apr 18, 2017 2:05 am
by DrTeeth
Sorry to reply to my own post, but I just started an IfW backup. I
deselected the new metadata hash option and the backup started in
about 5-10 seconds like it used to.

I just need to try a backup with the item selected again and see if it
takes a long time to get started. The change just now is, to be
honest, conclusive...a very dramatic and fast start.
--
Cheers,

DrT

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

Re: IfW - Gathering information

Posted: Tue Apr 18, 2017 2:40 am
by mjnelson99
I set Phylock to 5 min.
Sometimes it has helped if I restart an image.
Not sure that would work now, though.
Mary

On Mon, Apr 17, 2017 8:54 PM, DrTeeth wrote:
> Hi,
>
> When the v3 line first came out, we noticed that the backup started
> much more quickly than the v2 line using PHYLock.
>
> Since 3.06a was released, it takes a long time for the backup to
> start. Even the Image for DOS in BIBM took longer to start than the
> blink of an eye it used to take. I had to resort to using IfDOS as I
> could not get the backup to start using IfW earlier today.
>
> I think I know what the issue is.
>
> I have set both the hash file options to be used. The problem is that
> the backup does not start as PHYLock is "gathering information". Image
> for DOS took over a minute to start instead of a second or so in the
> past. I could not get IfW to start in spite of waiting for over 20
> minutes.
>
> This "gathering information" stage used to be quite fast. I do have a
> few more tests to run, but something is definitely not right. Just to
> clarify, it is not IfW waiting for the disks to stop.
>
> Sorry if this is not clear as it is early morning (02:53 hrs).
>

Re: IfW - Gathering information

Posted: Tue Apr 18, 2017 8:29 am
by Brian K
DrT,

Just some background. How much data is in the partition you are imaging? Is it a Full or a Changes Only backup?

Re: IfW - Gathering information

Posted: Tue Apr 18, 2017 10:12 am
by DrTeeth
On Tue, 18 Apr 2017 01:29:31 PDT, just as I was about to take a herb,
Brian K disturbed my reverie and wrote:

>Just some background. How much data is in the partition you are imaging? Is it a Full or a Changes Only backup?

Hi Brian and Mary,

Mary, it is not a PHYLock issue at all.

Brian, I am imaging a complete hard drive with two versions of windows
and 7 Linux distros. The backups are changes only. The disk is 2TB and
is half full.

I have confirmed this behaviour on another PC. Setting the "Use
Metadata Hash Files" GREATLY extends the Gathering Information stage.
This was well over 20 minutes the other day and Image for DOS is also
affected, but not as much. Mary, PHYLock is currently operational at
this stage and PHYLOCK.SWP is being generated so changing the PHYLock
delay would have no effect.

With this option NOT selected, the backup starts in a blink of an eye.

--
Cheers,

DrT

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

Re: IfW - Gathering information

Posted: Tue Apr 18, 2017 1:01 pm
by Logman
Maybe a tad off topic but I only use IFL or IFD as I found doing it with IFW is too time consuming.

Re: IfW - Gathering information

Posted: Tue Apr 18, 2017 9:55 pm
by Brian K
DrT,

I don't have your issue with IFW 3.06a. I ran a test on a SSD with 25 partitions and about 150 GB of data.

Using /usemd to create a differential backup...

start 0:00
PHYLock 0:05
end gathering 1:15
end /b 3:04
end /vb 3:15


Not using /usemd to create a differential backup...

start 0:00
PHYLock 0:05
end gathering 0:06
end /b 8:45
end /vb 8:56

So the backup phase using /usemd took 1:49 and the backup phase not using /usemd took 8:39. Gathering with /usemd took 1:10.


I did the same differential test with IFL and /usemd...

start 0:00
end gathering 0:04
end /b 1:10
end /vb 1:18

IFL without using /usemd

start 0:00
end /b 5:11
end /vb 5:19


I did the same differential test with IFD and /usemd...

start 0:00
end gathering 1:28
end /b 3:39
end /vb 4:00

IFD without using /usemd

start 0:00
end /b 9:22
end /vb 9:43

Re: IfW - Gathering information

Posted: Wed Apr 19, 2017 10:56 am
by DrTeeth
On Tue, 18 Apr 2017 14:55:35 PDT, just as I was about to take a herb,
Brian K disturbed my reverie and wrote:

>I don't have your issue with IFW 3.06a. I ran a test on a SSD with 25 partitions and about 150 GB of data.

The problem is related to not only the amount of data being backed up,
but to the speed of the drive. I am backing up 1TB of data on a SATA
2. I am not surprised that you are not seeing an issue - it is a bit
like saying you don't have a problem with your Ferrari going over 100
mph so you cannot understand why my Trabant cannot go faster than 60
(LOLZ) :-).

I have tested this issue on three setups and selecting metadata hash
*always* results in more of a delay than not using it. Deselecting
this option removes the delay. My main PC has the largest amount of
data to be backed up - the backup itself takes 1.5 - 2 hours. With
metadata hash selected, the longest I have waited for the backup to
start was 20 mins - the backup did not start then, I just gave up
waiting.

I have noticed that the PHYLOCK.SWP file seems to be 1GB in size -
exactly so. I find exactitude in such scenarios suspicious. That 1GB
is allocated when the file is created - as soon as the 'gathering
information' stage starts.Maybe 1GB might not be enough? I have
managed to get a backup to fail when heavily using the PC during the
backup - the error message did confirm that the source drive had
changed too much for the backup to continue.

I can toggle this delay with 100% reliability. Selecting metadata hash
does produce a delay and deselecting it brings back the snap-start
noticed with v3.

--
Cheers,

DrT

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

Re: IfW - Gathering information

Posted: Wed Apr 19, 2017 11:22 am
by Brian K
"Gathering information" takes time but the total time of the Changes Only backup is shortened when using /usemd. See my examples with IFW, IFL and IFD.

Are you saying the total time of your Changes Only backup is lengthened when you are using /usemd?

Re: IfW - Gathering information

Posted: Wed Apr 19, 2017 5:12 pm
by DrTeeth
On Wed, 19 Apr 2017 04:22:36 PDT, just as I was about to take a herb,
Brian K disturbed my reverie and wrote:

>Are you saying the total time of your Changes Only backup is lengthened when you are using /usemd?

The "gathering Information" stage took over 20 mins (that's when I
gave up waiting) with metadata hash enabled and well under 30 secs
with it disabled. I do not know how long the backup takes as when it
starts, I go to bed ;-).
--
Cheers,

DrT

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