On Wed, 26 Mar 2008 09:06:42 +0000, Phil Cook wrote:

> And lo on Tue, 25 Mar 2008 22:42:52 -0000, Jim Henderson
> <nos### [at] nospamcom> did spake, saying:
> 
>> On Tue, 18 Mar 2008 16:40:28 +0000, Phil Cook wrote:
>>
>>> It all depends on what level of back-up you require. Heh my favourite
>>> is a company that used the grandfather/father/son system of backups,
>>> but reused the tapes. So Day 1: GF, Day 2: F, Day 3: S, Day 4: GF etc.
>>> I'm sure you can all see the problem.
>>
>> Palindrome?
> 
> I don't recall.

I recall from my NetWare 4 days that it was one of a very few that used 
GFS style backups - which are/were more popular on mainframe systems.

>> That's one package that used GFS backup strategies - it can work, of
>> course, as long as your tape rotation is appropriate to retaining the
>> full backups.
> 
> And that of course was the problem, if on day 4 the tape got screwed
> they'd have no full backup until they managed another backup. It never
> occurred to them if the server failed between those two points they'd
> have no full backup available.

Ouch.  That's gotta hurt.

Jim

Post a reply to this message

On Wed, 26 Mar 2008 11:09:58 +0000, Orchid XP v7 wrote:

> Jim Henderson wrote:
>> On Tue, 18 Mar 2008 11:43:43 +0000, Invisible wrote:
>> 
>>> What do you use to back up data?
>> 
>> Backup Exec is popular.
> 
> Not with me it isn't. But hey, that's what we have, so...

What problems do you have with it?  I know a guy in your area who used to 
work in BE support (he's now out of the business, but I still may have a 
few contacts inside Veritas - I think now a part of EMC, isn't it?)

Jim

Post a reply to this message

And lo on Wed, 26 Mar 2008 14:01:57 -0000, Jim Henderson  
<nos### [at] nospamcom> did spake, saying:

> On Wed, 26 Mar 2008 09:06:42 +0000, Phil Cook wrote:
>
>> And lo on Tue, 25 Mar 2008 22:42:52 -0000, Jim Henderson
>> <nos### [at] nospamcom> did spake, saying:
>>
>>> On Tue, 18 Mar 2008 16:40:28 +0000, Phil Cook wrote:
>>>
>>>> It all depends on what level of back-up you require. Heh my favourite
>>>> is a company that used the grandfather/father/son system of backups,
>>>> but reused the tapes. So Day 1: GF, Day 2: F, Day 3: S, Day 4: GF etc.
>>>> I'm sure you can all see the problem.
>>>
>>> Palindrome?
>>
>> I don't recall.
>
> I recall from my NetWare 4 days that it was one of a very few that used
> GFS style backups - which are/were more popular on mainframe systems.

Might have been Arcserve, rings bells.

>>> That's one package that used GFS backup strategies - it can work, of
>>> course, as long as your tape rotation is appropriate to retaining the
>>> full backups.
>>
>> And that of course was the problem, if on day 4 the tape got screwed
>> they'd have no full backup until they managed another backup. It never
>> occurred to them if the server failed between those two points they'd
>> have no full backup available.
>
> Ouch.  That's gotta hurt.

They were lucky in that it had never happened, the only 'problem' they had  
was the rolling restore window. The users had got into the habit of  
renaming files etc. out to ensure they wouldn't get overwritten before the  
archived monthly tape (yes they did at least have one of those) the amount  
of old crap that just got left on the servers [shakes head].

-- 
Phil Cook

--
I once tried to be apathetic, but I just couldn't be bothered
http://flipc.blogspot.com

Post a reply to this message

On Wed, 26 Mar 2008 14:33:14 +0000, Phil Cook wrote:

>> I recall from my NetWare 4 days that it was one of a very few that used
>> GFS style backups - which are/were more popular on mainframe systems.
> 
> Might have been Arcserve, rings bells.

I don't believe Arcserve has ever used GFS rotations - though I haven't 
looked at the latest versions, so maybe they added it.  Palindrone was 
bought by Arcadia IIRC, the original owners of Backup Exec.  They got 
bought by Veritas, who then was bought by EMC IIRC (might've been 
Symmantec, though, now that I think of it).

>> Ouch.  That's gotta hurt.
> 
> They were lucky in that it had never happened, the only 'problem' they
> had was the rolling restore window. The users had got into the habit of
> renaming files etc. out to ensure they wouldn't get overwritten before
> the archived monthly tape (yes they did at least have one of those) the
> amount of old crap that just got left on the servers [shakes head].

Users can be a pain, can't they?

Jim

Post a reply to this message

And lo on Wed, 26 Mar 2008 14:42:27 -0000, Jim Henderson  
<nos### [at] nospamcom> did spake, saying:

> On Wed, 26 Mar 2008 14:33:14 +0000, Phil Cook wrote:
>
>>> I recall from my NetWare 4 days that it was one of a very few that used
>>> GFS style backups - which are/were more popular on mainframe systems.
>>
>> Might have been Arcserve, rings bells.
>
> I don't believe Arcserve has ever used GFS rotations - though I haven't
> looked at the latest versions, so maybe they added it.  Palindrone was
> bought by Arcadia IIRC, the original owners of Backup Exec.  They got
> bought by Veritas, who then was bought by EMC IIRC (might've been
> Symmantec, though, now that I think of it).

Might not have been then, wasn't my section.

>>> Ouch.  That's gotta hurt.
>>
>> They were lucky in that it had never happened, the only 'problem' they
>> had was the rolling restore window. The users had got into the habit of
>> renaming files etc. out to ensure they wouldn't get overwritten before
>> the archived monthly tape (yes they did at least have one of those) the
>> amount of old crap that just got left on the servers [shakes head].
>
> Users can be a pain, can't they?

I think they got fed up with "I deleted/changed this four days ago can I  
get it back please?"- "No" so they worked around it. If the users start  
doing things like this it's time to look at why.

-- 
Phil Cook

--
I once tried to be apathetic, but I just couldn't be bothered
http://flipc.blogspot.com

Post a reply to this message

>> We do. On every backup run. And to this day, I have never, ever had a
>> problem restoring something that the backup software said was OK.
> 
> Do you do a checksum-based verification, or a compare-to-disk 
> verification?

It should be a compare-to-disk verification. (At least, it generates 
enough I/O traffic!) I would think anything less wouldn't tell you much...

>> Indeed, if the tape is wearing out, I usually get an error message
>> before the verify phase even begins. (Not sure how it manages that, but
>> it often reports a CRC error...)
> 
> I imagine that even without a verify, your software checks during the 
> backup process to make sure it was written properly.

Yeah, but I was thinking at the physical level. Maybe it's a 3-head tape 
system?

-- 
http://blog.orphi.me.uk/
http://www.zazzle.com/MathematicalOrchid*

Post a reply to this message

Jim Henderson wrote:

> What problems do you have with it?  I know a guy in your area who used to 
> work in BE support (he's now out of the business, but I still may have a 
> few contacts inside Veritas - I think now a part of EMC, isn't it?)

Well, apart from the general level of difficulty of getting any SCSI 
device to work properly in the first place, it *is* quite annoying how 
every single damn time BE has any kind of problem, it refuses to respond 
to any user input until you forcibly terminate all its services. I don't 
know *why* they bothered having a "cancel job" button; it never ever works!

-- 
http://blog.orphi.me.uk/
http://www.zazzle.com/MathematicalOrchid*

Post a reply to this message

On Wed, 26 Mar 2008 15:02:22 +0000, Orchid XP v7 wrote:

>>> We do. On every backup run. And to this day, I have never, ever had a
>>> problem restoring something that the backup software said was OK.
>> 
>> Do you do a checksum-based verification, or a compare-to-disk
>> verification?
> 
> It should be a compare-to-disk verification. (At least, it generates
> enough I/O traffic!) I would think anything less wouldn't tell you
> much...

I can certainly tell you it doesn't.  The only way to test the 
readability of a tape effectively is to actually read it and compare the 
data to disk.

>>> Indeed, if the tape is wearing out, I usually get an error message
>>> before the verify phase even begins. (Not sure how it manages that,
>>> but it often reports a CRC error...)
>> 
>> I imagine that even without a verify, your software checks during the
>> backup process to make sure it was written properly.
> 
> Yeah, but I was thinking at the physical level. Maybe it's a 3-head tape
> system?

Could be.

Jim

Post a reply to this message

On Wed, 26 Mar 2008 14:54:04 +0000, Phil Cook wrote:

>> Users can be a pain, can't they?
> 
> I think they got fed up with "I deleted/changed this four days ago can I
> get it back please?"- "No" so they worked around it. If the users start
> doing things like this it's time to look at why.

One of the many reasons people use NetWare - you delete a file, you don't 
have to go back to tape to recover it, as long as salvage is turned 
on. :-)

Jim

Post a reply to this message

On Wed, 26 Mar 2008 15:04:37 +0000, Orchid XP v7 wrote:

> Jim Henderson wrote:
> 
>> What problems do you have with it?  I know a guy in your area who used
>> to work in BE support (he's now out of the business, but I still may
>> have a few contacts inside Veritas - I think now a part of EMC, isn't
>> it?)
> 
> Well, apart from the general level of difficulty of getting any SCSI
> device to work properly in the first place,

Not difficult at all, as long as the bus is properly terminated.  90% of 
all SCSI issues are improperly terminated buses.  10% are probably 
mismatched SCSI technologies (speed, etc).

> it *is* quite annoying how
> every single damn time BE has any kind of problem, it refuses to respond
> to any user input until you forcibly terminate all its services. I don't
> know *why* they bothered having a "cancel job" button; it never ever
> works!

Well, if the problem is a hardware-related problem, that kinda makes 
sense - the system has to wait for the hardware to respond.  I don't 
think any backup program would work properly if there was a hardware 
issue...

Jim

Post a reply to this message