I AM ACTING AS AN IRRESPONSIBLE DATA OWNER BY PROCEEDING WITHOUT BACKUP

Heh. Would be funny if after you type that it completely wiped your data and displayed "I told you so."

The team that supports us needs their backups too.  Yesterday, when moving a new report into the QA environment, decided to click "REPLACE" rather than "ADD."  The result:  100 reports deleted, 1 report added.  Luckily they have backup.

Very cool - all systems need this!

If it had been me, I'd have just made the message read:

 

It has been %d days since your last backup. Since you obviously don't value your data, I have deleted it to free up space. 

Whoeever wrote that needs a raise.  Beautiful.

This is awesome!
Now, if only Microsoft integrated this into Vista's UAC... "I AM FULLY AWARE THAT INSTALLING THIS THIRD-PARTY GAME MAY DO BAD THINGS AND I WON'T CALL MICROSOFT COMPLAINING IF SOMETHING GOES HORRIBLY WRONG".

I tend to be a disagreeable person. Seeing this sort of a rant would make me inclined to refuse to do backups just out of spite.

zlogic:

This is awesome!
Now, if only Microsoft integrated this into Vista's UAC... "I AM FULLY AWARE THAT INSTALLING THIS THIRD-PARTY GAME MAY DO BAD THINGS AND I WON'T CALL MICROSOFT COMPLAINING IF SOMETHING GOES HORRIBLY WRONG".

Cancel or Allow?

I AM ACTING AS AN IRRESPONSIBLE DATA OWNER BY PROCEEDING WITHOUT BACKUP

Of course there is the WTF that they obviously expect the user to be using their application's built-in backup system. If you are using a third party backup tool, OS tool or RAID for backups then you will just be annoyed by their software whining.

No, 'tis not so deep as a well, nor so wide as a church-door; but 'tis enough, 'twill serve: ask for me to-morrow, and you shall find me a grave man.

This is probably a great idea to get through to the business type running the software that thinks to themselves "Self, I don't need to do backups. My computer works just fine. Why, I even dripped cheese in the fan one time and it's still running! This baby is indestructable!", hoses their system, and then proceeds to verbally berate the vendor for the next hour because their "shitty" software lost all of their work.

I AM ACTING AS A BLISSFULLY IGNORANT PEON BY PROCEEDING WITHOUT BACKUP SINCE I KNOW THAT OUR FILE SERVERS ARE BACKED UP HOURLY AND AS PART OF THE FILE SERVER ROLLOUT TEST CREW I STILL ACTUALLY HAVE ACCESS TO SAID BACKUPS IN REAL TIME--(THOUGH MAYBE I SHOULDN'T ANYMORE?)--SO BUG OFF.

 

 

Hmm, it's missing something. Something like case sensitivity, and random capitalization of every letter. Yeah, that'll do.

mallard:

Of course there is the WTF that they obviously expect the user to be using their application's built-in backup system. If you are using a third party backup tool, OS tool or RAID for backups then you will just be annoyed by their software whining.

Correct me if I'm wrong, but doesn't Windows keep track of backup information, so that (assuming the programmer is paying attention) this is a valid message? I could have sworn that there was an API for both asking Windows "when was this file last backed up" and telling it "I have backed up this file". I know Mac OS X has that capability, thanks to Google.

The Vicar:

mallard:

Of course there is the WTF that they obviously expect the user to be using their application's built-in backup system. If you are using a third party backup tool, OS tool or RAID for backups then you will just be annoyed by their software whining.

Correct me if I'm wrong, but doesn't Windows keep track of backup information, so that (assuming the programmer is paying attention) this is a valid message? I could have sworn that there was an API for both asking Windows "when was this file last backed up" and telling it "I have backed up this file". I know Mac OS X has that capability, thanks to Google.

Wouldn't that just be the 'archive' attribute?

medialint:

OUR FILE SERVERS ARE BACKED UP HOURLY

Hourly? 

dhromed:

medialint:

OUR FILE SERVERS ARE BACKED UP HOURLY

Hourly? 

No HOURLY

MasterPlanSoftware:

dhromed:

medialint:

OUR FILE SERVERS ARE BACKED UP HOURLY

Hourly? 

No HOURLY

A common mistake. 

dhromed:

MasterPlanSoftware:

dhromed:

medialint:

OUR FILE SERVERS ARE BACKED UP HOURLY

Hourly? 

No HOURLY

A common mistake. 

Happens to me all the time, glad I could set you straight.

dhromed:

medialint:

OUR FILE SERVERS ARE BACKED UP HOURLY

Hourly? 

 

Only files that change during the hour. No, the whole thing isn't backed up at that clip.  

MasterPlanSoftware:

The Vicar:

mallard:

Of course there is the WTF that they obviously expect the user to be using their application's built-in backup system. If you are using a third party backup tool, OS tool or RAID for backups then you will just be annoyed by their software whining.

Correct me if I'm wrong, but doesn't Windows keep track of backup information, so that (assuming the programmer is paying attention) this is a valid message? I could have sworn that there was an API for both asking Windows "when was this file last backed up" and telling it "I have backed up this file". I know Mac OS X has that capability, thanks to Google.

Wouldn't that just be the 'archive' attribute?

 
Anyway, who's to say that software is even on Windows. That screenshot is clearly of a terminal program, not a DOS window.

mallard:

RAID for backups

 RAID is a way to get redundancy. RAID is neither a backup strategy nor a replacement for one.

mallard:

Of course there is the WTF that they obviously expect the user to be using their application's built-in backup system. If you are using a third party backup tool, OS tool or RAID for backups then you will just be annoyed by their software whining.

The WTF is that you consider RAIDing disks as a backup tool. 

mallard:

[...] RAID for backups [...]

...

 RAID is *DISK* redundancy, not *FILE* backup. What happens when you delete data from one disk on the RAID?

 From my experience with RAIDs, I treat all the disks as if they were one and the same disk, merely mirrors of eachother.
 

tchize:

mallard:

Of course there is the WTF that they obviously expect the user to be using their application's built-in backup system. If you are using a third party backup tool, OS tool or RAID for backups then you will just be annoyed by their software whining.

The WTF is that you consider RAIDing disks as a backup tool. 

Well it does provide protection against disc failure. Coupled with a file-system revisioning system then it is an effective backup strategy.
 

Quinnum:

Anyway, who's to say that software is even on Windows. That screenshot is clearly of a terminal program, not a DOS window.

IIRC, "lightspeed" is the terminal emulator for the Wang Labs version of the s/360. Mainframe stuff. 

bdew:

mallard:

RAID for backups

 RAID is a way to get redundancy. RAID is neither a backup strategy nor a replacement for one.

Eeh, Seen it used as such tho with mirroring raid. Pull half the drives out, insert an other set. Also a neat trick when trying a risky upgrade without too much time for a rollback.

DigitalXeron:

mallard:

[...] RAID for backups [...]

...

 RAID is *DISK* redundancy, not *FILE* backup. What happens when you delete data from one disk on the RAID?

 From my experience with RAIDs, I treat all the disks as if they were one and the same disk, merely mirrors of eachother.

I have seen RAID controllers used for backup purposes. You make a RAID-1 volume and shove a hot-swap drive into one slot in the morning, then yank it out the next day and put a new drive in. You can make a 3-disk volume if you still want RAID-like redundancy. It works, although it's a little crude.

You can also separate the "hardware failure" and "user error" backup cases and handle them differently. If you tackle hardware failure by having suitable redundancy in your hardware, then you can simply make duplicate copies of the data on the volume overnight. I do this anyway: all of our fileservers have a hot copy of the past three weeks of nightly copies, stored in a hardlink forest, which the users can directly access. I never have to restore a file for a user who deleted it: they can simply fish it out of the backup volume.

asuffield:

I have seen RAID controllers used for backup purposes. You make a RAID-1 volume and shove a hot-swap drive into one slot in the morning, then yank it out the next day and put a new drive in. You can make a 3-disk volume if you still want RAID-like redundancy. It works, although it's a little crude.

Two things:

1)  wasn't that featured on the front page?

2)  That sounds very expensive, unless you subsequently pull the data off the hot-swap, compress it, archive it on another drive, then re-use the hot-swap the next morning.

belgariontheking:

2)  That sounds very expensive, unless you subsequently pull the data off the hot-swap, compress it, archive it on another drive, then re-use the hot-swap the next morning.

You haven't looked at the prices of backup tapes lately, have you? It's actually a heck of a lot cheaper, if you can fit all your data on a drive no larger than about 250Gb. Economy of scale has been heavily applied to hard drives over the past decade, but not so much to tape technology.

Tapes don't become cost-effective until you head up into the multiple-terabyte range - the cutover point has been hovering around 2-3 times the size of the largest hard drive on the market for the past few years.

 

mallard:

tchize:

mallard:

Of course there is the WTF that they obviously expect the user to be using their application's built-in backup system. If you are using a third party backup tool, OS tool or RAID for backups then you will just be annoyed by their software whining.

The WTF is that you consider RAIDing disks as a backup tool. 

Well it does provide protection against disc failure. Coupled with a file-system revisioning system then it is an effective backup strategy.
 

I take a shotgun to your server. Do you still consider RAID to be a backup strategy?

Some genius of a sysadmin types "rm -rf /". Do you still consider RAID to be a backup strategy?

Lightning strikes. Do you still consider RAID to be a backup strategy?

After replacing a failed disk, the RAID controller rebuilds the array using the existing data from the new disk and the blank space on the old disk. Do you still consider RAID to be a backup strategy?

Carnildo:

I take a shotgun to your server. Do you still consider RAID to be a backup strategy?

Some genius of a sysadmin types "rm -rf /". Do you still consider RAID to be a backup strategy?

Lightning strikes. Do you still consider RAID to be a backup strategy?

After replacing a failed disk, the RAID controller rebuilds the array using the existing data from the new disk and the blank space on the old disk. Do you still consider RAID to be a backup strategy?

If the building burns down, does that mean tape backups don't count as backups?

If a meteor strikes the Earth, does that mean backups stored in the same hemisphere don't count as backups?

There are obviously degrees of backups, from "burn things to a cd once in a while" to "automatically backup nightly to 10 different locations around the globe".

Carnildo:

I take a shotgun to your server. Do you still consider RAID to be a backup strategy?

The most recent drive from the offsite storage is retrieved and the volume reconstructed. Yes, still valid.

Some genius of a sysadmin types "rm -rf /". Do you still consider RAID to be a backup strategy?

Last night's filesystem image is copied back from the backup volume, which is mounted read-only. Still valid.

Lightning strikes. Do you still consider RAID to be a backup strategy?

The UPS soaks up the lightning strike and the servers carry on working happily. No problems here.

After replacing a failed disk, the RAID controller rebuilds the array using the existing data from the new disk and the blank space on the old disk. Do you still consider RAID to be a backup strategy?

It's a RAID-1 volume, that can't happen...
 

The real WTF is that photobucket has an absurdly low bandwidth limit and I can only see a "Bandwidth Exceeded" graphic.

powerlord:

The real WTF is that photobucket has an absurdly low bandwidth limit and I can only see a "Bandwidth Exceeded" graphic.

 

I think this pic is also on Digg and/or /. so it getting hammered..... 

asuffield:

Carnildo:

Some genius of a sysadmin types "rm -rf /". Do you still consider RAID to be a backup strategy?

Last night's filesystem image is copied back from the backup volume, which is mounted read-only. Still valid.

One read-only volume and one read-write volume isn't any form of RAID I've encountered.

Lightning strikes. Do you still consider RAID to be a backup strategy?

The UPS soaks up the lightning strike and the servers carry on working happily. No problems here.

What fantasy world are you living in? A UPS will protect against a near miss, but a direct hit on the building will turn any real-world UPS into a smoking ruin and continue on to wreck the server.

After replacing a failed disk, the RAID controller rebuilds the array using the existing data from the new disk and the blank space on the old disk. Do you still consider RAID to be a backup strategy?

It's a RAID-1 volume, that can't happen...
 

Software never has flaws?

Carnildo:

Lightning strikes. Do you still consider RAID to be a backup strategy?

asuffield:

The UPS soaks up the lightning strike and the servers carry on working happily. No problems here.

Carnildo:

What fantasy world are you living in? A UPS will protect against a near miss, but a direct hit on the building will turn any real-world UPS into a smoking ruin and continue on to wreck the server.

Actually, asuffield is mostly right, the UPS should not transfer the lightning strike. Theoretically it COULD happen, but it would have to be HUGE and direct. It is highly unlikely. There are ratings on most UPS devices that will tell you how many ka and kv the UPS can isolate in this situation. If selected and manufactured correctly, you should be in good shape.

However, I do agree asuffield is wrong that "the servers carry on working happily". Nope, UPSes would all be trashed, the servers are not going to magically keep working. I have yet to see a UPS withstand a decent lightning strike and happily continue service. They are meant to be the expendable member in this type of situation.

You will be down, but at least you minimized damage to the servers.

Carnildo:

asuffield:

Carnildo:

Some genius of a sysadmin types "rm -rf /". Do you still consider RAID to be a backup strategy?

Last night's filesystem image is copied back from the backup volume, which is mounted read-only. Still valid.

One read-only volume and one read-write volume isn't any form of RAID I've encountered.

Regardless of your lack of experience, this is my normal server configuration (even if there's also a secondary tape backup). It's the fastest-recovering system I've been able to come up with - usually back up from total failure in under an hour, if all the parts are on hand.

Lightning strikes. Do you still consider RAID to be a backup strategy?

The UPS soaks up the lightning strike and the servers carry on working happily. No problems here.

What fantasy world are you living in? A UPS will protect against a near miss, but a direct hit on the building will turn any real-world UPS into a smoking ruin and continue on to wreck the server.

Stop buying cheap-arse UPSes. All the good ones have a spark gap, which will ground out any lightning strike that doesn't actually immolate the building on the spot. It's  as effective as any lightning conductor.

After replacing a failed disk, the RAID controller rebuilds the array using the existing data from the new disk and the blank space on the old disk. Do you still consider RAID to be a backup strategy?

It's a RAID-1 volume, that can't happen...
 

Software never has flaws?

RAID-1 doesn't have a 'rebuild', it just copies the contents of the active volume onto newly added drives. Can't have flaws in code that doesn't exist. 

if someone competent is given more than a penance to set up a server that needs high availability then a lightning strike probably won't take down your server. 

tster:

if someone competent is given more than a penance

ITYM "pittance" 

asuffield:

RAID-1 doesn't have a 'rebuild', it just copies the contents of the active volume onto newly added drives. Can't have flaws in code that doesn't exist. 

If the RAID controller doesn't use code to perform the copy operation, what does it use?  Magic? 

operagost:

asuffield:

RAID-1 doesn't have a 'rebuild', it just copies the contents of the active volume onto newly added drives. Can't have flaws in code that doesn't exist. 

If the RAID controller doesn't use code to perform the copy operation, what does it use?  Magic? 

You might as well argue that because calc.exe has code in it to add numbers and code can have bugs, adding numbers in calc could destroy your RAID volume. It's crack. 

asuffield:

operagost:

asuffield:

RAID-1 doesn't have a 'rebuild', it just copies the contents of the active volume onto newly added drives. Can't have flaws in code that doesn't exist. 

If the RAID controller doesn't use code to perform the copy operation, what does it use?  Magic? 

You might as well argue that because calc.exe has code in it to add numbers and code can have bugs, adding numbers in calc could destroy your RAID volume. It's crack. 

Calc.exe, while it is powerful enough to calculate numbers in the BILLIONS,  doesn't have direct access to storage; nor does it even open a file for write access.

operagost:

asuffield:

operagost:

asuffield:

RAID-1 doesn't have a 'rebuild', it just copies the contents of the active volume onto newly added drives. Can't have flaws in code that doesn't exist. 

If the RAID controller doesn't use code to perform the copy operation, what does it use?  Magic? 

You might as well argue that because calc.exe has code in it to add numbers and code can have bugs, adding numbers in calc could destroy your RAID volume. It's crack. 

Calc.exe, while it is powerful enough to calculate numbers in the BILLIONS,  doesn't have direct access to storage; nor does it even open a file for write access.

THERE is the TRWTF. How can you multiply without a temp file???

WTF?!

TRWTF is.... Bandwidth limits