Does leaving a program open cause your network backup procedure to fail? The complany I work for recently sent out an email with the following in it:
Since August 16 there have been 10 successful backups and 9 failures. The failures are due to the staff leaving programs, mainly Outlook, open at night. I have stated this time and time again that you need to close your email and other programs before you go home at night. Leaving the programs opens causes the backup to fail,
Is this typical of network backup systems? I suspect it’s not, not because of any knowledge of the systems but rather from some other issues that regularly occur with our network. I would elaborate but I’m afraid I would have to do so in the Pit.
It can certainly happen but it isn’t normal especially for business backup software. Your IT people either have some crappy backup software or they don’t have it configured correctly. Even my home backup software doesn’t care if you leave things open while it is making backups.
Likely they’re not using any backup software at all, but just write a quick-and-dirty batch script to do it. Actual backup software has had this problem solved for decades.
“Decades”? Let’s see, that brings us back to at least '92, Win3.1 days. Under DOS/Win31/BSD/DECOS/takeyerpick, if another process has a file open and is continuously writing to it, what kind of magic allows you to get a true non-corrupt copy of that binary file?
The software can’t do it without a little help from somebody higher up the chain. And of course, there is that possibility that I am mistaken. Waiting for somebody to explain it to me.
Unixes will happily let you do whatever you want with the file, which results in a file in an inconsistent state. If it was that easy you could hot-backup huge mysql databases without having to jump through allllll kinds of hoops, some of which are worse than just taking the production database down briefly.
If you have some trick, there are hundreds of thousands of mysql guys who would like to know it.
As in “DEC OS”, the space would have made it look messy in that sentence. Back in '92 I believe it would have been still called DEC/OSF1 or something. Hell if I remember; I was drunk most of the time back then at work. The boss was kind enough to stock the Coke machine with beer.
FWIW, this still happens in consumer-grade backup software like Syncplicity. I guess they’re supposed to use Volume Shadow Copy, however that works, but it doesn’t always and usually breaks when large files are opened by another program.
The message received by the OP is an itsy-bitsy-bitchy. Anyone who’s done enterprise scale backup work knows that most backups aren’t “successful”; that is, they don’t successfully backup 100% of the target data. This is perfectly acceptable if you’re in an environment that runs (for example) differential backups nightly and a weekly full backup. Basically, some OSs, applications, and databases take “ownership” of data in when it’s being use. They tell the backup software to bug off. The backup solution can then come back and get it the next time. No biggie unless it happens too frequently with the same dataset. In this example, that would be the employee who’s had Outlook open for 3 years without the backup software getting a chance to grab a good backup.
Three questions:
Why the hell aren’t people logging out at the end of the day?
Why the hell is the company backing up user desktops?
Why the hell isn’t the data server-resident?
If they need a consultant, I know a good one (he lives in my clothes).
I do environmental work. Sometimes we are preparing to travel and need to check an email before we leave and then forget to close out of the program (i.e. we’re human).
Desktops aren’t being backed up. Outlook runs on the desktop but email is stored on a server.
I’m literally at this very moment taking a break from writing a proposal document that talks about how great we are… the backspace key is getting heavy use as I remove the subconscious-induced statements like, “we will overcharge you for services that your IT department can already provide, and will take those excessive profits to buy additional 'objective third-party reports” that help us dupe other suckers."
Every place has human employees (much to the chagrin of their employers), and you’re right, they do fairly frequently fail to log out (whether accident or laziness is another question).
If the email is stored on the server, then having Outlook open should be an issue (I’m a former Exchange administrator - trust me on this one).
I think the thing that struck me was just the tone of the message you cited. Email would typically be the last thing that would cause a backup to crap out. On the other hand, if a database is being leveraged heavily when the backup runs, then you’re going to have a lot of records/tables that aren’t backed up.
Think about this (not trying to take a snotty tone) - if having applications open during backups always made the backup “fail” (and I mean totally fail), no organization would be able to operate 24/7. This actually, from an IT perspective was the case 20 years ago (and still is in some instances)… but, seriously, unless you’re running old-school VAX systems all over the place, it’s not an issue.
Actually the tone of the email deteriorated after the part that I quoted. Fortunately it’s not me that has caused the problem, but one coworker said “Looks like I’ve been kicked off the island.” Turns out her email is no longer being backed up because she has caused too many problems.
It’s an excellent point about operating 24/7. Maybe when emotions aren’t so high I’ll bring it up and suggest they look for a solution. It’ll probably get me kicked off the island though.
Back in my IT days, that email would have been a fireable offense under one CIO I worked for. He had his issues, but customer service was priority one and that crap just would not have flown. We worked with the customer (internal employees were customers - his philosophy was that we treated them right or they would get IT services elsewhere, and that would be bad).
It’s been six years since I worked in the field and even back then there is no reason why you couldn’t get a good backup. Something tells me the architecture is screwed up. But you fix your infrastructure problems, don’t blame end users for using the stuff you provide.
Upgrading infrastructure costs money that isn’t generated by the IT department. I don’t know any IT admin that wouldn’t be able to put more money to good use. The problem is going to the non-IT person that controls the budget with a proposal for a good backup system that costs $X plus $Y per user plus $Z per year maintenance and convincing them it’s better to spend the money than “just tell people to log off”.
So accurate!