I have no idea - what is your estimate? I can only go by the information in the linked article, that it will take a few days to sort out: "Torvalds is attempting to recover the dead hard disk but at the moment it doesn't appear easy and subsystem maintainers who have outstanding pull requests may need to re-submit their requests in the coming days."
Oh my God! They will each have to execute a single push command! The horror!
Making other people pay for your own mistakes is bad karma.
I can't agree. Having to deal with a courtroom situation where your lawyers can't sweep it all away with a private settlement must cause corporate lawyer nightmares. There are probably all sorts of insurance and contractual implications to being convicted of something.
Dunno about you, I would notice if my garden hose was on continuously for 5 days... and it's not spouting wastewater.
They may very well have been negligent - that's what the grand jury decided, anyway. I was responding to the comment about locking someone up. I meant to put the scale of the spill in perspective - this isn't some grand environmental disaster.
You might not notice your garden hose running, depending on where the water was going. It was raining the day the inspectors arrived. Just being the devil's advocate.
They ARE being charged with criminal offenses: "XTO Energy Inc. is charged with five counts of unlawful conduct under the Clean Streams Law and three counts of unlawful conduct under the Solid Waste Management Act."
They were held accountable. They had to remove soil and paid a fine. They are also being charged under environmental laws in the state of PA - I'm just not sure how jail time would help here. I tend to be anti-jail for nonviolent offenders, so maybe I'm just biased.
Any human endeavor will involve mistakes. Sending some poor slob to jail for forgetting to put a plug in is not really justified IMHO. Sending his manager to jail for not double-checking is once removed from that. Sending his VP to jail starts to get even more absurd. The company needs to own up and make good, but jail wouldn't be very effective here.
Note that my wife is a doctor, so I may have a skewed view of the tort system. I'd hate to see her go to jail for an inevitable mistake, or one by one of the staff she supervises. Everyone makes mistakes, and all man-made systems have flaws. You want a disincentive, to be sure, but ruining someone's life for an honest mistake seems heavy-handed.
Where does it say that? I see in the summary and linked article where it says he's trying to recover the drive and if he can't then people will need to re-submit over the next couple of days.
No, seriously. I think a garden hose is around 300 GPH, and this works out to 416 GPH. It's not that big of a flow and I could buy their claim that they didn't know they were leaking.
But they didn't deliberately deface anything - they left a plug out of a tank, which leaked contaminated water at a rate which may not have seemed significant. They seem to have made good on the cleanup. Intent matters - that's why we have murder and manslaughter.
I certainly agree that I'd like criminal charges applied to decision makers at companies like this. BUT...
50,000 gallons isn't that much water. It was a 10,000 gallon-per-day spill. That's garden-hose territory. The fact that they were made to clean it up and pay fines seems reasonable to me - I'm not sure jail time is warranted here.
I'm sure all of the data still exists somewhere - git is very nice for that. However, there is time and work involved with corralling all of it and getting back to where he was. Had he been using even the most brain-dead, consumer-level backup schemes like Dropbox he wouldn't have lost any work at all. If you read through those kernel mailing lists, the git pushes often seem quite eventful:)
I don't think 10 good drives with an average retirement age of about 3 years is terribly exceptional. Good luck for sure, but not "against all odds" stuff. I've only had 1 drive (coincidentally a WD Green) go bad on me in that time period, if you exclude physical damage. And even then, the WD Green committed suicide by parking its head every 5 seconds for a year. I didn't realize this was occurring until my server started sending me SMART alerts. The replacement has the patch so that it won't do that:) I have some drives that are probably 9 years old.
I'm even more paranoid and use Crashplan for the "cloud" and a native backup (Windows Backup, Time Machine, etc) for local backups to a local server. That way, if Crashplan screws up I still have the other backup method going.
You offered "solution" after "solution" that is no solution at all,
You can say that all you want, but if he were running Crashplan, he'd have lost maybe 20 minutes of his working day. Switch to one of his other computers, restore his working directory to a new directory and keep on truckin'. Instead he's struggling to recover data from a dead drive. Dropbox or Sparkleshare wouldn't even require a restore step - just walk over to the other computer and pick up where you left off. Do you dispute this fact?
and seem to have missed the first line of his post: "I had pushed out _most_ of my pulls today, so realistically I didn't lose a lot of work."
And you seem to have missed the part where he's been trying to recover the dead drive and told people that they will need to re-push their changes over the next couple of days. Does that sound like a trivial amount of work lost to you? Days of recovery?
There is no reason, and no advantage to, using a separate git repo and storing your whole git repo in a git repo.
Except, you know, that it would be backed up if the drive died.
I can never decide if it is hilarious or sad that people think they could teach Linus a thing or two about software development.
I don't think "running a backup application" is part of "software development", and it's clearly something he needs help with. The man is clearly not as infallible as you seem to think he is. His mistake now requires work from other people to clean up.
Huh? Dropbox is a local folder. A daemon monitors the folder for changes and then uploads those changes to the Dropbox servers. It keeps a version history for each file, and so you can restore to a point in the past. In my experience, it does not use much in the way of system resources. Sparkleshare does the same thing, but to a server of your choosing and it is completely open source - and in fact based upon git.
That is a ridiculous statement. Work is lost every time a drive fails unless it happens to fail immediately after a backup.
Hasn't Apple's Time Machine sort of set the bar a little higher on frequency of backup? Even if his working directory just sat in a Dropbox (Sparkleshare if you prefer) folder, he'd be better off. My computers all have access-based remote and local backup. I could move over to another computer and have my working directory restored in 10 minutes.
Like I said, he likely didn't lose any code because, hopefully, the contributors' drives didn't also crash and so they can just re-push their changes. But he did lose work, because now everyone has to re-push their changes and he has to co-ordinate it all. This isn't just a delay - people have to go back and re-do stuff they've already done.
I'm starting to look like a shill in the comments for this story, but what the hell. I've been happy with Crashplan. They charge you to store on the cloud, but it's free to your own computers, or friends computers. It also runs a hash against everything it backs up, so data integrity is supposedly "assured". Another commenter has been going on about Bacula, but I've never tried it.
Git is fantastic and you are _technically_ right, in that no code was probably lost. However, work _was_ lost. And not just Linus's work, but the work of kernel contributors. Now people have to redo a lot of effort, because he was too lazy or arrogant to install one of the many effortless backup systems available.
I'm quite happy with Crashplan, but it is obviously not the only versioning backup system. Sparkleshare is an open source Dropbox clone with a git backend.
I'm not sure why being a fanboy of good software is a problem - you are a fanboy of open software. Did that hurt? No? Now you know how I feel.
Slashdot Mirror
How many minutes of his working day did he lose?
I have no idea - what is your estimate? I can only go by the information in the linked article, that it will take a few days to sort out:
"Torvalds is attempting to recover the dead hard disk but at the moment it doesn't appear easy and subsystem maintainers who have outstanding pull requests may need to re-submit their requests in the coming days."
Oh my God! They will each have to execute a single push command! The horror!
Making other people pay for your own mistakes is bad karma.
I can't agree. Having to deal with a courtroom situation where your lawyers can't sweep it all away with a private settlement must cause corporate lawyer nightmares. There are probably all sorts of insurance and contractual implications to being convicted of something.
Dunno about you, I would notice if my garden hose was on continuously for 5 days... and it's not spouting wastewater.
They may very well have been negligent - that's what the grand jury decided, anyway. I was responding to the comment about locking someone up. I meant to put the scale of the spill in perspective - this isn't some grand environmental disaster.
You might not notice your garden hose running, depending on where the water was going. It was raining the day the inspectors arrived. Just being the devil's advocate.
You don't seem to be aware that they are facing criminal charges.
They ARE being charged with criminal offenses:
"XTO Energy Inc. is charged with five counts of unlawful conduct under the Clean Streams Law and three counts of unlawful conduct under the Solid Waste Management Act."
It just doesn't involve jail.
They were held accountable. They had to remove soil and paid a fine. They are also being charged under environmental laws in the state of PA - I'm just not sure how jail time would help here. I tend to be anti-jail for nonviolent offenders, so maybe I'm just biased.
Any human endeavor will involve mistakes. Sending some poor slob to jail for forgetting to put a plug in is not really justified IMHO. Sending his manager to jail for not double-checking is once removed from that. Sending his VP to jail starts to get even more absurd. The company needs to own up and make good, but jail wouldn't be very effective here.
Note that my wife is a doctor, so I may have a skewed view of the tort system. I'd hate to see her go to jail for an inevitable mistake, or one by one of the staff she supervises. Everyone makes mistakes, and all man-made systems have flaws. You want a disincentive, to be sure, but ruining someone's life for an honest mistake seems heavy-handed.
NOTHING about this being the maximum release.
That's a good point. But I can't guess better than the grand jury.
Where does it say that? I see in the summary and linked article where it says he's trying to recover the drive and if he can't then people will need to re-submit over the next couple of days.
No, seriously. I think a garden hose is around 300 GPH, and this works out to 416 GPH. It's not that big of a flow and I could buy their claim that they didn't know they were leaking.
How can anyone make such a claim and be taken seriously? A warrant will always let the government have your data.
But they didn't deliberately deface anything - they left a plug out of a tank, which leaked contaminated water at a rate which may not have seemed significant. They seem to have made good on the cleanup. Intent matters - that's why we have murder and manslaughter.
I certainly agree that I'd like criminal charges applied to decision makers at companies like this. BUT...
50,000 gallons isn't that much water. It was a 10,000 gallon-per-day spill. That's garden-hose territory. The fact that they were made to clean it up and pay fines seems reasonable to me - I'm not sure jail time is warranted here.
I'm sure all of the data still exists somewhere - git is very nice for that. However, there is time and work involved with corralling all of it and getting back to where he was. Had he been using even the most brain-dead, consumer-level backup schemes like Dropbox he wouldn't have lost any work at all. If you read through those kernel mailing lists, the git pushes often seem quite eventful :)
I don't think 10 good drives with an average retirement age of about 3 years is terribly exceptional. Good luck for sure, but not "against all odds" stuff. I've only had 1 drive (coincidentally a WD Green) go bad on me in that time period, if you exclude physical damage. And even then, the WD Green committed suicide by parking its head every 5 seconds for a year. I didn't realize this was occurring until my server started sending me SMART alerts. The replacement has the patch so that it won't do that :) I have some drives that are probably 9 years old.
I'm even more paranoid and use Crashplan for the "cloud" and a native backup (Windows Backup, Time Machine, etc) for local backups to a local server. That way, if Crashplan screws up I still have the other backup method going.
You offered "solution" after "solution" that is no solution at all,
You can say that all you want, but if he were running Crashplan, he'd have lost maybe 20 minutes of his working day. Switch to one of his other computers, restore his working directory to a new directory and keep on truckin'. Instead he's struggling to recover data from a dead drive. Dropbox or Sparkleshare wouldn't even require a restore step - just walk over to the other computer and pick up where you left off. Do you dispute this fact?
and seem to have missed the first line of his post: "I had pushed out _most_ of my pulls today, so realistically I didn't lose a lot of work."
And you seem to have missed the part where he's been trying to recover the dead drive and told people that they will need to re-push their changes over the next couple of days. Does that sound like a trivial amount of work lost to you? Days of recovery?
There is no reason, and no advantage to, using a separate git repo and storing your whole git repo in a git repo.
Except, you know, that it would be backed up if the drive died.
I can never decide if it is hilarious or sad that people think they could teach Linus a thing or two about software development.
I don't think "running a backup application" is part of "software development", and it's clearly something he needs help with. The man is clearly not as infallible as you seem to think he is. His mistake now requires work from other people to clean up.
Huh? Dropbox is a local folder. A daemon monitors the folder for changes and then uploads those changes to the Dropbox servers. It keeps a version history for each file, and so you can restore to a point in the past. In my experience, it does not use much in the way of system resources. Sparkleshare does the same thing, but to a server of your choosing and it is completely open source - and in fact based upon git.
You'll notice that I singled out frequency. :)
That is a ridiculous statement. Work is lost every time a drive fails unless it happens to fail immediately after a backup.
Hasn't Apple's Time Machine sort of set the bar a little higher on frequency of backup? Even if his working directory just sat in a Dropbox (Sparkleshare if you prefer) folder, he'd be better off. My computers all have access-based remote and local backup. I could move over to another computer and have my working directory restored in 10 minutes.
Like I said, he likely didn't lose any code because, hopefully, the contributors' drives didn't also crash and so they can just re-push their changes. But he did lose work, because now everyone has to re-push their changes and he has to co-ordinate it all. This isn't just a delay - people have to go back and re-do stuff they've already done.
I'm starting to look like a shill in the comments for this story, but what the hell. I've been happy with Crashplan. They charge you to store on the cloud, but it's free to your own computers, or friends computers. It also runs a hash against everything it backs up, so data integrity is supposedly "assured". Another commenter has been going on about Bacula, but I've never tried it.
Git is fantastic and you are _technically_ right, in that no code was probably lost. However, work _was_ lost. And not just Linus's work, but the work of kernel contributors. Now people have to redo a lot of effort, because he was too lazy or arrogant to install one of the many effortless backup systems available.
On a drive somewhere. It can be on an NAS or something locally attached.
I'm quite happy with Crashplan, but it is obviously not the only versioning backup system. Sparkleshare is an open source Dropbox clone with a git backend.
I'm not sure why being a fanboy of good software is a problem - you are a fanboy of open software. Did that hurt? No? Now you know how I feel.