Data recovery – when should I give up? A learning experience in progress

My daughter’s laptop decided a week or two ago that it was a good day to die, and so its hard drive refused to mount. The symptoms were basically those of a startup file problem, so I hoped to be able to refresh the system and move on, but unfortunately it soon became apparent that the hardware itself was at fault – and there was no backup available.

In the past I’ve run a disk utility against errant drives and salvaged what I could, but my usual utilities were unable to get the drive to mount, so that approach was a no-go. I decided instead to try a program new to me which has received good reviews, Data Rescue 3. I ran a demo version briefly to see if it could recover anything, and managed to pull a small file out of the drive. On the strength of that I purchased the full package and aimed it at the failed drive, trying to get a clone of the entire disk, warts and all.

The software immediately reported it had found a problem – slow reads – which suggested the drive was about to fail. OK, I knew that, but at least the software knew something was wrong, and I told it to keep going. About 24 hours later it had processed slightly over 75% of the drive, but somehow, magically, the disk appeared to have mounted – though a system message told me that the system software – not Data Rescue – couldn’t repair the disk, and that I should immediately try to recover as many files as possible. So I did.

The disk directory opened up, and I tried to copy a folder. The system tried to copy, but couldn’t get a handle on the files to be copied – the disk was just in too bad a condition. Oh well, I’d just have to work through Data Rescue. Which was no longer reporting any progress.

Perhaps it was in the middle of a very slow read.

I waited, but the block count which is the underlying progress indication remained unchanged. I waited some more, but there was still no progress. I cancelled the program, but it refused to quit, kept going even though no progress was being reported, and the lights on the recovery drive continued to flash – which should have been a clue. Figuring Data Rescue had simply crashed, I forced the program to quit and tried to restart it. It refused to start, telling me that background activities were still going, and I should try to log out and log back in. Instead, I restarted the system – my second mistake – and fired up Data Rescue again. Wait – where was the 75% of the disk I had already recovered? Evidently in abandoning the process I had left the disk image incomplete, and my system had deleted it, recovering the space on the recovery drive. And so, I made my third mistake, and told Data Rescue to begin again. What I should have done was undeleted the previous disk image first, and tried to recover files from that, but by starting over I over-wrote the previous disk image, making recovery from that image impossible. And the new image is being created from a disk which has had more time to go bad – dying even as Data Rescue was working on it – and no longer contained as much good data as before.

So far, after three days of work, Data Rescue has been able to recover only 30 Gb of “good” disk space – which does not mean good files! – and has slowed down dramatically after getting only 28% through the recovery process. It now projects that it will be weeks, if not months, before the disk is fully processed, and I am considering stopping the cloning and working on what currently survives to recover what I can.

Lessons learned: don’t mess with a disk while recovery is underway; try to recover from partial copies if that’s all that remains; and – for my daughter – back up regularly. In which regard, I will say that I myself do backups, but if the backup and original drives fail together, I’m still dead in the water – so I should review my processes.

OK, it’s been a day….

The old blog turns out to have an incompatible data structure with the new blog, so first I have to upgrade the old blog, then export the content, port it to the new servers, and then import to the new blog. That’s going to take longer than anticipated (and it’s not as though I updated on a regular basis, though I had promised to be a little more active just recently).

Patience, Grasshopper.

…..and a week later, images are still outstanding, but text is up!

A change is gonna come…

It’s been a long while since I posted with any frequency. I’ve been meaning to do something about that, but as with any retired person, I’ve been busy (no, really, it’s very strange how every retired person I know is busy as hell within a few weeks of retirement).

A couple of weeks ago my network connection went down for a few hours, during which all my services – mail, web, FTP, and others – went offline. During the outage I spent a lot of time talking to tech support, being told there was nothing wrong while my modem kept insisting there was no PPoE server available. Eventually I set a monitor to sound an alarm if ever service came back and wandered off to read a book. Two hours later the chimes drew me back to the disaster area that is my office, and service was back to normal. Next day I had a repeat of the outage and once again was told the network was just fine. This is typical for a weekend outage. Weekday outages are handled faster, but it’s always a frustrating exercise.

So. Heads will roll somewhere, but I’ve decided to move my services over to HostPapa – let them deal with outages – and move my connection from a DSL to a fast cable connection. While the DNS moves over there may be interruptions, but at least the cause will be known, and eventually the drive will be a little smoother. And I promise to update more often…

WordPress updater FAIL

I have been updating the blog through the WordPress app on my iPad, but for some reason I am no longer able to connect through that app. I did manage to update my other blog location with it, so it’s not a total fail, but it’s a mystery to me why it works on one but not the other. Not a big deal, since I can still update using the web interface, but it’s still irritating (and as my wife will testify, lots of other things are irritating to me).

Worrying symptom

Earlier this morning I found the web site was hung, along with the mail server. I had to shutdown the system and repair the system disk structures. I have my fingers crossed that this was just a temporary and transient glitch, but if there are problems over the next few days it’s probably due to a more serious failure.

Kicking and screaming

I’ve written before this of my growing frustration with Apple’s Aperture as compared with Adobe’s LightRoom. I’m not the only one – Nick Santilli has written a plea for a major upgrade to Aperture, and I agree with him 100%. LR 3.0 is expected in April – and if Aperture is still stuck at 2.1 I will probably be dragged kicking and screaming over to the Other Side. I want to see at least a point upgrade which delivers 64-bit performance (after trying 3.0 I now find Aperture unpleasantly slow) and better/more editing features. Santilli mentions Faces and Places, and I could get behind that as well, but what I would really like to see is better support for additional themes. Changing an Aperture web page around is painful right now, while LR seems to have third-party themes (both paid and free) coming out the proverbial wazoo (what is that, anyway? something scatological, I betcha).

Show Aperture users a little love, Apple.

Pressure

Every now and then someone tries to get into my system using a simple password attack. I used to ignore them, since there are no simple passwords on my system, but they’ve been clogging up my logs recently, to the extent that I felt it necessary to install blocking software. Since then I’ve been keeping an eye on the number of attacks, which is currently running around 2 or 3 a day. I think of it as network pressure, and it’s an indication of how worthwhile selective late post-parturitive abortions would be (and answering Pirate Jenny’s question, “kill them now, or later?”).

Ouch! Reverting to an older backup

My long-suffering hard drive finally died yesterday, and I was forced to revert to a backup dating back six months. There was very little activity during that time due to weather, and the most significant event was personal rather than astronomical, in that I was caught up in a layoff round and I decided to retire. Timing was an issue, though, and funds are tighter than I would prefer (but then, funds are always tighter than I would prefer).

I’ve been keeping the server back at an earlier OS release in order to keep some legacy software running, but that has meant that I did not avail myself of the backup features built into the later versions of the OS. I’m rethinking that decision in light of the fact that I was unable to recover using the manual backups taken about a month ago. The automated setup would have been more recent and I think more reliable.

ADDED LATER: Everything from November 2008 to August 2009 is irretrievably gone, including notes and images of the local tornado damage. A nightly backup is now running, but at some point I need to test it. Sigh. I thought I was done with Disaster Recovery Exercises when I joined the leisure classes.

Addicted!

Shortly after 10 am yesterday my internet connection failed. I noticed right away, but I waited twenty minutes before contacting my ISP. Tech support there told me nothing was wrong, but after about forty minutes of struggle, I was able to convince them otherwise, and was passed on to ‘second tier’ tech support, from whom I received a problem ticket number and the admission that when they tried my DSL settings from their own office, the connection failed to complete. Six hours later the problem was ongoing, and I called tech support again, to find that the time for resolution could be as long as forty-eight hours due to a shortage of staff. It’s a long weekend here in Ontario, and this was not news I wanted to hear. In the final event, it took almost nineteen hours to restore service, shortly before 5 am this morning.

The time taken to resolve the problem was not, I suppose, particularly long, but I was astonished at the level of frustration and pain I felt during the outage. I have to admit that I am well and truly addicted to the net. It’s going to be interesting to see how I make out from Thursday through Sunday of the coming week, as I attend Starfest. There is supposed to be a shared wireless connection – I’ll have to take a laptop!