My workstation was acting up, not shutting down unless power was shut off but on booting again not getting the usual message about Windows not being shut down properly. I ran the full Dell Online Diagnostics and the Backup HDD which is always connected for continuous backup failed the Targeted Read Test and the SMART Short Self Test. I ran CHKDSK /f /r and kept getting the message "disk does not have enough space to replace bad clusters in file xxx", and all the named files were backup files, not other stuff I have on there. So firstly, I know my backups are messed up. What I don't understand is I have 267 GB of free space on the disk, why can none of the bad clusters be replaced? Would appreciate advice on what I should do. Obviously I don't want to lose the years of data I have on there and since HDDs aren't expensive nowadays, I'm leaning to buying a new one to be sure my stuff is safe.
I can't help with data recovery. But if you're constantly backing up large amounts of data I suggest using a RAID setup instead of a single drive. There are arrangements that allow for 1 or 2 drives to fail yet data is still retained due to redundancy. At work we use QNAP products which are cost-effective "network attached storage" solutions.
It seems that some people just love to learn the hard way. Let's summarise the "laws of backup" : (1) If its important, back it up to more than one device, typically the "rule of thumb" is that for anything critical, you should backup to minimum 3 independent devices, and preferably one of the devices is located at a different geographical location to your other backup devices and/or main computer. (2) RAID MUST be considered to be ONE logical device. Therefore backing up to more than one device, as per rule number one, would mean multiple RAID devices. Trust me RAID is not infallible, I've seen lots of nasty corruption on RAID, hence RAID IS ONLY ONE DEVICE ! (3) Replace the word "RAID" in (2) with "cloud storage provider". Same rule. Don't put all your eggs in one basket. (4) Test your backups regularly. An untested backup is NOT a backup. (and I mean manual human testing, not hitting the "verify" button in your software). P.S. Added bonus... for those wondering "why minimum 3 copies" its called N+1 redundancy. If you only made two copies, you would be at N redundancy, which is fine ... until one of your backup devices fails. Then you're down to one, and you have no idea when that remaining device might fail. Therefore with N+1 redundancy (or three copies in laymans terms), you would be down to two working independent copies which buys you more safe time to fix the third.
The on-board drives are configured RAID 0, something Dell decided as I knew nothing about RAID at the time, which is only slightly less than what I know now. The failing drive is an external backup drive so as long as I sort it out before anything happens with the on-board drives, I won't lose any data. I don't think I'm backing up large amounts of data, 40+ symbols of FX and IB Futures data. This might change if I decide to go for Barcharts tick data as I feel IB data isn't prompt enough for short term trading. Market orders result in pretty horrible slippage, something IB is well aware of as I'm constantly getting emails from them warning of the dangers of market and stop orders and suggesting limit/stop limit orders while noting the limitations of those as well.
Thanks Tim, I wasn't aware of some of this. For the really critical stuff, ie my Excel spreadsheets which represent over a thousand hours work in constructing them, I do a separate backup to a thumb drive daily. Do it for peace of mind. I've always had the nagging fear that when I need the backup, it might not work, and this experience reinforces that fear because the backup software did not flag the failing drive and if I didn't have shutdown problems I wouldn't have known. Re (4) above do you mean testing the disk with something like CHKDSK or the Windows equivalent?
BTW Tim your post appears blank on my iPad (Safari), I only managed to read it when I hit reply to tell you, and I don't see it quoted in my reply either. Strange.
1. RAID 0 not a good choice for "important" data. When you have to reload your OS from scratch, suggest you scrap RAID altogether or at least use RAID 1. 2. As a general rule... at the 1st sign of any trouble with a drive, back up the data and replace it. Trying to "save" or "fix" a drive which is acting up is often wasted effort and puts your data further at risk. Besides, with the long warranties for drives, it's better to have the manufacturer replace it than screw with it.
I highly recommend to add to the whole backup solution storage in the cloud as well. Data can be store in encrypted ways hence security should not be a concern. Amazon cloud drive allows virtually unlimited amounts of data for around 60 bucks a year. My setup looks as follows : My windows machines run Storage Spaces (mirrored) though a raid setup would be even better. Then all machines store frequent backups on my main server machine in the local network. The storage on the server itself is raid as well and then the server pushes all new backups on a monthly basis to the cloud. I believe this is slight overkill but I can sleep at night knowing that it's virtually impossible to lose data. Very frequently changing data I run on a Dropbox folder and if any file gets accidentally overwritten then Dropbox allows me to quickly access old file versions rather than having to do a complete folder restore from my backups. Am open to criticism or suggestions if anyone thinks that this can be improved, though it has served me very well so far.
Thanks Scat, appreciate the direct advice. It's out of warranty alas, I bought it in 2013. Time flies at my age, I thought it was two years ago but I keep all my IT receipts in one place and that proved otherwise. I checked online for the replacement price, it's almost 20% cheaper today so I might just spring for a 2 TB drive.