Unfortunately after a major server crash and being held up in a server room working for 36 hours straight, I wrote some bad scripting and I am HOPING that I didn't damage or corrupt the file structure. Any help in how and why I may be ok or not ok would be greatly appreciated!
Essentially, this is what happened...
1. We operate a virtualized window server and utilize 7-zip to perform data backups from several critical data areas on one of the hard drives... We store these backups on an external hard drive.
2. In the past I have used a simple batch script I wrote using 6-10 in-line windows start commands to simultaneously compile a bunch of 7-zip files for archiving on that external hard drive...
3. Unfortunately tonight, I added a file rename (appending the date) at the end of the "start" sequences.
4. I thought about it after I finally left and went home for the day, but logged in remotely and there was no backup there... I am assuming that for a couple of seconds that all of these 7-zip archives were assembling and all of the sudden a file rename command pulled the rug out from underneath all of them and it must have erred out or something.
My question is, should I be concerned about anything? Could I have damaged the file structure on the removable? How about the NTFS file system from which I was copying from?
Again, any advise as to why I would be ok or not ok would be greatly appreciated... After this major crash and 36 hour day, I am super paranoid about stability at this point!
Another way to put this I suppose so that I can make this determination is, "What exactly is happening when taking files from one drive and compressing them into a file on a different drive?"
Does it begin with a table of contents write to place the file name and then is it compressing in memory and then periodically "dumping" onto the other drive and is that why the file size remains zero until the compression normally completes? If so, I'm probably fine because the files were renamed a second or two after the compressions were started and wouldn't have had time to fill up the server's memory.
OR, Does it immediately start writing, sector by sector, but NOT write to the table of contents and sector start and end locations until the end of the compression. If this is the case, I'm probably still ok because its making changes to sector addresses on the hard drive that have not been assigned to a file yet.
OR, Does it simultaneously write to the file, sector by sector, while updating the table of contents sector start/sector end? If this is the case, I probably need to run chkdsk because a fragment of a file would have been written and the system would not know what to do with it.
OR, (And I really hope not), the compression occurs on the drive that files are being copied from (in one of the manners listed above, 1,2,or 3) and then transferred to the drive specified to compress to in the command line... If this is the case and in the manner of 2 or 3, I suppose I would have to run chkdsk on the source hard drive which would be the worse case scenario.
Thanks again for any possible help or advice on this.
Log in to post a comment.