vBulletin Search Engine Optimization
| |||||||
| Register | FAQ | Members List | Calendar | Search | Today's Posts | Mark Forums Read |
| ||||
| Currently have SCO Enterprise 5.0.6 running with some production software installed on a Compaq ProLiant. We have just aquired a second hard disk the same size as the first and wondered if it was possible to do a software mirror of the current hard disk. The machine does not have a raid controller. We were thinking maybe there is something we can put into the crontab to run overnight which would do a full mirror of the hard disk. Thanks for any help Martin Roclawski |
| |||
| In article <3f0bfe7a@news.userve.net> "Martin Roclawski" <m.roclawski@XremoveXdenisonmayesgroup.com> writes: $We have just aquired a second hard disk the same size as the first and $wondered if it was possible to do a software mirror of the current hard $disk. You could add Virtual Disk Manager to your server. This is software RAID and is an optional (i.e. extra-cost) SCO product. I've never used it at all, let alone tried to migrate to it from a single-disk configuration, so I can't tell you if there are any gotchas to be aware of. $We were thinking maybe there is something we can put into the crontab to run $overnight which would do a full mirror of the hard disk. Sure, you could do something like that, but there are gotchas with most of these approaches, too. If the new disk is no smaller than the first one (not all "18.4 GB" drives are exactly the same size, for example), you could look at using dd to make an image of the whole first disk onto the whole second disk. Since the whole disk includes things like the masterboot block and partition table, you should (at least theoretically) be able to make a duplicate disk that will be bootable, though I haven't tried this. One possible gotcha with this approach is that if filesystem data is changing while you're doing it (even something as simple as a nightly job that cleans up old files, or an email message being received), you could end up with corruption on the second drive, because you could (for example) copy the inode table before the data changes, but the directory entries and the actual file data after the change takes place, and then you will have inconsistencies that could make some files vanish or be corrupt. You could also set up filesystems on the second drive that match what's on the first drive, then copy the data at the filesystem level (e.g. using cpio -p). You won't end up with metadata problems on the new filesystem, but if you're in the middle of copying a file while it's being updated, the file data could still be messed up, and since different files will be copied at different times, you could have problems if the data in two files (e.g. your customer list and the customer order history) is supposed to match but one got updated before the copy and one after. You'll also have a bit more work to do to get the second drive set up and bootable. You could set up a job that restores your nightly tape backup to the second drive. This has similar problems to the file-by-file copy approach. -- Stephen M. Dunn <stephen@stevedunn.ca> >>>----------------> http://www.stevedunn.ca/ <----------------<<< ------------------------------------------------------------------ Say hi to my cat -- http://www.stevedunn.ca/photos/toby/ |
| ||||
| Cheers for the info Steve. It seemed alot easier when discussing the concept with my people here!!! Stephen M. Dunn <stephen@stevedunn.ca> wrote in message news:HHrzv5.28G@stevedunn.ca... > In article <3f0bfe7a@news.userve.net> "Martin Roclawski" <m.roclawski@XremoveXdenisonmayesgroup.com> writes: > $We have just aquired a second hard disk the same size as the first and > $wondered if it was possible to do a software mirror of the current hard > $disk. > > You could add Virtual Disk Manager to your server. This is > software RAID and is an optional (i.e. extra-cost) SCO product. > I've never used it at all, let alone tried to migrate to it > from a single-disk configuration, so I can't tell you if there > are any gotchas to be aware of. If software raid works like hardware they have to be in an array prior to attempting to rebuild one from another. However software might be different. > > $We were thinking maybe there is something we can put into the crontab to run > $overnight which would do a full mirror of the hard disk. > > Sure, you could do something like that, but there are gotchas > with most of these approaches, too. > > If the new disk is no smaller than the first one (not all > "18.4 GB" drives are exactly the same size, for example), you > could look at using dd to make an image of the whole first > disk onto the whole second disk. Since the whole disk includes > things like the masterboot block and partition table, you should > (at least theoretically) be able to make a duplicate disk that > will be bootable, though I haven't tried this. One possible > gotcha with this approach is that if filesystem data is changing > while you're doing it (even something as simple as a nightly > job that cleans up old files, or an email message being received), > you could end up with corruption on the second drive, because > you could (for example) copy the inode table before the data > changes, but the directory entries and the actual file data > after the change takes place, and then you will have inconsistencies > that could make some files vanish or be corrupt. > > You could also set up filesystems on the second drive that > match what's on the first drive, then copy the data at the > filesystem level (e.g. using cpio -p). You won't end up with > metadata problems on the new filesystem, but if you're in > the middle of copying a file while it's being updated, the file > data could still be messed up, and since different files will > be copied at different times, you could have problems if the data > in two files (e.g. your customer list and the customer order > history) is supposed to match but one got updated before the copy > and one after. You'll also have a bit more work to do to get > the second drive set up and bootable. The dd approach seems a possibility, especially if I can stop services through a script and then start them after the dump has taken place. The only thing that may cause a problem is try to figure out what is running in order to stop it!! May try experimenting with this first. > > You could set up a job that restores your nightly tape backup to > the second drive. This has similar problems to the file-by-file > copy approach. > -- > Stephen M. Dunn <stephen@stevedunn.ca> > >>>----------------> http://www.stevedunn.ca/ <----------------<<< > ------------------------------------------------------------------ > Say hi to my cat -- http://www.stevedunn.ca/photos/toby/ |