Question

  • Creator
    Topic
  • #2211404

    Offsite backups and other issues within same…

    Locked

    by smartacew0lf ·

    Carbonite doesn’t seem as though it is going to be suitable…
    A brief sysnopsis of the details.

    Have 2 servers, neither serve the role of PDC, both use Server 2003 and either or both servers are running Quick Books, SQL Server 2005, and a couple of proprietary applications related to the Freight industry. Users do not have accounts on the machines, they do however have shared folders on either or both servers.
    Each machine has aproximately 25G each in total drive usage for System and Data. This usually equates to approximately 16 to 18G of backup files per full backup to per machine.
    I have been working on trying to implement a reasonably efficient Offsite backup solution. Have tried Carbonite. Here is the stumbling block with online backup services. We have T1 access to the Internet. The size of my pipe to the net severely limits that which I perceive to be an acceptable backup policy using Online services. All of the above inspires several questions.
    1.) Given the total of almost 50G total uncompressed data to backup between the 2 machines, I have my doubts about our ability successfully upload the entire weekly backup within the time constraints of the weekend. My reservations regarding the actual implementation of a total system recovery via an online service are multiplied even greater. Am I missing something here? Does anyone else see a viable solution within my limitations?
    2.) Does anyone know of a relatively trouble-free methodology of maintaining an offsite backup?
    3.) And finally, for efficiency, I have pursued this objective with the idea in mind of implementing it by utilizing the same backup policy in use locally. Should I perhaps create an entirely different backup job to more efficiently preserve the amount of data necessary to upload to such a service?

    A couple of last minute bits of background information regarding our current backup policy. Each machine currently runs a full backup policy nightly to its own dedicated external 500G drive, deleting older jobs as necessary to create free space. One machine uses NT Backup via a front end package. (Backup Assist) and the other uses Backup Exec.
    Thanks in advance for your time and help.

All Answers

  • Author
    Replies
    • #2864449

      Clarifications

      by smartacew0lf ·

      In reply to Offsite backups and other issues within same…

      Clarifications

    • #2863615

      Hybrid

      by rkuhn040172 ·

      In reply to Offsite backups and other issues within same…

      Use external drives for say weekly offsite backups and then use Carbonite for say daily differential backups. That should greatly reduce the amount of data being backed up and make it more realistic.

      • #2862031

        What leaves me apprehensive about this…

        by smartacew0lf ·

        In reply to Hybrid

        Is that in the event I needed to do a full recovery I would need to get every differential backup made for the week up to the point of the disaster in order to do a full backup. Not sure what kind of lag time would be involved. I do know this though, given the time ever presents itself, I will be under some great pressure in getting it all back up and running as efficiently as possible. I reckon I should setup a “lab” test just to see what kind of timeline all of this would entail.
        I have just about decided to do the greater brunt of our offsite backup via external drive and caddy setup though. Another issue I have observed with this situation is that in using Backup Exec, I have noticed that all of the differential backups are nearly equal in size to the main backup. Does this seem proper? If so it begs the question, “Why even bother with differential backups as opposed to full?” Thanks for your input sir.

        • #2862020

          differential versus full, and incremental

          by churdoo ·

          In reply to What leaves me apprehensive about this…

          Differential backup uses the archive bit of a given file’s properties to determine if the file should be backed up, so your full backup needs to reset the Archive bit.

          Check your Full Backup job definition and make sure it’s a “Normal” type backup and not a “copy” as the normal resets the Archive bit, and will prepare for a proper differential (or incremental) backup. As files are modified during normal operations after a Full backup, the Archive bit is set, and that’s how the differential backup job knows what to back up. Also, understand that a differential backup is any file that’s been modified since the last FULL backup, so really you only need the last full backup and the most recent diff backup to restore to a point in time.

          Incremental vs differential: incremental is those files that have been modified since the last full OR incremental backup (incremental resets the archive bit as well, i.e. if a file is changed on Tuesday it is backed up Tuesday and if the same file is NOT changed on Wednesday, then an incremental backup will NOT backup the file on Wednesday night, whereas in a differential strategy the file will be backed up Tuesday, Wednesday, and every day thereafter until the next FULL backup. With an incremental backup strategy, you need the latest full backup plus EVERY subsequent incremental to restore to a point in time.

          Anyway I’m not sure why I launched into that lesson about full, differential, incremental backups, but getting back on your original topic, you have some very valid points. At 1.544mbps inbound, in a restore scenario that is not enough bandwidth to bring down a full restore in a reasonable timeframe with, as you said, everybody breathing down your neck screaming “DONE YET?! DONE YET?! DONE YET?!”.

          Perhaps with a broadband service which may give you 6mbps or 10mbps or even more inbound bandwidth, might be practical in a full restore scenario, so short of managing removeable media, maybe the answer is to get a broadband service as a BACKUP to the T1 in the event of a disaster/restore scenario. With this backup circuit in place, not only would it be available to pull down a full restore in a reasonable timeframe, but the rest of the time you could potentially utilize it for non-critical operations and alleviate some burden from the T1.

        • #2861925

          Thank you so much Churdoo…

          by smartacew0lf ·

          In reply to differential versus full, and incremental

          I am confident that with this knowledge I can go back and figure out exactly what I need to do with the Backup Exec issue I spoke of. Very good explanation.

    • #2863612

      Send physical backup to service

      by wdewey ·

      In reply to Offsite backups and other issues within same…

      The better services allow you to send them an initial backup on some sort of physical media (External HD I believe). From that point on they do a differential and use that to build a new full backup. As long as you don’t have too much data that changes every day then you should be able to complete a differential over the T1 every night which will give you a full backup. The issue comes if you need to do a full restore in that it could take several day to do. I don’t know if any of the service will put the restore on to a hard drive and ship it back to you.

      Bill

    • #2862013

      suggestions

      by lei fan ·

      In reply to Offsite backups and other issues within same…

      Local Backup then pick up the most critical files for offsite. T1 should be fine, most of the offsite backup like MozyPro should do incremental, so the first time will take long(But again it can run in the background, you can specify bandwidth throttling as well) but after the full copy made, the future backup will only pick up the changes.

      • #2861874

        Onsite – Waterproof/Fireproof safe?

        by jasonjb1222 ·

        In reply to suggestions

        Is an onsite, waterproof and fireproff safe not an option?
        Put your tapes in there, with a rotational cycle. Just make sure to test your backups every now and again.

        • #2861727

          Yes it is, however…

          by smartacew0lf ·

          In reply to Onsite – Waterproof/Fireproof safe?

          As we use Hard Drives for backup purposes it still requires (or depends on) human intervention. This is compounded by the complacency of most employees in fulfilling that role. The good news is that I have persuaded the CEO of the company to take on the responsibility of that role himself. I am setting up a hard drive caddy system in his office and combining that with online backup as well as the fire safe thanks to some of the tips I have received here. Thanks for the tip Jasonjb. 🙂

Viewing 3 reply threads