Tape Labeling in AMANDA

Alright, I’m once again appealing to people with more experience than I have.

Waist deep in learning AMANDA, I have questions about tape labeling. I’m attempting to find the best practices for labeling tapes, both physically (via barcode stickers on the tape) and the “logical” label written using the ‘amlabel’ command.

Anyone who knows better, please correct the following overview: For anyone who doesn’t know (this included me 2 hours ago), AMANDA uses a program called ‘amlabel’ to write a label to the beginning of the tape so that it always knows which tape is inserted. It keeps records of every tape that it ever writes, so that in the future when you attempt to recover data, you can insert the specific tape that it asks for.

Since this electronic label can be in nearly any format, many people that I’ve seen commenting in various places have the knee-jerk reaction to label their tapes “Monday”, “Tuesday”, etc etc. Many of the old-timers discourage that, because if a tape runs over, you’ve screwed up the labeling system. Many people suggest “Daily1, Daily2, Daily3”, etc etc.

My question is this: What are your tapes labeled, and do your AMANDA labels match the physical labels on your tapes? Also, if you have a barcode reader, do you order customized labels to take advantage of that, so that your barcode matches your AMANDA label, or do you have some sort of mapping between the two in a spreadsheet (or AMANDA itself)?

By the way, if you’re looking for customized labels, I found that https://www.labelarchitect.com has some very configurable options. Color, # of characters, etc. Feel free to share your sources if you’ve got some suggestions!

  • David Magda

    We have a ‘Weekly’ pool (AWF, BWF, CWF, DWF), and an ‘Offsite’ pool (NV8, DE8, JA9, FE9, MR9, AP9, AY9, etc.).

    This week we’re sending data to the CWF weekly set (it’s the third week of the month), and come Monday I’ll switch to the offsite set that will be labelled MR9. When March 30 rolls around the MR9 set will go offsite, and the AWF set will be put in the library.

    So the LTO barcode labels are CWF### for this week, where ### starts at 001 and goes as high as you need it to (lately we’ve been reaching the low seventies). And on Monday night the first set of offsite will go to MR9001 (I’ll rotate the sets when I get in on Monday–baring a major disaster I have to deal with).

    The MR9 will then go to an offsite provider for a period of time, and will recycled in the future. At the end of April I will take an old offsite set and put labels on it that say AP9 and relabel things. In 6-12 months the MR9 set will be brought back in and it too will be re-stickered and relabelled in software.

    Depending on far you want to be able to go back, and how many tapes you want to purchase, you can have your Offsite pool be offsite for as long as you want: 1 month, 3 month, 1 year, 2 years, etc.

    To minimize wear on the weeklies (A, B, C, DWF), we use them for a year, and then turn them into month-ends (buying new media for the weeklies). Every 3-4 years we upgrade tape drivers (LTO-1 to -2, to -3, etc.) and simply order a new batch of tapes.

    LTO is nice since it can read two generations back, so you don’t have to clone volumes to the new media to be able to access your old save sets (to use the NetWorker parlance).

    I’m not sure if this is the best or most efficient method, but it’s what I inherited and it seems to work pretty well. I’m hoping for a VTL this year, so we may be able to re-jig things a bit.

    For colour, have the ### white-on-black for easy reading. Have the prefix something unique, consistent across the entire prefix, i.e., all the letters in “AWF” are a single colour, as would all the characters in “MR9”–‘M’ is not different colour than ‘R’. I’ve found that have each letter or number a different colour (1s are always blue, 2s are always green, etc.) really annoying and hard to read–prefix one colour, sequence number another, is the best balance IMHO.

    Before we completely de-rack a machine we power it up one last time at the end of the month to get a full month-end back up of it. That way if five months from now someone notices some critical data we need from it we can go back and restore it. If we know ahead of time that a machine will be retired, we may make an extra one or two ‘full’ backups to the Offsite set, and then it’s “safe” to retire it mid-month.

  • Matt

    wow, David, thank you so much for that. Your post is exactly the kind of response I was looking for: In depth explanation of a policy that a lot of people can use.

    I really do appreciate the time it must have taken to type this in.

    It really seems like a good, solid plan. You’ve got a lot of tapes at this point to go through. How often do you find yourself digging back to find old data, and have you ever had AMANDA (or your backup solution) fail?

  • Josh

    Our set up is not as complex as David’s above, but it’s served our university department well for the five years that I’ve been here.

    We run tape backups twice a week, always level 0. Each run takes two tapes out of a pool of about 30 tapes (not counting archived tapes).

    In the past, before we had a barcode reader, we let Amanda automatically write the electronic label onto new tapes. Our labeling scheme is PREFIX###. Whenever a new tape got an Amanda label, we would print out a sticker to match the Amanda label.

    Now that we have a barcode reader and preprinted barcode stickers, we manually set the Amanda label to match the barcode number whenever we get a new set of tapes.

    At the beginning of every month, we archive the most recent backup run, and mark the tapes as no-reuse with amadmin. We keep archive tapes for two years, so each month we also take out any older tapes out of the archive, mark them as reusable again, and put them back in the active pool. This way, barring tape failure, our active pool stays at a constant size.

    Amanda does a good job scheduling tape reuse, so we don’t try to force it to go in any particular order. For example, our tape changer is currently holding tapes 10, 24, 48, and 58. Every Monday, we take a look at what tapes Amanda expects to use and load up our tape changer accordingly. If we ever need to look up the tape used on a particular date, we can either use “amadmin find …” or check the tapelist file.

    It’s a pretty low-maintenance system, and so far we haven’t had any trouble with it. Let me know if you have any questions.

  • David Magda

    Not that often, though it has been convenient a few times.

    This way dates back to previous company ownership (we got bought out so things may change soon), and Legal wanted data going back a while back. So we ended have each month-end be offsite for two years. Probably a bit excessive, but that’s what the higher ups wanted and were willing to pay for. (They actually wanted Offsites to be around “forever”, but when we gave a price for new tapes every month and offsite contracts, they went with two years.)

    So I’ve got two years’ worth of tapes for the Offsite pool, plus four sets for the Weekly pool: about 3,000 tapes all together. Too much really.

    We’re still using LTO-1, so it’s about 100-200 GB per tape depending on the data type. We could shrink that a lot by using a newer LTO generation, but before we upgrade the tape drives we need to get a VTL or something. Feeding LTO-1 is easy (15 MB/s), LTO-3 and -4 can’t be realistically streamed straight from the client IMHO: you need to go to disk first, and then perhaps clone to tape.

    We’re using NetWorker right now, because when all of this was set up it (and NetBackup) were really the only multi-platform games in town.

    Our new corporate overlords have NetBackup as a corporate standard, so as part of the upgrade that will (hopefully) occur in the next year we’ll probably go to that (and keep the old system up for a little while for ‘archival’ purposes).

    When deciding on a policy it really has to come from Business, as generally it’s their data that you’re backing up. If they want two years’ worth of backups (“just in case”) then you calculate the cost and hand them the bill for the tapes (drives are a sunk cost, but you should plan ahead for generational upgrades every few years I guess).

    Also, even though we have an Offsite pool, it doesn’t actually have to be offsite for as long as you keep the set. So even if we keep the March 2009 back up set around for two years, it doesn’t have to be at Iron Mountain (or where ever) for two years. You can send it away for (say) six months, and on the seventh just put it on a shelf someplace (so you can go back).

    This way you have the long-term data around for a while, but don’t incur the full cost of keep it offsite. In a DR scenario you’ll probably only want the last few weeks’ worth of data, which you can call up and order back as part of your DR scenario.

    Though if you do keep media on-site, it may be worth getting a data-rated safe. LTO tapes have a storage range of 16°C – 35°C (20-80% RH), but most “data safes” are only UL Class 125 or 150 (internal temperatures can reach 125F/52C or 150F/66C to meet that standard), so double-check with the vendor about how appropriate these things actually are.

    Most hard drives are rated for a max of 70C storage, so the two Classes are fine for them (in case you do an rsync to USB/FW drives).

  • Larry Olin Horn

    My recommendations:

    Physical and logical labels should match. Maintaining a cross-reference is unnecessary work.

    Meaningless IDs, say “T1001”, “T1002”, …. That avoids overflow issues and future angst when you discover that your labeling scheme no longer matches your current practice. Some would also consider that a security (by obscurity) feature.

    Re-use tapes at random from the scratch pool, so each tape will likely get varying usage patterns; for example, incrementals only ever using the first 20% of a tape (if that applies to your type of backups).

  • Anonymous

    We bought a Brother Dymo label writeer who allso does barcodes. One of the cheaper models.
    So we can set the barcode value to oure choice.