Dismiss Notice
I hate the colors. What do I do?

At the far bottom of the page, on the left, is a menu or link that says, "Forum Default." Click on that and choose a different Style.

Backups. *Sigh*

Discussion in 'Pete's Computer Corner' started by pete, Sep 6, 2014.

  1. pete

    pete Brassica Oleracea Staff Member Administrator

    I am relatively enthusiastic about backups. I tend to do full backups at least once a month for everything. I've got at least 5 terabytes of hard disk space -- probably more -- lying around my house. Every month, I also back up this website and my own website because I don't trust the automatic backups that the webhosts use. Hey, IRL I essentially get paid to be paranoid about computer stuff.

    So, yesterday, I'm cruising my website and I see a broken picture in my Yanagisawa album. No biggie. I'll delete it. I do. The entire Yanagisawa album goes *poof*. About 5000 pictures. Joy. Well, I did a backup a week ago, but let's see what the webhost's most current backup is. Ah. Two minutes ago. Maybe I'll be in luck ....

    Nope. Probably what happened is that the backup was happening at the exact same moment as I was browsing my Gallery.

    OK, so I'll see if there's a webhost monthly backup. Nope. From February.

    *Sigh* OK. Let me check my last month's backup I downloaded. Erm. Where's the Yanagisawa album? Did it drop into another directory? *Poof*?

    Last weekend, I uploaded about 2500 new pics into my Gallery, so I really have no desire to have to re-upload all my raw files (which would be my "last resort" backup). Is there something else I can do? Yes ... there's a utility in Gallery 3, the software I use, that can automagically rebuild all your albums. Also, if I do the rebuild, the little notes I put in some albums, like the sale price, etc., should be preserved, which isn't in my "last resort" backup. Sounds like a plan, then.

    So, I start the utility. Says I have 18,500 photos and it's rebuilding ... two files a second. That's 6.4 days. I've also got about 1/2 hour or so I've got to use to reset permissions on my webserver's directories.
  2. Helen

    Helen Content Expert Saxophones Staff Member Administrator

    OMG... I'm so sorry to hear that Pete. What a pain in the ass.

    The same weekend you did all your work on your gallery, I was busily working on mine. (Am still working on it as a matter of fact). I would scream if everything that I did were to just go "poof".

    I don't have nearly the back-ups that you do. This post should be titled "A cautionary tale".
  3. pete

    pete Brassica Oleracea Staff Member Administrator

    Right now, I'm running an automated cleaning routine on my picture database. The "two pictures a second" number became about a picture a minute. So, after the clean -- which looks like it'll be done tomorrow -- I'm going to go back to my raw backup. That might take another day to re-upload -- it's close to 8000 pics (really!). I'll have a few broken links on my blog, but that's not too bad.

    The raw files I have are named in the form of:

    A-901 Eb Alto - Lacquer - sn 00123456 (1980)

    So, when they re-upload, the albums will be named right. I'm going to have to manually fix some of my text captions. That'll be an ongoing project.

    SOTSDO Old King Log Staff Member CE/Moderator

    It helps to be paranoid when dealing with backups. I've been doing them on my computer system since the early 1980s on critical stuff (financial, personal data and the like), but I got into them in a huge way when I started my music group.

    Some of you may know of the fire that destroyed the entire music library (the only copy that existed) of the Charlie Barnett Orchestra back in the day. His group lost everything (instruments, charts, equipment), and only was brought back from the dead by a number of other groups cooperating in loaning instruments and charts.

    Having read of this, I resolved to ensure that my music (the only really hard to replace item belonging to a musical group) was thoroughly protected. From our very beginning (an ancient jazz band library, with many stock hotel arrangements, augmented by a paltry twelve original charts, all that I could get arranged in time for our first appearance), I have "backed up" my music library with full "fair use" copies kept in manila envelopes, kept in turn in large plastic filing tube. Lose a chart, and it's simple as pie to pull the envelope, pull the chart, and create a new copy.

    (All of the originals are preserved as well, but only in their original form. We have boxes and boxes of them in storage right now; the whole assembly would fill a good sized van if transported all at once.)

    Once things were stable, I also added a digital backup of the scanned PDF files for each and every arrangement. Starting small at first, we have slowly built the collection up to a total of eight CDs worth of scans. Copies of these CDs are kept in the safe at home, and in the bank safe deposit box.

    Then there's the "originals" kept on the computer, for instant access at a job when someone (by which means, I know not - we usually pull all of the charts, but invariably someone will have a missing one a half hour before the downbeat) comes up short. With the laptop and the portable printer and tape up board, a new copy is ready in a couple of minutes.

    But, backing up onto CDs is cumbersome and takes a lot of time. I also have everything on a hard drive (kept separate from everything else) and on a "wart" USB drive that stays plugged into the laptop at all times. When a new chart is readied, a copy goes into the "Music Scans" file on the laptop, and a duplicate into a similar file on the wart. With a capacity of 16 GB and a library size currently at 7 GB or so, there's still a lot of room left.

    The only other things that I back up with similar regularity are the financial files and the contact information, the files used to build the website and one or two other personal items.

    We've talked about dumping the paper copies along the way - they currently fill sixteen long filing boxes, boxes of a type that we cannot find any longer. Maybe after the move...
  5. pete

    pete Brassica Oleracea Staff Member Administrator

    As far as music charts go, I'm kinda surprised I don't see more folks scanning the files to PDF and bringing their iPad or other tablet to a performance. Heck, the tablet has its own built-in light. You can also buy an original iPad for around $125 or a new (junkier) Android tablet for under $75.

    I've got one of those scanner/printer/copier thingies with a sheet-feed scanner. I really should scan my stuff.

    SOTSDO Old King Log Staff Member CE/Moderator

    Reading music off of an iPad-sized screen may seem like a good idea, but unless your eyesight is extraordinarily acute, you are going to have real troubles making it out.

    When we cruised last fall around this time, the ship (one of the Disney monsters) possessed almost zero "live music". A couple of very skillful duos (with automatic chorusing software), plus a pair of soloists.

    The guitar soloist did all of her work from memory, but the pianist had to have charts for her stint in the grand foyer. She used a standard sized iPad, and she had to lean over and squint to make things out. (She was a poor pianist, for what that was worth.

    My alto player has one of those music tablets (as seen in the pages of the International Musician), and while it will display one page at an almost readable size, page turns are cumbersome at best, and DC/DS maneuvers even moreso.

    Harry Connick did it right, with full sized computer screens (all Apple equipment, from what I've been told) and a centralized computer running things. However, the amount of equipment is/was formidable, and one computer fault and you're sitting there without the "soul" of your musical group.

    Paper isn't all that modern, but it's still the most foolproof of the technologies that we have. I'm even in the process of rendering all of the Comstock charts that we have (almost all of which are single pagers) from the book that they came bound in into individual charts. Not as convenient of a container, but a lot easier to shift things around for a gig.
  7. What a pain, Pete. It must be hugely frustrating to have done so much, with probably lots more to do, and have to spend time recovering a lost position.

    WRT paper v. digital, I scan my music to PDF and use iPad with Forscore app.

    Advantages are - all music for saxes and clarinets are in one place and easily accessable; backlit for performances; foot pageturner by bluetooth is a half a page at a time, so next page can appear at top of screen while still playing bottom page; easily crop music to fill screen maximally; does not blow off the music stand in the wind.

    Disadvantages are - not good in the rain; not very good in direct sunlight; need music stand slightly closer to read from, but I use glasses with a prescription specifically for my music stand distance; if many of the band have a difficult page turn all at the same time, I can be the only one still playing for a few bars...

  8. tictactux

    tictactux Distinguished Member Distinguished Member

    One of the (for me) important criteria when selecting a "web page engine" (whatever you call your content management system) was the back-end: database? flat file?
    Within "image album systems" there's the question whether or not the system stores the pictures as DB objects (BLOBS) or just references a file within a folder structure.

    Back in the day when Microsoft's SharePoint Portal server was young I had the questionable honour of maintaining this sorry excuse for a content management system. When I discovered that everything, yes, everything was stored within a huge monolithic database I instantly knew that this was not suited for web pages that are hosted someplace else where you don't have physical access to the system.

    So for my own needs I settled for a plain old flat file system, every page content (not its form!) is stored in human readable form as a file, every file and every image exists "as is" within the file system. Backups are a breeze, as is portability - I just copy the whole tree over to a new system and off I go. Performance is not necessarily worse as file system caching is often better than DB caching, at least on shared systems. Bulk editing is easy, all you need is an editor capable of doing search/replace work in any number of open files.

    Databases are evil, I tell you.
  9. pete

    pete Brassica Oleracea Staff Member Administrator

    RE: iPad stuff, you also have to remember that the Retina Display iPads are significantly more crisp than the non-Retina ones. You also can make the print bigger. In the somewhat near future, Apple is supposed to be bringing out a 12.2" iPad. The current "big" one is 9.7". However, they didn't announce that today: bigger/new phones, the Apple Watch and their credit card/payment thing.

    RE: Databases, Gallery 3 has a MySQL back-end, so a fully-relational database. Photos are stored in separate directories, with pointers to those directories, so no BLOB. However, I don't have that much of an opinion on evil databases, other than to say all software sucks. It just sucks in different ways. In this case, it's obvious that deleting something when a backup was going on was the problem. It's unfortunate I wasn't told when said backup would have been running. If I paid more for the service and/or was losing money because of this, I'd be mad. Considering I'm not, I'm just kinda annoyed.
  10. pete

    pete Brassica Oleracea Staff Member Administrator

    And that is that. I restored the entire section. That was fully 1/2 of my picture gallery. I'll fix some of the attributions over the next few days/weeks.

    Something should be said about restoring from backup. For my home computers, I have external hard drives that are the exact size (in both gigabytes and shape) of my internal hard drives, except that they're mechanical drives, rather than solid state (SSD). So, if my computer's hard drive fails, all I have to do is shut down, pop in the backup and I'm good to go. I'm never more than a week out of data. In this case, as mentioned, I had to a restore from a completely different medium -- a backup hard drive to somebody else's webserver. That was a very slow process. The restore did approximately one directory of pics ("albums") every 5 seconds. Let's see. I have 31 models of Yani. Each of those model directories has an average of about 15 horns (i.e. subdirectories). That means 15*31 or 465 directories. That's about 1.6 days. That sounds about right. At least it was a semi-automatic practice and I could do something else, provided I still had a Chrome window open to view the process.

    Lesson: even if you have good backups, it'll take you awhile to recover. You need to budget this time. Especially if we were talking about a business. FWIW, one of my various computer certifications is "Security+." This topic, disaster recovery, is heavily emphasized.

    A lesson I learned is that I need to start being a bit more consistent in how I name my picture galleries. While I really love the naming format I use, the album titles can get too long. I'm probably going to stop writing out things like "A-900 Eb alto" and just leave it at "Eb alto." I'm also going to stop using the word "unknown" -- which I use in reference to serial numbers -- and use "unk" or some variation thereof. Or stop using it altogether, except if I have multiple albums that end up like "A-900 Eb Alto - Lacquer - sn unk1." I have to have a unique name for every horn's directory or the Gallery 3 software spits at me.
  11. SOTSDO

    SOTSDO Old King Log Staff Member CE/Moderator

    Any large project like an extensive website or a retail merchandise tracking system can lead you down this path if you don't plan ahead. I learned ("learnt") a long time ago that any extensive project, like a website or my music cabinet (which really looks magnificent now, with all of the labels in place), requires at least 10% of the time spent planning rather than doing.

    Once you lay things out (on paper for me, other may prefer doing it on the computer), you can cypher out the best way to name objects in the program, the best way to cut and finish wood, and so on.

    In effect, design is arguably the most important phase of any project. I know that we all tend to just start in doing something, and that you and I are not the only ones who have learned/learnt this the hard way. But, getting the message out (like what you are doing by describing what you are doing) is a good thing for others.
  12. Helen

    Helen Content Expert Saxophones Staff Member Administrator

    +1 on the planning and naming... My gallery is behemouth with its 21787 images in 2270 categories, and yes, some of the names are crazy-ass long.

    When I planned out my gallery, I thought I did a good job. (I spent hours plotting it all on paper using flow charts.) But then as the gallery grew in scope and size, I realized what I had was not really adequate. Now I'm facing a dilemma on how to retool it a bit. Then I read about your nightmare Pete, and I get the cold sweats wondering how exactly I should be protecting all my work. :-( ... A rhetorical question, but nonetheless one that gnaws away at me in the quiet moments...
  13. pete

    pete Brassica Oleracea Staff Member Administrator

    +5 billion on the planning and naming. I will qualify, though, that my original plan was to have a few really good pics of a few horns. It got out of control when I realized that I better start archiving pics of, say, Virtuoso Deluxe-finished Conns, because no one else does.

    I'm thinking about the backup idea. What I'd like to do -- probably -- is have one really good, stable, working photo gallery on one server (thesax.info) and a month old backup that people can't see here (woodwindforum.com). So if one goes *poof*, I can just slap a redirect HTML in http://thesax.info/gallery and you'll come over here. I don't think this would cause a bandwidth problem. Helen, you could do the same. That would eliminate the problem of missing comments, captions and/or other attributes. I've done a variation of this when I was the techie geek at SOTW: I just duplicated the SQL databases and monthly backups of everything else. I couldn't do a full duplication because of lack of space and bandwidth.

    Another option that's not as "fault tolerant" is to make a copy of everything on one server. The reason I haven't done that is because I think that my webhost would boot me out for using too much space. They might not mind that much, anymore. As I've mentioned, storage space is now dirt cheap. Server-class 4TB hard drives are $150 or less. Hey, 1TB SSDs (solid-state drives) can now be found in the $450 range. Imagine a RAID server with SSDs. Mmmm. Speedy.

    Anyhow, I did get everything working again, for a brief time. I then made a backup and tried to do something that's a good idea: update. Update, of course, failed and corrupted some files. I'm now restoring from backup. Again. *Sigh*
  14. Helen

    Helen Content Expert Saxophones Staff Member Administrator

    Well all this talk about back-ups prompted me to do a much overdo manual backup of my entire site yesterday... It took all day. I started the process around 10:00 am and final file transferred at around 7:00 pm. No, I don't have a slow Internet connection. As a matter of fact I have one of the fastest connections available in our region. It took so long because my site is so friggin' big... It was around 15.7 GB of files that were copied to my PC! (94,369 Files, 8,976 Folders, according to the info in the Windows Explorer folder). :eek: My first PC had a hard drive that was 4.5 GB.

    I was thinking yesterday about how technology has just moved at the speed of the bullet train in the last decade. Ten years ago it wasn't uncommon to still see 3 1/2" floppy disks around offices, which of course held a grand total of 1.44 MB (slightly less in reality) of material. Nowadays our digital cameras routinely take images that are more than 3 to 4 times the size of that.
  15. saxhound

    saxhound Moderator Staff Member CE/Moderator

    Storage prices continue to amaze me. You can now get a 5 Terabyte USB 3.0 drive on Amazon for $175. Even though it is a much slower connection than SAS or SATA, you can set it up to copy overnight or whenever it's convenient. Not a bad "belt and suspenders" approach to load it up and take it off site to a safe deposit box or a friend's house.
  16. pete

    pete Brassica Oleracea Staff Member Administrator

    I'm now starting to see 6TB drives coming out. A desktop-class one is about $275.

    There is a kind of hard drive that uses the PCIe slot in your computer -- it's what the new Mac Pro uses. A really fast SSD (Solid State Drive) has a random read/write speed of around 30,000 IOPS. A really fast PCIe drive is around 135,000 IOPS. One caveat: a "cheap" PCIe drive that will hit those specs is $650 for 250gb -- and not all PCIe drives work with all computers. You can buy a really fast 1TB SSD for under $600.

    Regarding website backups, I rather like the system the host that the WF is on has: you press a button to start a backup a compressed file to the webserver, then you can download the backup via FTP. If you want. It takes maybe 2 hours total to do this. However, the backups here are around 3gb, compressed. The host I use has you press a button and the backup downloads automatically from your browser. My website's around 6gb, compressed, and it takes about two hours to do.
  17. Helen

    Helen Content Expert Saxophones Staff Member Administrator

    Yah, and if I could figure out GD's user panel, maybe I'd know if they feature the compressed download option as well.

    I literally just got off the phone with one of their tech support guys due some 500 errors I'm getting on the back end of my blog. During this discussion I found out--after nearly 2 years with them--that they have a 30 day backup of all my sites that I can roll back to. Who knew? Not me, but then their user panel still frightens and confuses me...
  18. pete

    pete Brassica Oleracea Staff Member Administrator

    I'm fairly sure most webhosts have a full monthly backup, if not weekly, and incremental daily backups. You sometimes have to ask.

    The GoDaddy control panel is very minimalistic. The WF, on KnownHost, and I, on Bluehost, have a really comprehensive version of Cpanel. I've got at least a dozen more options than we have at the WF, but mine are mostly in the form of one-click installs for things like phpBB, ZenPhoto, etc.
  19. SOTSDO

    SOTSDO Old King Log Staff Member CE/Moderator

    Dammit, Pete - speak English...
  20. pete

    pete Brassica Oleracea Staff Member Administrator

    phpBB is a forum program. We used to use it on Sax-on-the-Web. ZenPhoto is photo gallery software, somewhat similar to the software I mentioned earlier, but with fewer options. It really doesn't matter. These were just examples.

    A one-click-install is exactly that: click one button and everything installs for you. Well, mostly.
Our staff's websites: