INTERACT FORUM

Please login or register.

Login with username, password and session length
Advanced search  
Pages: [1]   Go Down

Author Topic: Backup time  (Read 6550 times)

tuneup

  • Galactic Citizen
  • ****
  • Posts: 281
Backup time
« on: December 10, 2015, 12:54:57 am »

I just made my first backup of 973 GB of FLAC files from a Synology DS214play to a Seagate Backup Plus on USB 3.0 using a USB 3.0 cable. It took 14 hours. As this seemed overlong to me I called Synology tech support and they surprised me when they said they didn't know how long such a backup should take. Can a few of you please tell me how long (per 100 GB or per terabyte) your first backups over USB 3.0 have taken (even if it's a different NAS and different USB backup drive)?
Logged

JimH

  • Administrator
  • Citizen of the Universe
  • *****
  • Posts: 72438
  • Where did I put my teeth?
Re: Backup time
« Reply #1 on: December 10, 2015, 06:49:35 am »

You could look up USB 3.0 speeds on Wikipedia.  That seems pretty fast for almost a terabyte of data.

If you need to do it regularly, there are programs that will only transfer files that have changed (after the first time).  These have been discussed here.
Logged

robydago

  • Citizen of the Universe
  • *****
  • Posts: 518
Re: Backup time
« Reply #2 on: December 10, 2015, 07:54:17 am »

@tuneup

Are you sure your slowest link is the USB3 interface?

Last time I checked, single hard drives were slower than the USB3 speed.
On the other hand, a RAID config should be able to kill a USB3 link.
Anyway, your disk setup speed is what I would investigate first.
Logged

mwillems

  • MC Beta Team
  • Citizen of the Universe
  • *****
  • Posts: 5234
  • "Linux Merit Badge" Recipient
Re: Backup time
« Reply #3 on: December 10, 2015, 08:01:22 am »

If you were a) achieving maximum theoretical USB 3 throughput (640MB/sec)and b) doing a straight copy operation with a fast computer you'd expect to transfer a TB in about 30 minutes.  However, it's impossible that you're achieving item a) with a single spinning disk drive and b) is non-trivial in a NAS context.

Specifically, "fast" spinning disk drives might allow for throughputs of up to 200MB/sec or 250MB/sec under the right conditions, but those are typically the fastest the disk can possibly read or write and don't reflect actual sustained disk performance (especially for external drives).  Writes are also slower than reads, often by half. Realistically you can expect something more like 100MB/sec a realistic average write throughput for most external spinning drives (although many are slower than that).  So that bottleneck right there will take your TB data transfer from half an hour to more like 3 and a half hours.  A straight sequential file copy with an average spinning disk drive will take about that long (assuming that the copying computer is capable of supplying data fast enough to saturate the drive's write capacity).  I can confirm that that's about how long it takes for me (typically) to copy a TB of data to an external spinning disk drive over USB 3 (I do this regularly because my music collection is just over a TB).

However, that speed is the maximum "best case" speed if you're doing a straightforward file copy and your CPU can keep up with the I/O.  With a NAS that might not have the fastest CPU in the world, you should expect additional delays.  Furthermore, if your backup solution is anything other than a straight file transfer it will also take longer (does your backup solution use compression? does it de-duplicate?  does it do rolling checksumming for some reason?).  When I've done first time backups with complex deduplicating backup solutions (for example borgbackup for linux) it took me on average about 10 hours per terabyte, and this was with a desktop i7 handling the compression and deduplication.  

So bottom line, unless you use an SSD over USB 3 you're unlikely to be able to move a TB of data in under 3 hours in the best case.  If your backup solution is more complicated than "just copy stuff" and/or the CPU on your NAS is not super robust you can easily see that timeframe double or quadruple or more.  14 hours is slower than I would be willing to put up with, but your equipment may be functioning as intended.
Logged

blgentry

  • Regular Member
  • Citizen of the Universe
  • *****
  • Posts: 8014
Re: Backup time
« Reply #4 on: December 10, 2015, 08:09:54 am »

I use basic USB3 external drives in single enclosures. No NAS.  No RAID.  Nothing fancy; just a couple of external drives connected to my Macbook Pro.

I achieve somewhere between 50 and 90 MB/sec on these drives, sustained.  For long term averages, due to mixed file sizes and file system performance, I tend to see something closer to 50MB/sec... maybe even slightly slower.  But I haven't done a really big full backup in a long time; I do incrementals which are just a tiny fraction of the time.

So, the math on 50 MB/sec works out to about 180 GB per hour, or roughly 5.5 hours per TB.

Brian.
Logged

tuneup

  • Galactic Citizen
  • ****
  • Posts: 281
Re: Backup time
« Reply #5 on: December 10, 2015, 09:16:41 pm »

Thank you all for the responses.

mwillems, I called Synology and they said there is no compression, de-duplication, rolling checksumming or encryption in their backup software. I will get in touch with Seagate to find out their HDD's write capacity.

Brian, there is no RAID in the Seagate backup. The NAS is RAID 1 (2 mirrored drives), but I suspect this won't affect the backup speed.

robydago, by disk setup speed do you mean something other than RPM? If so, I'm not familiar with the term. Please give me a synonym or a little more explanation. I put Western Digital Red HDDs in the Synology. According to a review: "The WD Red NAS Hard Drive has a SATA 6Gb/s interface, 3TB capacity, 5K spindle speed, 1TB platters and 64MB cache."

Jim, I forgot to ask Synology today if the next backups will be incremental and how use anybody else's backup software to back up their NAS. I know it can be done because I saw a mention of using Apple's Time Machine.
Logged

jctcom

  • Citizen of the Universe
  • *****
  • Posts: 690
  • Rush - Styx - Yes - Porcupine Tree - Staple Food!
Re: Backup time
« Reply #6 on: December 11, 2015, 03:56:49 am »

Hi Tuneup.  Not sure if this will help because I am not sure if my backups now are incremental or if it is re-backing up everything.

However.  My backup logs show that my weekly backup for my 1.75TB music folders (Includes music videos and concerts as well) took the following times in the last few backups.

1 Hr. 7 Min.
1 Hr. 9 Min.
1 Hr. 4 Min.
0 Hr. 55 Min.
1 Hr. 6 Min.

I am running a Synology DS1512+ with a 5 drive Hybrid RAID array  (5 X 4TB WD RED drives).  Backup is to a 3 TB WD RED drive which is plugged into a 4 bay Startech USB 3.0 HDD Dock  (Model SDOCK4U33).

I am guessing that it is performing a full backup rather than an incremental only because the times are so consistently close to an hour and a bit and I do not add Music and stuff that consistently? also for the most part the times seem to be climbing at a rate that might be close to the rate I am adding new music / Music videos etc...?

I think the fluctuations such as they are might be due to torrents downloading or some other activity that might be going on at 3am. (When the backup of the music starts)

Hope that helps.  Let me know when you check with Synology again though if the backups are incremental or complete?  I would be curious to know for sure.

Carl
Logged
Carl's Music: https://cloud.clz.com/jctcom/music
Carl's Movies: https://cloud.clz.com/jctcom/movies

Some of Carl's Equipment:  Yamaha RX-A2A, i7-11700K, 128GB, PCIe X4 2TB M.2 SSD, GTX-970, SMSL DL200

Arindelle

  • Citizen of the Universe
  • *****
  • Posts: 2772
Re: Backup time
« Reply #7 on: December 11, 2015, 05:38:57 am »

I just copied slightly over 3Tb of flac files using a type of speed copying utility (4tb drives over USB 3 to a an external dock) the speed's show up and vary between 70Mps max at 105Mps (don't understand why it varies it just does :p ). I am using a reasonably fast i5 PC 3.1Gghz with the OS on an SSD with 8Gb of RAM. going from a green drive 5900RPM to a 7200RPM drive .. not sure if that would affect rates or not.

When using my normal backup software (Syncback) it does a double write verification and is incremental and it runs a LOT slower - around 30Mps which doesn't bother me because I'm not backing up all the files. When doing a full archive (just straight copying), I have Syncback verify the copied files just to be sure

So if running on a slow NAS or using any backup software which does verification, seems pretty much normal to me. Of course the OP can do incremental backups afterwards.

@jctcom - those have to be incremental  -- I'd say its just coincidence and you are writing playback info to the files or retagging your media.

PS- I tried an e-sata dock and I thought I might get some faster rates, but I didn't find it appreciably speedier than USB3 .. anyone know why?
Logged

dtc

  • MC Beta Team
  • Citizen of the Universe
  • *****
  • Posts: 3119
Re: Backup time
« Reply #8 on: December 11, 2015, 08:27:29 am »



...   When using my normal backup software (Syncback)  ...

Another vote for Syncback. I run it every night on all my music files, the MC library and other files.  Since I do not change much most days, it usually just takes seconds to run. There is a free version as well as a more advanced pay version ($40, I think).
Logged

tuneup

  • Galactic Citizen
  • ****
  • Posts: 281
Re: Backup time
« Reply #9 on: December 11, 2015, 05:53:07 pm »

Carl - I just spoke to Synology and all backups after the initial one are incremental. Given how long your incremental backups take, I can see where 14 hours per terabyte is normal for a Synology NAS.

Arindelle - Good to say Hello again. Everyone reporting faster speeds is backing up from a computer. Is it possible that the faster CPU in a computer compared to what is likely to be found in a NAS is responsible for the faster backup speeds? I put WD Red drives, which spec 6 GB/s so I don't think those are slowing things down.

In conversation with Seagate, they made the distinction between transfer speed, the info passing through USB, and backup speed, the slower speed of the backup software compiling the files and putting them to bed.
Logged

jctcom

  • Citizen of the Universe
  • *****
  • Posts: 690
  • Rush - Styx - Yes - Porcupine Tree - Staple Food!
Re: Backup time
« Reply #10 on: December 11, 2015, 08:27:25 pm »

Carl - I just spoke to Synology and all backups after the initial one are incremental. Given how long your incremental backups take, I can see where 14 hours per terabyte is normal for a Synology NAS.


I figured it must be after reading everyone else's results.  I just can't believe I am that consistent in my Additions / Changes?

Carl.
Logged
Carl's Music: https://cloud.clz.com/jctcom/music
Carl's Movies: https://cloud.clz.com/jctcom/movies

Some of Carl's Equipment:  Yamaha RX-A2A, i7-11700K, 128GB, PCIe X4 2TB M.2 SSD, GTX-970, SMSL DL200

blgentry

  • Regular Member
  • Citizen of the Universe
  • *****
  • Posts: 8014
Re: Backup time
« Reply #11 on: December 12, 2015, 05:07:37 am »

^ In incremental backups, all files must be compared to the previous set of files to see if any need to be backed up.  This takes time.  Depending on the various factors, these comparisons can take longer than the actual transfer of new files.  If your backup software produces a log, that would be the place to look to see what it's doing and how long it's doing it.

Brian.
Logged

mwillems

  • MC Beta Team
  • Citizen of the Universe
  • *****
  • Posts: 5234
  • "Linux Merit Badge" Recipient
Re: Backup time
« Reply #12 on: December 12, 2015, 08:00:47 am »

^ In incremental backups, all files must be compared to the previous set of files to see if any need to be backed up.  This takes time.  Depending on the various factors, these comparisons can take longer than the actual transfer of new files.  If your backup software produces a log, that would be the place to look to see what it's doing and how long it's doing it.

Brian.

Depending on the backup solution, incrementals don't necessarily compare files.  For example, many backups solutions just check the file size and date modified stamp on the files and if they haven't been touched since the last backup they just skip them (rsync and many other backup solutions work this way by default).  That allows some file-based incremental solutions to move absurdly fast (I have one that can do a daily incremental on 8TB in less than half an hour, and running another one immediately after one run completes takes about a minute and a half). 

However, many incremental solutions either don't trust the filesize/timestamps (so they checksum all the files all the time) or they do disk image based backups rather than file backups (which requires serious analysis every time). 

In addition to checking the logs, I would advise doing some research on your backup method to find out more about how it works (and whether a different one might work better/faster for your needs).  We're living in something of a renaissance in backup software and modern deduplicating backup methods are getting pretty close to the holy grail of backups (fast, compressed, deduplicated, incrementals forever, etc.)

I put WD Red drives, which spec 6 GB/s so I don't think those are slowing things down.

6 Gb/s is the sata link speed, not the drive speed (that's the maximum theoretical speed that your motherboard could support).  I run several WD Reds in various sizes and at their fastest they do 180MB/s and at their slowest about 80 MB/sec.  On average I see about 120MB/sec in throughput with single WD Reds, so that's definitely part of the story, but that's not your main bottleneck as with those kinds of speeds you'd be done with a TB in three hours or so.
Logged

robydago

  • Citizen of the Universe
  • *****
  • Posts: 518
Re: Backup time
« Reply #13 on: December 13, 2015, 06:13:03 pm »

robydago, by disk setup speed do you mean something other than RPM? If so, I'm not familiar with the term. Please give me a synonym or a little more explanation. I put Western Digital Red HDDs in the Synology. According to a review: "The WD Red NAS Hard Drive has a SATA 6Gb/s interface, 3TB capacity, 5K spindle speed, 1TB platters and 64MB cache."

You can ignore the interface speed as it is much faster than the disk physical read\write speed.
You need to check the actual performance data for your hard disk.
Page 3 and 4 of this review for example:
http://www.legitreviews.com/wd-red-3tb-nas-hard-drive-review_2092

Logged

blgentry

  • Regular Member
  • Citizen of the Universe
  • *****
  • Posts: 8014
Re: Backup time
« Reply #14 on: December 13, 2015, 06:48:11 pm »

Depending on the backup solution, incrementals don't necessarily compare files.  For example, many backups solutions just check the file size and date modified stamp on the files and if they haven't been touched since the last backup they just skip them (rsync and many other backup solutions work this way by default).

I'm continually surprised at how deep your knowledge goes Mwillems.  :)

I used to run a fairly large backup solution at a company I worked for that happened to use "incremental forever" as it's driving philosophy.  So I'm somewhat versed in how this kind of thing works.  My idea was that the comparisons take some time.  This might involve calls to the backup system's database, reading file metadata, etc.  It takes *some* amount of time.  If there are many many files, these things might take longer.

These days I use rsync a lot too.  It's wonderful.  :)

Brian.
Logged

mwillems

  • MC Beta Team
  • Citizen of the Universe
  • *****
  • Posts: 5234
  • "Linux Merit Badge" Recipient
Re: Backup time
« Reply #15 on: December 14, 2015, 07:52:38 am »

I'm continually surprised at how deep your knowledge goes Mwillems.  :)

I've been doing a lot of research and/or tinkering with backup solutions because I needed a space efficient one for daily workstation backups.  I found an open source one that was almost perfect (attic), and I dug down into the code (and the code of a few other backup projects) to see if I could fix the "almost" part and learned quite a bit in the process.  Meanwhile, someone forked attic and not only fixed my "almost" but added a whole bunch of even more awesome stuff.  So my preliminary attempt to fix the code fell by the wayside, but I now have an awesome backup solution (borg) and I know more than I probably want to about the guts of various linux backup programs  ;D


Quote
I used to run a fairly large backup solution at a company I worked for that happened to use "incremental forever" as it's driving philosophy.  So I'm somewhat versed in how this kind of thing works.  My idea was that the comparisons take some time.  This might involve calls to the backup system's database, reading file metadata, etc.  It takes *some* amount of time.  If there are many many files, these things might take longer.

Definitely comparisons do take time and if a backup solution is doing the comparison every time incrementals will always be fairly slow.

Quote
These days I use rsync a lot too.  It's wonderful.  :)

I use rsync for "must-succeed" backups, but if you have any Unix or Linux boxes, check out borgbackup (or it's parent project attic); I think there's a dev port to OSX but I don't know how well tested it is.  The feature set is completely insane.  It's under heavy development so I don't use it for deep storage must-succeed backups, but its features make it great for daily (or even hourly) backups, and it's saved my bacon a few times.

Things borgbackup does:

1) Globally deduplicating with a user specifiably chunk size: as it backs up, it compares 2MB (or whatever sized) chunks and if it's ever seen one with the same hash before it doesn't back it up and just includes a reference in the database.  This makes it much more space efficient than traditional backup solutions as rsync-diff type solutions can deduplicate incremental changes to a file, but borg can deduplicate two copies of the same or similar files!  For example, if you back up multiple machines to the same repository system libs will only be stored once drastically reducing the size needed for extra machines.
2) Rafts of compression options
3) Because of properties 1) and 2) initial backups are often much smaller than the size occupied on a drive, and daily incremental backups are very, very small on a typical workstation (a few hundred MB reflecting browser cache changes most days), so you can retain incrementals indefinitely (as space allows).
4) But you can also prune old backups with more or less arbitrary criteria or by name (i.e. keep 6 monthly, 4 weekly, 7 daily, etc.)
5) You can mount any backup as though it was an external drive via fuse and browse through and just copy what you want.  This feature alone makes it much more useful than other similar solutions.
6) Encryption supported, and can be done client-side so an untrusted backup server never sees anything but an encrypted blob.
7) Excellent memory and speed performance by default, but it's also user tunable.  For example, if you have limited RAM available you can tune the chunk size and take a modest hit on deduplication performance in exchange for lower RAM use.  It took a little over 2GB to backup 8TB with the now default settings, but I could have increased the chunk size to get it lower, etc.  In terms of speed, it's about half as fast as rsync at default settings, which given how much it's doing is pretty impressive.

I'm currently using it for daily backups of all my workstations, and I'm piloting it on a redundant backup of my entire 8TB media collection. So far every restore has gone perfectly, and I've done five incremental backups on the 8TB collection and the incrementals typically take less than half an hour over a gigabit link.  That said the initial 8TB backup took about 3 days :-)  On incrementals with more reasonable sizes (i.e.a  typical workstation) it's fast enough that you could use it for hourlys without much of a performance hit.  
Logged
Pages: [1]   Go Up