Comment by ghshephard | Hacker News Reader

Comment by ghshephard | Hacker News Reader

Arq, (For backups to AWS - though obviously supports every cloud back end under the sun) and "Data Backup" by ProSoft engineering (For backups to USB) are my goto backup tools for day-day ensuring all my work documents are kept up to date.

Yes, I have Crashplan (for the last couple years, backblaze for the three years before that) - but the constant chewing up of CPU cycles gets annoying after a while. And both crashplan/backblaze "everything for $5" come with massive caveats (like deleting backups of Hard Drives that haven't been plugged in for 6 months - I've got Arq Backups of Hard Drives that I haven't plugged in for a couple years, safe and sound) - and I've never had an AWS backup bill in excess of $3.00, ARQ does a wicked good job of keeping your backups on a tight budget.

Also - awesome win for ARQ - when I moved to Singapore, I simply added a AWS Singapore S3 Bucket and wowza - fast backups on my gigabit ($49/month) link from MyRepublic. Really feel like I'm living in the future.

I think once I switch away from Aperture over to Photos, which presumably has a rock solid backup to iCloud photos, then simply doing a quarterly backup or so with CarbonCopyCloner + Arq to AWS + DataBackup to USB key will have my OS X backups covered.

like deleting backups of Hard Drives that haven't been plugged in for 6 months

Wait - what?

If your hard disk gets lost or destroyed, how do you protect those backups?

You need to keep them connected periodically otherwise if you lose it without reconnecting... lost.

For backblaze it's even less time.

I believe for CrashPlan, this ONLY applies to CLOUD backups. If you backup to local media or across computers, there is no "auto-delete" timeframe.

Yea I think so, it's only for them to keep you from storing everything up there forever when it could just be stale data.

I think it's fair if these are supposed to be daily backups to do this esp with cloud data and having to provide "unlimited" space to everyone using the product.

Don't know whether I'm grand-fathered into a different terms but I've just pulled some files from a back-up on a machine that was last active 5.4 years ago.

Hello Everyone,

Wanted to jump in here to confirm.

This policy only affects devices that have not connected to CrashPlan Central in 6 months or longer. This does not affect volumes that have not connected to the device in that period of time. (i.e. an external hard drive that has not connected in 6 months.) Additionally, there is no minimum connection time for local CrashPlan backups.

It’s important for CrashPlan users to consistently connect their device(s). Part of CrashPlan’s ability to maintain the archive health and integrity relies upon regular connection from the device. CrashPlan is able to routinely perform maintenance on the archive by comparing checksums between both device and CrashPlan Central.

https://support.code42.com/Administrator/3/Monitoring_And_Ma...

Please let me know if I can provide additional clarity.

Best regards,

Jarrod

Is this also the case for crashplan business or PROe? (ceejay in europe).

Since the infrastructure for CrashPlan's backup engine is the same between our Business/Consumer clients, we recommend that all users routinely connect their devices to the backup destinations. That being said, this policy only affects CrashPlan for Home subscribers at this time.

"You need to keep them connected periodically otherwise if you lose it without reconnecting... lost. For backblaze it's even less time."

No way. What ?

You're saying that if I dutifully pay my $X/mo for unlimited backup space, but fail to connect, or perform an update, the remote data is removed ?

What ?

Yeap, most if not the overwhelming majority of backup services work like that. Some like Mozy have as low as a 30 day counter. [1]

--

[1] At least they had that 30 days limit a few years ago, I haven't used them in a while.

Well, I know one that doesn't work like that :)

It took 3 years of disconnection before I got a notice the other day that one of my old backed up PCs was going to be deleted.

It makes sense; they're not an open-ended data archival company. If you don't have a copy of the data locally, you can't expect CrashPlan to keep the one true copy indefinitely. Connecting the data to the internet twice a year doesn't seem too onerous to show evidence of that, particularly for a cloud backup, the entire premise of which is that you're connected.

Absolutely agree with you - I think CrashPlan/BackBlaze are acting entirely reasonably when they delete old hard drives, particularly if they give a bit of grace after sending the email that they are about to nuke them.

It's that just for some of us, who like to archive something like a 100 GB Hard Drive onto Amazon Glacier, for $0.007/GB/Month. (Roughly $0.70/month + $5 upload fees for a 100 GB Hard Drive Archive) - and just leave it there, presumably for decades, are better served by Arq + Glacier than we are by CrashPlan/BackBlaze - they are entirely different tools for different purposes.

On the Flip Side, backing up users two 5+ Terabyte Hard Drives on S3 with Arq gets a little pricey... How crashplan/backblaze manage to do it for $5 is beyond me. Presumably it's because most its users are sending in < 100 Gigabytes (after deduping)

Yev from Backblaze here -> Absolutely. They key difference is backup vs. archive. Backblaze was designed as a backup solution, it's intended to be a 1:1 copy of your user data, and if your data set changes we change it on our end as well, with a 30-day history for accidental deletes. We need to reclaim that space to keep costs down, and we're not intended nor designed to be an archive (keeping data forever).

Backblaze B2 is designed differently and can be used as an archival system. The philosophies are different, but one of the reasons that we created it was to give folks the option of making actual archives they could keep in the cloud.

We hit the $5/month price-point by having our own server design, and by reclaiming space on occasion when data sets are removed. On the B2 side, since you're paying per GB, we can afford to keep that data for longer stretches. Hopefully Arq will integrate with B2 in the future and you'd be able to use their system to pick and choose what you want archived and have B2 as a possible repository.

I think your comparison is out of date [1], AWS S3 infrequent access has $0.0125 per GB now not $0.022+ that you have listed. Of course yours is still much better with $0.005 per GB.

Also, it does not look like you have any kind of consumer offering for it, as I have to contact sales to get anything at all. So there's really no point for Arq to even try adding support. I like your prices though.

[1] https://www.backblaze.com/b2/cloud-storage-providers.html

We don't write out the infrequent access pricing as it's not what we're competing with, also why we don't price out Nearline or Glacier, it's not at 1:1 for what we're trying to offer. The consumer offering is available from the docs: https://www.backblaze.com/b2/docs/. You can use the web interface if you want, or tie in with CLIs/APIs. The sales channel is primarily for people that want to integrate it in to the apps that they are building. Hope that clears some things up!

Thanks for the clarification.

These pages make it clear how it works and it seems reasonable:

https://help.backblaze.com/hc/en-us/articles/217665398-Backi...

https://help.backblaze.com/hc/en-us/articles/217664898-What-...

”On the Flip Side, backing up users two 5+ Terabyte Hard Drives on S3 with Arq gets a little pricey”

I send my Arq backups to Amazon Cloud Drive using the ”unlimited” plan. It costs $59.99/year. So far I have only backed up < 100 GiB however, so I don’t know how well it handles backups that are multiple TiB.

[Edited to correct confusing typo.]

Well, I am fairly certain that Amazon S3 will handle multi-terabyte backups, mostly because their pricing tiers are for 1 TB, 50TB, 500 TB, 1 Petabyte, 4 Petabytes, and 10 Petabytes - and they make more money the more you store. Performance (at least in Singapore) is also pretty awesome if you have a gigabit connection.

I'll be interested in hearing of any experiences (particularly around performance) of someone attempting to backup on the order of 10 TBytes on the Amazon Cloud Drive. My guess is that if more than very few people do this, then either (A) Amazon puts an end to "unlimited" (and yes, I appreciated the scare quotes), or, (B) They rate limit uploads after a certain size to the point at which it just frustrates people.

For some of us, dealing with a vendor who finds greater usage on your part to be a desirable behavior, such that they actually give you price breaks the more you use, creates a business relationship that is worth more than the several hundred dollars/year you'll end up saving. (Of course, this is coming from the guy who has a $36/year AWS bill, $24 of which is S3 storage)

(Side note - When talking about Storage, it's very rare to use GiB/TiB - Data rates and Storage are almost always GB or TB).

How much are you storing for those $3.00/month? Sounds too good to be true :-)

We're currently just rsyncing our pictures and stuff at home to two 3TB USB-drives (one active and one at my parents' place; using LUKS and btrfs with compression, snapshots). But even after running deduplication, they're filling up (raw files …), so I'm always on the lookout for other options. Upgrading to 2x4TB is a bit expensive, but I haven't yet found anything that'll cost me less than that while still having client-side encryption and Linux support. E.g. tarsnap seems to be about $250/TB-month, http://jotta.no/ is unlimited but has no Linux/encryption support, and I never understood Glacier pricing :-)

(and it's really convenient to be able to just restore from a local USB drive instead of having to wait for the network, though of course it's less convenient not having backups when travelling …)

AWS storage with S3 is $0.03/GB/Month. I have probably around 20 Gigabytes, stored in Singapore and US-East, versioned back 2+ years + another 58 Gigabytes in Personal Photos Backed up on Amazon Singapore (which I really should move to Glacier). It's compressed when stored on S3, so total storage is only 54 Gigabytes on Amazon Singapore, and 12 Gigabytes on Amazon-East.

I shoot a ton of pictures with my SLR, but, at the same time, I don't shoot raw, my camera is an EOS 10D (6.3 megapixel resolution) and I'm hyper aggressive about deleting all but the top 5% of my shots each day. I may shoot 300 pictures and keep 10-15.

So, I guess the major difference is I'm backing up about 78 Gigabytes of Data (though 20 of which is in two locations).

Reply on news.ycombinator.com