My scheduled backup ran successfully in Fremont last night (this morning). Times are PST:

Started                  Finished               Duration                   Status
2011-06-23 01:10:51      2011-06-23 01:23:47    12 minutes, 56 seconds     successful

So you are saying Linode service is so bad that we can not expect even the basic things to operate as promised, and should constantly manually recheck absolutely everything Linode service is supposed to provide? I know some people have problems with Linode, but overall I personally had a very positive experience, so I don't agree with your implication that Linode can not be trusted to do even the basic things right.

This backup fiasco dragged on for quite some time, and constitutes a serious failure. At the very least Linode should send email notifications when Linode backup fails.

What I'm saying is if the data or service that your linode is providing is that sensitive, critical, or otherwise important, and you have ONE method of backup, which you've never tried, and which may not be the best backup method for your situation, then why arent you checking your backups daily?

tl;dr: you need to lookup the best backup methods for your needs. In my experience, linode snapshots are best for short term backups of the OS disk and are rarely the most effective method of backing up actual customer or critical data that you may need to move to another system at short notice.

you're right, dont trust anyone backup method. That includes linode.

EDIT: added word

In any case, is this the proper forum for this post?

One user's problem belongs on a trouble ticket (or with his wallet - if he doesn't like it, walk away), not in Customer Testimonials.

This is just an under handed dig at Linode without any context or rebuttials.

I'm sure the positive testimonials out number the negative by several thousand to one.

Kudo's to Linode for not just deleting this thread as a troll.

Anyone even considering giving any credence to anything the OP of this thread posts should really read his history. troll is an understatement

Nevertheless, the problem was real, many others experienced the exact same issue.

What drags these discussions on and on is the constant attempts of some posters to always "defend" linode no matter what actually happened, and blame users instead, in this case with ridiculous assertions like "you should have manually checked status of linode backups daily".

neo++!

Although it is irresponsible to solely rely on Linode's backup system -- it's not meteor- or bankruptcy-proof -- and not to occasionally check that it's working.

_________________
Matt Nordhoff (aka Peng on IRC)

Perhaps, but it really should notify you when it fails.

I'd like to share a recent experience I had with myhosting.com, on June 8th 2011 my VPS went down together with 30 other VPSs on this particular host node. The cause of the failure was according to their status blog multiple disk failures resulting in data corruption. It took three days of outage before they could migrate my VPS to another host node. I spent my three days trying to appease irate clients who could not access any of their websites. One even threaten to file a law suit! I had nightly website backups occurring and stored locally on the VPS including a paid backup which was supposedly a base backup for Open Vz containers. None of which helped in this unfortunate incident. I learned an invaluable humbling lesson have OFF SITE BACKUPS!

I have since then improve my backup procedure.
Linode VPS
Linode backups nightly
Websites backup locally Monday - Sat
Weekend backup locally Sunday

Full Website backups SCP to another server in VA - Sunday
Incremental backups SCP to same server in VA - Monday - Sat

Do a test restore at a different location.

I get an alert e-mail if any of the above website schedules fail, except for the ones created through Linode. I would agree that there should be some alert notification when Linode's backups fail.

I only use snapshots just before applying any huge updates or major setting changes.

I have to point out to be fair to myhosting.com they did issue me credit for two months. I now use my VPS account with myhosting.com solely for testing until my credit expires.

All of my production websites have been moved to Linode. Knock on wood, so far so good.

I do the same thing. I use virtualmin to do a full backup on Sunday and differentials Mon-Sat. They are then encrypted and copied to an SSHFS (Fuse) mount that is mounted to my backup server that is actually a VPS on its own at another much-cheaper provider.

The backup VPS is not one I would use on a daily basis as my production box (again, notice the cheap in my last sentence), but it doubles as not only a place to put my backups on but also bring online as my production server fairly quickly if the need arises.

I have no pity for somebody that does not create their own offsite backups. Lack of knowledge is no excuse... learn if you don't know how!

Linode should definitely notify when a backup has failed. Not building this feature into the backup system is in my view a major oversight especially if they allow failure to linger or *cough* "maintenance" to linger 6 days. 6 days fing days = we fucked up and dont know how to fix it yet, we dont consider backups important so we won't let you know 'you should be checking on us', strange given how Linode would certainly have a full port-mortem if any VPS nodes were down for 6 minutes.

However, ideally, one should still check that backup are being taken, and occasionally verify their backups - even if it costs 1 day of Linode time.

It reminds me of time when I worked in a small IT department some 15 years ago. We had a IT manager, 2 competent IT sysadmins, 1 developer for Excel/VBA etc. We had one of component IT sysadmin change the backup tapes every day at 4.30pm, and walk to other side of campus and put them in a safe, occasionally they would be verified. Now I think of this situation as Linode customer = IT manager, IT backup boy = Linode backup service. Whilst it is feasible that the IT backup boy might use those 20 minutes to go smoke some pot in some hidden area of campus and do the backups once every dozen times instead (and not mention it), it would be a major offence, the IT manager would be pissed and surprised and all hell would break lose when the offence came to light. Sure the IT manager hold the ultimate responsibility (and it rests on his head lalala) and verification by the other sysadmin is no doubt a good idea, but we dont really expect a competent sysadmin to using the backup time to go smoke pot and not tell us, we trusted this guy at least in his professional capacity. The pot smoking IT sysadmin needs to come forward, needs to volunteer that he has not been making backups, needs to admit that's not just a controllable hobby (maintenance) but that's he mentally messed up and needs some time off to get back in order, in the meantime the IT manager needs to give the backup responsibility to someone. He should probably be disciplined as well, possibly a salary cut.