Harpur KD Hill's picture

Hi. I’m struggling with a recently discovered peculiarity of tklbam. It appears that I cannot restore a backup generated in AmazonWS ‘availability zone’ s3-eu-west-1, corresponding to Region: Virginia, on a server originally backed up in AmazonWS availability zone s3, corresponding to Region: Ireland.

 

If I do traceroute to s3-eu-west-1.amazonaws.com it gets through consistently (I’m in the UK), whereas if I traceroute s3.amazonaws.com the packets seem to just get dropped in the ether.

 

I can tklbam-backup OK – the files go to server: s3://s3-eu-west-1.amazonaws.com/tklbam-

But I can’t tklbam-restore, because it looks to server: s3://s3.amazonaws.com/tklbam-

 

It might in fact be worse – that tklbam-restore looks to s3 generically whereas backup send to s3-eu-west-1. I can’t test this as I’m already in a difficult position on a production server which didn’t restore when required.

 

The issue could be related to this: https://www.turnkeylinux.org/forum/support/20150722/tklbam-restore-s3-failing-due-timeout, although there appears to be plenty of space. It could also be related to this, https://www.turnkeylinux.org/forum/support/20130910/tklbam-s3-error-417 but I don’t actually get error 417, I just get a timeout.

 

Any suggestions will be greatly appreciated.

 

Regards,

 

Lee

 

 

 

Forum: 
Jeremy Davis's picture

It is a weird one... It's probably not ideal but what about editing your hosts file as a dirty workaround? Something like ping s3-eu-west-1.amazonaws.com to get an IP. Then add an entry to your hosts file (/etc/hosts); something like this:
54.231.136.0 s3.amazonaws.com

Its a long way from a proper fix (as the IP might change) but should hopefully get you out of your current predicament...

Harpur KD Hill's picture

Thanks Jeremy, I'll try that.

I think TKL and tklbam are great products, so it was a shame to be let down when I needed it most. We just about scraped through the emergency, with a couple of days disruption and some data loss (I was able to restore from an older backup held in the same availability zone).

I'll post back an update on the workaround when the dust settles.

Lee

Jeremy Davis's picture

It's on our (GitHub) issue tracker: https://github.com/turnkeylinux/tracker/issues/458

Add new comment