duplicity-talk
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Duplicity-talk] Problems with restore of LARGE backup from S3


From: Roy Badami
Subject: Re: [Duplicity-talk] Problems with restore of LARGE backup from S3
Date: Mon, 14 May 2012 15:17:14 +0100
User-agent: Mozilla/5.0 (Windows NT 6.1; rv:12.0) Gecko/20120428 Thunderbird/12.0.1

On 23/04/2012 15:32, address@hidden wrote:
On 23.04.2012 16:26, Roy Badami wrote:
I'm trying to perform a test restore of the S3 backups I've been doing of a 
LARGE fileserver (approx 700GB , approx 100,000 files) which comes to 236 
volumes (with volsize set to 2000) in a full backup.  Duplicity version is 
duplicity-0.6.17-1.el6.x86_64

I've made two attempts to perform a full test restore to another machine, and 
in both cases it eventually gives up.  It got to volume 127 and then keeps 
failing with
Download s3+http://...//duplicity-full.20120304T000513Z.vol127.difftar.gpg 
failed (attempt #9993, reason: IncompleteRead: IncompleteRead(0 bytes read, 
1622573204 more expected))

Then eventually, when it hit the retry limit of 9999, it printed the following, 
and now is hung doing nothing

BackendException: Error downloading 
s3+http://...//duplicity-full.20120304T000513Z.vol127.difftar.gpg

A couple of other things I notice.  The duplicity process has grown to almost 
4GB - is this expected? (however, the machine has lots of RAM and swap, so this 
in itself shouldn't be a problem)

root     31336  9.1  1.8 3801340 613756 pts/2  Sl+  Apr18 673:44 
/usr/bin/python /usr/bin/duplicity --s3-use-new-style --num-retries 9999 
--tempdir /opt/restore-tmp --archive-dir /opt/restore-archive restore 
s3+http://.../ /opt/restore-test

Also, there are lots of gpg processes running (presumably left behind)?  
However no where near one per volume (there are 38 gpg processes, and we've 
processed 126 volumes before we hit the problem)

Any thoughts?  Anyone else successfully using duplicity on filesystems this 
large?

on the first look it is a downloading issue. could you try to download your 
backup chain to a local drive and restore from there?

second idea is that we had a memory leak between 0.6.13-17 . so please try 
0.6.18 and see if that helps. install from tarball as described here under TIP
http://duply.net/?title=Duply-documentation
if your distro does not have the latest.


Thanks, and sorry for delayed response. A downloading issue was my first thought too, hence why I tried again with even with num-retries set very high. The fact that it failed again even with num-retries set high makes me suspect there may be another problem (this is over an enterprise net conenction - not DSL - and I saw no other evidence of a network outage).

Thanks for your suggestions, I will give them a go and report back.

Incidentally, is the large number of gpg processes that was seeing considered normal, or is that perhaps a sign of some problem somewhere?

Regards

roy

--
Roy Badami
Roboreus Ltd
Third Floor
The Place
175 High Holborn
London WC1V 7AA





reply via email to

[Prev in Thread] Current Thread [Next in Thread]