Re: [SLUG] rdiff-backup
On Thu, 2009-06-04 at 14:29 +0800, jam wrote: rsync -xa -e ssh --exclude=.gvfs [...] tigger:/home/ [...] rsync -xa -e ssh --exclude=.gvfs [...] tigger:/root/ [...] .gvfs is mounted as a FUSE filesystem, so that if you include --one-file-system (or -x), you won't need to exclude .gvfs explicitly, and you get the added bonus of not backing up any other random mountpoint that happens to manifest itself within your filesystem. signature.asc Description: This is a digitally signed message part -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
[SLUG] rdiff-backup
Wearing my I wish somebody had said this, useful, sharing cap and not my dragons and swords one I declare that, except for those interested in pages- n-pages of python stack traces or very simple systems, rdiff-backup is unmentionably awefull. I backup daily 4 systems with a couple of 100s G each. rdiff-backup does not play with other versions, the client and server machines are different, and rdiff-backup needs to be on both machines. Some work, some dont. All versions complained about obsolete modules. As you may expect, winter power fails during the backup happened. This resulted in chaos that took significant manual input to fix. http://www.sanitarium.net/golug/rsync_backups.html shows rsync is easy and justworks(tm) Here is a snip of my rsync backup script: It uses 1 backup+30 tinybits of space for the backup A tiny bit of logic to avoid regetting everything after a failure would help ie don't cycle and copy unless all suceeded #! /bin/bash DAY=`date +%a` Cycle all the daily backups rm -fr /backup/tigger.etc.30 rm -fr /backup/tigger.home.30 rm -fr /backup/tigger.mail.30 rm -fr /backup/tigger.root.30 ... for (( i=29 ; i ; i-- )) do let j=$i+1 if [ -d /backup/tigger.etc.$i ] [-d /backup/tigger.etc.1 ]; then mv /backup/tigger.etc.$i /backup/tigger.etc.$j fi if [ -d /backup/tigger.home.$i ] [-d /backup/tigger.home.1 ]; then mv /backup/tigger.home.$i /backup/tigger.home.$j fi if [ -d /backup/tigger.mail.$i ] [-d /backup/tigger.mail.1 ]; then mv /backup/tigger.mail.$i /backup/tigger.mail.$j fi if [ -d /backup/tigger.root.$i ] [-d /backup/tigger.root.1 ]; then mv /backup/tigger.root.$i /backup/tigger.root.$j fi ... done tigger echo tigger /tmp/mail date /tmp/mail rsync -xa -e ssh --exclude=.gvfs --delete --link-dest=/backup/tigger.home.2 tigger:/home/ /backup/tigger.home.1 rsync -xa -e ssh --exclude=.gvfs --delete --link-dest=/backup/tigger.root.2 tigger:/root/ /backup/tigger.root.1 rsync -xa -e ssh --delete --link-dest=/backup/tigger.etc.2 tigger:/etc//backup/tigger.etc.1 rsync -xa -e ssh --delete --link-dest=/backup/tigger.mail.2 tigger:/var/spool/mail/ /backup/tigger.mail.1 ssh tigger df -h | grep '^/dev' /tmp/mail ... James -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] rdiff-backup
2009/6/4 jam j...@tigger.ws: Wearing my I wish somebody had said this, useful, sharing cap and not my dragons and swords one I declare that, except for those interested in pages- n-pages of python stack traces or very simple systems, rdiff-backup is unmentionably awefull. We are just starting to deploy Bacula after some examination. The matter with backups is not just to make them - but also to manage a catalog to help you find the right file when you have to restore things. #! /bin/bash DAY=`date +%a` Cycle all the daily backups rm -fr /backup/tigger.etc.30 rm -fr /backup/tigger.home.30 rm -fr /backup/tigger.mail.30 rm -fr /backup/tigger.root.30 ... for (( i=29 ; i ; i-- )) do let j=$i+1 if [ -d /backup/tigger.etc.$i ] [-d /backup/tigger.etc.1 ]; then mv /backup/tigger.etc.$i /backup/tigger.etc.$j fi if [ -d /backup/tigger.home.$i ] [-d /backup/tigger.home.1 ]; then mv /backup/tigger.home.$i /backup/tigger.home.$j fi if [ -d /backup/tigger.mail.$i ] [-d /backup/tigger.mail.1 ]; then mv /backup/tigger.mail.$i /backup/tigger.mail.$j fi if [ -d /backup/tigger.root.$i ] [-d /backup/tigger.root.1 ]; then mv /backup/tigger.root.$i /backup/tigger.root.$j fi ... done Consider looking at savelog or logrotate. I'm not sure where savelog popped up from but it appears to be a script-friendly interface to one-off logrotates. Cheers, --Amos -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] rdiff-backup
2009/6/4 Amos Shapira amos.shap...@gmail.com: The matter with backups is not just to make them - but also to manage a catalog to help you find the right file when you have to restore things. On that note, you can use the rdiffWeb[0] tool to browse your rdiff-backup revisions and recover files. [0] http://www.rdiffweb.org/ Lindsay -- http://holmwood.id.au/~lindsay/ (me) -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
[SLUG] rdiff-backup - does it work as advertised?
I've been advised to use rdiff-backup for backing up (!) and it looks very nice. Has anyone any experience with this tool? If so, how do you test it's integrity and what if any problems did you encounter? I've spot checked the mirrors that it creates and they look perfect (correct byte counts, user/groups look good, spot checking files look good) but I don't want to put my eggs in this basket and find out the hard way that the basket has holes. Also, I have no idea how to test the restore functions without actually restoring, which I only want to do in the case of a crisis. Does anyone have suggestions about this? -- David McQuire -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
