Hi all, I've built a new backup server for our linux-clients. Is Amanda the way to go for a backup-solution? It seems to be pretty powerful, if a bit finickety to set up initially. The way we currently do backups is to use rsync from the clients to two folders on an older server that rolls over every other week. This worked fine for a while, but the rsync is cumulative and the users generate a tremendous amount of data every day after having had a client upgrade with newer and hilariously fast computers for calculation. The previous *nix-admin set it up this way with rsync, meaning that we in the long run have data that is way obsolete and get increasingly difficult to maintain. As the backup solution must be next to free, ie "free beer", Amanda looks suitable. What do you use for backing up data? -- BW, Sorin ----------------------------------------------------------- # Sorin Srbu [Sysadmin, Systems Engineer] # Dept of Medicinal Chemistry, Phone: +46 (0)18-4714482 >3 signals> GSM # Div of Org Pharm Chem, Mobile: +46 (0)701-718023 # Box 574, Uppsala University, Fax: +46 (0)18-4714482 # SE-751 23 Uppsala, Sweden Visit: BMC, Husargatan 3, D5:512b # Web: http://www.orgfarm.uu.se ----------------------------------------------------------- # () ASCII ribbon campaign - Against html E-mail # /\ # # MotD follows: # One tentacle, one vote. -------------- next part -------------- A non-text attachment was scrubbed... Name: smime.p7s Type: application/x-pkcs7-signature Size: 5106 bytes Desc: not available URL: <http://lists.centos.org/pipermail/centos/attachments/20100113/b60b015d/attachment-0001.bin>
From: Sorin Srbu <sorin.srbu at orgfarm.uu.se>> I've built a new backup server for our linux-clients. > Is Amanda the way to go for a backup-solution? > It seems to be pretty powerful, if a bit finickety to set up initially. > The way we currently do backups is to use rsync from the clients to two > folders on an older server that rolls over every other week. This worked fine > for a while, but the rsync is cumulative and the users generate a tremendous > amount of data every day after having had a client upgrade with newer and > hilariously fast computers for calculation. The previous *nix-admin set it up > this way with rsync, meaning that we in the long run have data that is way > obsolete and get increasingly difficult to maintain. > As the backup solution must be next to free, ie "free beer", Amanda looks > suitable. > What do you use for backing up data?Used amanda in the past (a decade ago)... I use bacula now. JD
Sorin Srbu wrote:> Hi all, > > I've built a new backup server for our linux-clients. > > Is Amanda the way to go for a backup-solution? > > It seems to be pretty powerful, if a bit finickety to set up initially. > > The way we currently do backups is to use rsync from the clients to two > folders on an older server that rolls over every other week. This worked fine > for a while, but the rsync is cumulative and the users generate a tremendous > amount of data every day after having had a client upgrade with newer and > hilariously fast computers for calculation. The previous *nix-admin set it up > this way with rsync, meaning that we in the long run have data that is way > obsolete and get increasingly difficult to maintain. > > As the backup solution must be next to free, ie "free beer", Amanda looks > suitable. > > What do you use for backing up data?You can check out BackupPC as well. That's what I use. Regards, Max
On Wed, Jan 13, 2010 at 5:04 AM, Sorin Srbu <sorin.srbu at orgfarm.uu.se> wrote: <snip>> The way we currently do backups is to use rsync from the clients to two > folders on an older server that rolls over every other week. This worked fine > for a while, but the rsync is cumulative and the users generate a tremendous > amount of data...<snip> You might want to check out the rsync switches --backup-dir and --suffix. Using them some thing like this: --delete --backup --backup-dir=$MIRROR_DIR/RsyncBckups --suffix=".$DATE" allows you to keep an exact duplication of the original directory and keeping the original files that were either deleted or overwritten in a seperate backup directory with dated suffixes, which can be archived on some regular basis. This should allow you to keep the simplicity of rsync and control the cumulative size. Brett
> On Wed, Jan 13, 2010 at 5:04 AM, Sorin Srbu <sorin.srbu at orgfarm.uu.se> wrote: > <snip> >> The way we currently do backups is to use rsync from the clients to two >> folders on an older server that rolls over every other week. This worked fine >> for a while, but the rsync is cumulative and the users generate a tremendous >> amount of data...I use rsnapshot .. which manages sets of rsync backups using hardlinks. IT may be similar to what you are using already ..
Sorin Srbu wrote:> Hi all, > > I've built a new backup server for our linux-clients. > > Is Amanda the way to go for a backup-solution? > > It seems to be pretty powerful, if a bit finickety to set up initially. > > The way we currently do backups is to use rsync from the clients to two > folders on an older server that rolls over every other week. This worked fine > for a while, but the rsync is cumulative and the users generate a tremendous > amount of data every day after having had a client upgrade with newer and > hilariously fast computers for calculation. The previous *nix-admin set it up > this way with rsync, meaning that we in the long run have data that is way > obsolete and get increasingly difficult to maintain. > > As the backup solution must be next to free, ie "free beer", Amanda looks > suitable. > > What do you use for backing up data?Backuppc is good for this - it can use rsync for the transfers (or tar or smb) but all duplicate data is compressed and pooled even if found on different machines and it has an easy setup to control how long old copies are retained. -- Les Mikesell lesmikesell at gmail.com
On 13.1.2010 12:04, Sorin Srbu wrote:> Hi all, > > I've built a new backup server for our linux-clients. > > Is Amanda the way to go for a backup-solution? > > It seems to be pretty powerful, if a bit finickety to set up initially. > > The way we currently do backups is to use rsync from the clients to two > folders on an older server that rolls over every other week. This worked fine > for a while, but the rsync is cumulative and the users generate a tremendous > amount of data every day after having had a client upgrade with newer and > hilariously fast computers for calculation. The previous *nix-admin set it up > this way with rsync, meaning that we in the long run have data that is way > obsolete and get increasingly difficult to maintain. > > As the backup solution must be next to free, ie "free beer", Amanda looks > suitable. > > What do you use for backing up data?I think nobody has yet mentioned rdiff-backup. I have very good experiences with it. Easy to setup and control (only remember first to install the required packages, and I think rsync-devel was not mentioned but is required). Rdiff-backup keeps up an exact copy of the source director(ies), plus it maintains a separate directory for deleted/changed items. With an appropriate command you can restore the source directory as it was at a given point of time. Very neat, and space-saving. http://rdiff-backup.nongnu.org/ - Jussi -- Jussi Hirvi * Green Spot Topeliuksenkatu 15 C * 00250 Helsinki * Finland Tel. +358 9 493 981 * Mobile +358 40 771 2098 (only sms) jussi.hirvi at greenspot.fi * http://www.greenspot.fi "Remember that if you can't understand an answer, it's okay, we'll just keep going forward as if you do." -- Peter Gulutzan, "MySQL 5.0 Stored Procedures"
On Wed, Jan 13, 2010 at 8:36 AM, Sorin Srbu <sorin.srbu at orgfarm.uu.se> wrote:>>-----Original Message----- >>From: centos-bounces at centos.org [mailto:centos-bounces at centos.org] On > Behalf >>Of Barry Brimer >>Sent: Wednesday, January 13, 2010 2:15 PM >>To: CentOS mailing list >>Subject: Re: [CentOS] Backup server >> >>I use rsnapshot .. which manages sets of rsync backups using hardlinks. >>IT may be similar to what you are using already .. > > Somewhat similar, thanks. I think however I need to get away from this sort > of backups. They're just to space-consuming. > -- > /SorinDo you currently need that disk space for something else? If not, then there is no reason to get rid of the old files. An empty disk is a wasted disk, so you would be creating free space on a perfectly good disk that you already paid for. Having blank space gains you nothing, but deleting a file that you could need in 6 months might cost you. Pick a percentage of the disk that you are comfortable using for backups, and then let the backups grow to that size. Only when you reach that size or need the space for something with a higher priority should you worry about recovering the space.
On Thu, Jan 14, 2010 at 9:37 AM, Sorin Srbu <sorin.srbu at orgfarm.uu.se> wrote:>>-----Original Message----- >>From: centos-bounces at centos.org [mailto:centos-bounces at centos.org] On Behalf >>Of Benjamin Franz >>Sent: Thursday, January 14, 2010 2:12 PM >>To: CentOS mailing list >>Subject: Re: [CentOS] Backup server >> >>If you have any budget at all, invest in bigger drives. 7200 RPM 1 TB >>RAID rated drives can be bought for $160 each. Desktop rated 5900 RPM >>1.5 TB drives (which you can probably get away with in a dedicated >>backup server since you don't care a lot about speed and can tolerate >>long pauses for sector repair) can be bought for $110 each. Check Newegg. > > I haven't got a budget really. Today I asked for a new group-printer today and > the boss looked pained... 8-} > > I opted for the proven 500GB-sized disks and got more of those instead. I've > had a handful of 750GB-drives die on me recently. Somehow it feels the > technology isn't quite there yet for the bigger drive-sizes. Anybody remember > the IBM Deskstars in the early 00's...? > > Also, my experience is the more smaller disks you have, the faster they get. > Less to write to each I guess. > > >>Second, to maximize 'depth' of backups you should use a 'Tower of >>Hanoi'-like backup system. > > Good advice, thanks! > -- > /SorinThere seems to be a persistent conception among managers that anything "IT related" is a huge capital expenditure (as it used to be), and there's all sorts of resistance to buying anything new. However, you probably spend more on printer paper in 1 week than it costs to buy a 1TB drive. This kind of equipment is a disposable commodity, even though the accounting department still prefers to write it off over 7 years. However, IT also has a reputation of always wanting to buy new toys. Many times these toys are not needed, even though the IT person insists that they are. So you need to be able to walk the fine line between these two. To put it into perspective, ask the manager how much it would cost the business if this data was unrecoverable? After that, if they still don't want to spend a few hundred $$s on the insurance, get it in writing that your manager understands the risk and print it out and post it on your office wall.
On Wed, Jan 13, 2010 at 5:20 PM, Jussi Hirvi <listmember at greenspot.fi> wrote:> I think nobody has yet mentioned rdiff-backup. I have very good > experiences with it. Easy to setup and control (only remember first to > install the required packages, and I think rsync-devel was not mentioned > but is required). > > Rdiff-backup keeps up an exact copy of the source director(ies), plus it > maintains a separate directory for deleted/changed items. With an > appropriate command you can restore the source directory as it was at a > given point of time. Very neat, and space-saving. > > http://rdiff-backup.nongnu.org/ > > - Jussiyes, i can agree. also using rdiff-backup and can say that it is very good sollution. even tested it by restoring few terabaytes of information.