[Bacula-users] Cleaning LOG.MYD
I have a 17GB LOG.MYD file in my MySQL's bacula database. This is because a previous configuration our bacula instance was using MySQL as a log destination. We no longer want that and have set our logs to go to text files that can be rotated. As I understand it, Bacula will purge these logs when their associated job entries are purged. We are using Bacula to perform archives as well as backups so we are keeping data around for a very long time. I would like to continue to keep the data in volumes with the ability to restore it but disregard log entries about it. So I'm left with a 17GB table. Of course, our MySQL backups are backing this table up with regularity and I'd like to stop that and free up some disk space in the process. This file has not been modified in a month so I'm confident nothing else is writing to it. Is it safe to clean this table out? If so, what is the safest or easiest way to clean that table out? (bonus points if safest and easiest are the same!) What are the repercussions of wantonly wiping out all entries in this table? Eric -- Write once. Port to many. Get the SDK and tools to simplify cross-platform app development. Create new or port existing apps to sell to consumers worldwide. Explore the Intel AppUpSM program developer opportunity. appdeveloper.intel.com/join http://p.sf.net/sfu/intel-appdev___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Noob user impressions and why I chose not to use Bacula
On Mon, Dec 5, 2011 at 11:36 AM, Pablo Marques pmarq...@miamilinux.netwrote: Thanks you Jesse for the feedback. Regarding the disaster recovery, I have a suggestion for the bacula team: Why not make the director write the bacula config files and any relevant bsr files at the beginning of each tape? The space wasted on the tape to save these file would be very small. Well, the first problem here is that the Director would have to know how much space it was going to need for BSR files. Of course, it could pre-allocate a fixed-size block of, say, 1MB for BSR files. Agree, 1 MB is basically nothing on a tape and it can accommodate easily a huge amount of bsr files. My /etc/bacula is 88k uncompressed. The second problem, it seems to me, is that this would break compatibility with all older Bacula volumes and installations. not necessarily if you make this information at the beginning of the tape look like a volume file. It will be ignored by old directors because it will look the same as a failed job that took space on the tape. Pablo Or you could just decide that backward compatibility of that type is just not that important. For instance: versions x.0.0 and later use this format. Tapes written in this format are not accessible to older versions. It's OK to do that. It's also OK to have the option of writing in the older format from the newer directors. This will give you time to bring all your directors up to date before switching to the new format. -- All the data continuously generated in your IT infrastructure contains a definitive record of customers, application performance, security threats, fraudulent activity, and more. Splunk takes this data and makes sense of it. IT sense. And common sense. http://p.sf.net/sfu/splunk-novd2d___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Noob user impressions and why I chose not touse Bacula
On Mon, Dec 5, 2011 at 4:47 PM, Jose Ildefonso Camargo Tolosa ildefonso.cama...@gmail.com wrote: On Mon, Dec 5, 2011 at 6:55 PM, James Harper james.har...@bendigoit.com.au wrote: Regarding the disaster recovery, I have a suggestion for the bacula team: Why not make the director write the bacula config files and any relevant bsr files at the beginning of each tape? The space wasted on the tape to save these file would be very small. A script to email the bsr file to a gmail/Hotmail/whatever account would suffice. It's not like the file contains any sensitive information. Well, yeah, you could also rsync those files to any of the available online storage solutions (some of them free) but, I think this breaks the point of being able to recover from just the tape, or for instance: any storage (disks, for example)... Now, I wonder if any backup solution out-there allows you to do this... ie: recover from just backup media, you *always* need to get OS running again... so, there is no such thing as bare metal recovery. Unless you create something like an installer image that uses the backup to restore the machine... mmm maybe a life-bacula... that automatically rebuild the catalog from volumes... uh... is that possible? (automatically and completely rebuild catalog from volumes?). Ildefonso Camargo How about a special DR backup that dumps the database to the DR volume along with the server's config and each defined client's config? The volume would be a simple tarball you could store on a thumbdrive, tape, cloud, whatever. When you do your bare metal restore of your systems, you only need to retrieve that tarball. With it you should be able to completely recreate your catalog and clients from this sucker with a drbacula tool. As long as I don't have to cough for it, that should be a pretty good basis for bare metal restores of servers and clients from a live distro. Eric -- All the data continuously generated in your IT infrastructure contains a definitive record of customers, application performance, security threats, fraudulent activity, and more. Splunk takes this data and makes sense of it. IT sense. And common sense. http://p.sf.net/sfu/splunk-novd2d___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
[Bacula-users] Need help canceling a job
I tried restoring a file but the restore wanted to mount an offsite tape. I took another path to get the file I need so no longer need this restore done. However, it's still waiting on a mount so I cannot cancel it. Here is what happens. *status all --snip-- JobId 25604 Job RestoreFiles.2011-10-19_11.53.27_04 is running. --snip-- *cancel jobid=25604 JobId 25604 is not running. Use Job name to cancel inactive jobs. *cancel job=RestoreFiles Warning Job RestoreFiles is not running. Continuing anyway ... JobId 0, Job RestoreFiles marked to be canceled. *status all --snip-- JobId 25604 Job RestoreFiles.2011-10-19_11.53.27_04 is running. --snip-- I have another restore queued up that I do need this one is keeping the one I need from proceeding. How can I cancel this restore job? Thanks! Eric -- The demand for IT networking professionals continues to grow, and the demand for specialized networking skills is growing even more rapidly. Take a complimentary Learning@Ciosco Self-Assessment and learn about Cisco certifications, training, and career opportunities. http://p.sf.net/sfu/cisco-dev2dev ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Need help canceling a job
On Wed, Oct 19, 2011 at 2:17 PM, Ben Walton bwal...@artsci.utoronto.ca wrote: I'm not positive, but I believe that to reference this job by name, you'd want: RestoreFiles.2011-10-19_11.53.27_04 Not just RestoreFiles. Also, if you just say cancel, doesn't it prompt you with a menu driven choice? Thanks -Ben -- Ben Walton Systems Programmer - CHASS University of Toronto C:416.407.5610 | W:416.978.4302 Thanks for your reply! I've also tried using the full name as you mentioned and I get the same exact response and the job is not canceled. If I run cancel without options, I get this: *cancel No Jobs running. Yet, I can still see the job is listed as running in the status output. I just restarted bacula-sd and that did cancel the job. But that just seems like hitting a small nail with a 20lb sledge hammer. It also seems like the wrong solution give that it can affect other processes too. I've been looking through the documentation and it looks like it says I need to perform a mount of the volume before the job will cancel. That can't be right! The volume is offsite and not available right now which is why it is waiting on the mount in the first place. While I can go retrieve the offsite volume and perform the mount, that's a lot of legwork since I'm not on the same side of the city where my datacenter lives. Do I really have to mount the volume to be able to cancel the job? If so, I cannot make sense of that. Is there another way? Eric -- The demand for IT networking professionals continues to grow, and the demand for specialized networking skills is growing even more rapidly. Take a complimentary Learning@Ciosco Self-Assessment and learn about Cisco certifications, training, and career opportunities. http://p.sf.net/sfu/cisco-dev2dev ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Bacula, Offsite, and Restores
2011/9/17 Rodrigo Renie Braga rodrigore...@gmail.com: Well, in order to restore from the Copy tapes, yes, you have to purge the original tapes... But you don't have to purge the original tapes if you want to do a normal routine restore. Even if you pull out the Copy tapes from your library right after a Copy Job (which is the hole idea), your restores will work normally like there never was a Copy Job ran before... The original tapes will stay intact on your on-site location and you'll use these original tapes to do your restores. You will only use your copy jobs tapes in case of a disaster... Yes, but what I am looking for is the easiest method for my team to perform restores in case I'm not there, even if it is a restoration from offsite volumes. Having them go through the process of finding the right volumes to purge and then purging them is something I don't think they should have to learn how to do. It's nice that the ability is there in Bacula, but I think it's far too complicated for a robust solution. The solution needs to be easy and intuitive and merely running a second client makes it ridiculously easy and intuitive. I think it's the right way to go for now. Eric -- BlackBerryreg; DevCon Americas, Oct. 18-20, San Francisco, CA http://p.sf.net/sfu/rim-devcon-copy2 ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] How to use retention to limit disk usage
On Fri, Sep 16, 2011 at 7:20 PM, Eric Sisolak haldir.j...@gmail.com wrote: Hello, I am looking for a way to limit the amount of space taken up by backups by truncating/deleting volumes whose files and jobs have been pruned according to retention periods with Bacula 5.0.3. ActionOnPurge=Truncate looks like it could do what I want, but it seemed like there many people had issues with it. Has anyone else implemented this? How did you do it? --Eric I am doing something like this, but not using ActionOnPurge. I'm using vchanger as a VTL and initializing X number of volumes. Lets say X = 80. I tell Bacula to limit the volume size to 5GB meaning I'm now only using 400GB of disk space for backups. I also tell Bacula to use a volume retention period of one day and to recycle the oldest volume. Bacula will not use these settings unless it runs out of volumes so your data will be retained in older volumes until you run out of space in the pool. But, once you've filled up your 80th volume in this scenario, it will look at the volume with the oldest last written timedate stamp in the pool and see if the volume is past its retention period. If it is, it will purge the jobs and files and re-use the volume. The net result is that Bacula is now set to use 400GB of storage space and never exceed it. It will automatically cannibalize the oldest volumes and purge records associated with those volumes as needed. If you are a little queasy about a 1 day volume retention, you can set this to something higher like one month to insure you always have at least one month's worth of backups. Just be aware that any jobs attempting to use storage when all 400GB are allocated will hang waiting for volumes if there are no volumes past their volume retention period. You must make sure that you have enough storage to handle the actual retention period you want. This also makes the names of the volumes generic, so your volume names will no longer be indicative of their contents. I find that using volume names tied to their jobs and pools or whatnot to be useless for me. So using generic names for the volumes results in no loss, but I gain the ease of use of vchanger. Here are the relevant config sections I'm using to accomplish this: -- bacula-sd.conf -- Device { Name = PrimaryVTLDevice DriveIndex = 0 Autochanger = yes Media Type = File Device Type = File Archive Device = /var/lib/bacula/PrimaryVTL/0/drive0 Random Access = yes RemovableMedia = yes LabelMedia = yes } Autochanger { Name = PrimaryVTLAutoChanger Device = PrimaryVTLDevice ChangerDevice = /etc/bacula/PrimaryVTL.conf ChangerCommand = /opt/bacula/bin/vchanger %c %o %S %a %d } --- bacula-dir.conf --- Storage { Name = PrimaryVTLStorage Address = enter.your.bacula-sd.hostname.here SDPort = 9103 Password = Device = PrimaryVTLAutoChanger Media Type = File Autochanger = yes } Pool { Name = PrimaryVTLPool PoolType = Backup Storage = PrimaryVTLStorage AutoPrune = yes VolumeRetention = 1 day MaximumVolumes = 80 MaximumVolumeBytes = 5368709120 RecycleOldestVolume = yes } --- PrimaryVTL.conf --- changer_name = PrimaryVTL work_dir = /var/lib/bacula/PrimaryVTL virtual_drives = 1 slots_per_magazine = 80 magazine_bays = 1 magazine = /var/backups/PrimaryVTL -- Hope that helps! Eric -- BlackBerryreg; DevCon Americas, Oct. 18-20, San Francisco, CA http://p.sf.net/sfu/rim-devcon-copy2 ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Bacula, Offsite, and Restores
Thank you for your feedback, Rodrigo. I looked up the copy job information as you suggested. From what I can tell, you have to purge the original job before you can use a copy. This means to me that to do a restore, we have to: 1) identify all the jobs associated with all the files being restored. 2) purge those jobs from the database (which promoted their copies to a restorable state) 3) perform the restore Since I'm trying to make it as easy as possible for my coworkers to restore, having them go through the process of identifying the jobs for a given backup and purging them seems a bit much. I've decided to stick with the dual client method. I've implemented it and tested it. It is working beautifully! We're running CentOS and the Redhat/CentOS init script for bacula-fd as supplied will not allow multiple clients. I had to modify it to allow that. In fact, I will look into submitting my updated init script to the package maintainer. The init script doesn't use the PID files generated by bacula to manage the process and it should, even if you just want to run a single client. Other than that, the only drawback is that when we do perform offsite backups, we are essentially moving that data over the network twice (once for the normal backups and once for the offsite backups.) Since we're not moving a lot of data, this is a non-issue. Thanks again! Eric On Sun, Sep 11, 2011 at 5:55 PM, Rodrigo Renie Braga rodrigore...@gmail.com wrote: Well, I'm really starting to figure this bacula feature yet, but I'd recomend taking a look at Copy Jobs. The ideia would be only running your normal Full/Diff/Inc Backups and then, weekly, create a copy of them on your offsite storage. When restoring, it will require only your normal Full/Diff/Inc backups. Only when these were unavailable (like in a disaster!) that Bacula would automatically require the Offsite Storage. I lost all my documentation and links that I had about Copy Jobs (along with my 1TB HD), but I'm once again taking a look at this feature to implement it on a different company and as soon that I find these documentations again I will send them to you... 2011/9/8 Eric Pratt eric.pr...@etouchpoint.com I'm using Bacula with USB drives to perform offsite backups. I'm trying to create the simplest process possible so in the event I'm unavailable, my coworkers can perform restores with confidence without knowing a whole lot about bacula. Originally, our offsite backups were performed as just a part of the normal schedule for a job called JobName. The schedule was: Run = Level=Full 1st sun at 23:05 Run = Level=Differential 2nd-5th sun at 23:05 Run = Level=Incremental mon-sat at 23:05 Run = Level=Full sun at 10:00 Pool=OffsitePool Any scheduled runs that did not define a pool went to our normal VTL pool. The problem I had there was that the weekly fulls at 10:00AM each sunday became the new baseline for all following differentials and incrementals. The drive with offsite backups is cycled weekly so this means attempting to restore a directory that got blown away Tuesday meant bringing the offsite drive back before restoring. The intention for our offsite backups is purely to give us some form of disaster recovery ability. In this case, they're actually hindering us. Worse, the more we have to bring these onsite, the less likely they will be safe for DR purposes. I want to separate the offsite backups from the normal backups so I removed the last line of the schedule above. Then I created a new job called JobNameOffsite and a new schedule called JobNameOffsiteSchedule that looks like this (the pool is defined in the job to be OffsitePool): Run = Level=Full sun at 10:00 Now, the differentials and incrementals appear to be looking at the full from the 1st sunday of the month and not the weekly offsite fulls for the last full. However, restoring most files still results in attempts to require the offsite backup volumes since it was the last job to use that fileset for a specific path on a given client. Since the offsite disk is not available on any given week, this can pose problems. I can of course work around this and pull files from specific job IDs but I am trying to keep this simple for non-admins to perform restores in my absence. My next idea is to configure a second client on each machine just for offsite backups. If I do this, I can tell bacula to restore from ClientName or ClientNameOffsite. This should provide 100% separation of the normal and offsite backups as well as an easy method for restoring data for those who are not experts with bacula. They would simply choose to restore from ClientName and it's business as usual. Restores from ClientNameOffsite would be only in emergency situations. Even then, it's as simple as bringing the disk on site and choosing to restore from the proper client name. Before I start down this path, does anyone have any
[Bacula-users] Bacula, Offsite, and Restores
I'm using Bacula with USB drives to perform offsite backups. I'm trying to create the simplest process possible so in the event I'm unavailable, my coworkers can perform restores with confidence without knowing a whole lot about bacula. Originally, our offsite backups were performed as just a part of the normal schedule for a job called JobName. The schedule was: Run = Level=Full 1st sun at 23:05 Run = Level=Differential 2nd-5th sun at 23:05 Run = Level=Incremental mon-sat at 23:05 Run = Level=Full sun at 10:00 Pool=OffsitePool Any scheduled runs that did not define a pool went to our normal VTL pool. The problem I had there was that the weekly fulls at 10:00AM each sunday became the new baseline for all following differentials and incrementals. The drive with offsite backups is cycled weekly so this means attempting to restore a directory that got blown away Tuesday meant bringing the offsite drive back before restoring. The intention for our offsite backups is purely to give us some form of disaster recovery ability. In this case, they're actually hindering us. Worse, the more we have to bring these onsite, the less likely they will be safe for DR purposes. I want to separate the offsite backups from the normal backups so I removed the last line of the schedule above. Then I created a new job called JobNameOffsite and a new schedule called JobNameOffsiteSchedule that looks like this (the pool is defined in the job to be OffsitePool): Run = Level=Full sun at 10:00 Now, the differentials and incrementals appear to be looking at the full from the 1st sunday of the month and not the weekly offsite fulls for the last full. However, restoring most files still results in attempts to require the offsite backup volumes since it was the last job to use that fileset for a specific path on a given client. Since the offsite disk is not available on any given week, this can pose problems. I can of course work around this and pull files from specific job IDs but I am trying to keep this simple for non-admins to perform restores in my absence. My next idea is to configure a second client on each machine just for offsite backups. If I do this, I can tell bacula to restore from ClientName or ClientNameOffsite. This should provide 100% separation of the normal and offsite backups as well as an easy method for restoring data for those who are not experts with bacula. They would simply choose to restore from ClientName and it's business as usual. Restores from ClientNameOffsite would be only in emergency situations. Even then, it's as simple as bringing the disk on site and choosing to restore from the proper client name. Before I start down this path, does anyone have any other ideas to keep restores simple while still achieving offsite backups within bacula? Is there some simple way to tag a job as offsite to provide this level of separation without a second client on each machine? Are there any pitfalls with my proposed approach that I should be aware of? I am trying to avoid some things in this process. The primary one is avoiding doing anything outside of bacula. This is tied to the simplicity factor. I can document the restore procedures, but I want to avoid having people lookup volumes tied to JobIDs and performing an rsync to move the volumes into place. I want them to just go into bconsole, restore, and be done with it. Thanks, everyone! Eric -- Doing More with Less: The Next Generation Virtual Desktop What are the key obstacles that have prevented many mid-market businesses from deploying virtual desktops? How do next-generation virtual desktops provide companies an easier-to-deploy, easier-to-manage and more affordable virtual desktop model.http://www.accelacomm.com/jaw/sfnl/114/51426474/ ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] My backup schedule overlaps. Can it be fixed?
On Thu, Sep 1, 2011 at 11:57 AM, Dan Schaefer d...@performanceadmin.com wrote: In English, I want to do a full backup on the 1st and the 16th of every month. I also want to do a differential backup every Sunday and an Incremental Monday-Saturday. Here is my config: Run = Level=Full Pool=Full-Pool Developer-PC on 1,16 at 0:05 Run = Level=Differential Pool=Diff-Pool Developer-PC sun at 0:15 Run = Level=Incremental Pool=Inc-Pool Developer-PC mon-sat at 0:25 Since the Full backups are represented in numbers, the Diff and Inc backups always have overlap with the Full. For example, today Sept 1, a full backup was scheduled to run @ 00:05 and an incremental backup was scheduled @ 0:25. If in fact there were no changes to the files in the fileset, the incremental backup shouldn't have backed up anything. But it runs anyway. Is there a schedule that I can configure that will eliminate backup redundancy? One option is to do a diff on, say, the 8th and 22st of the month and the incremental on every other day not covered. I would, however, like to keep the differentials on Sunday. Thanks in advance. -- Thanks, Dan Schaefer Web Developer/Systems Analyst Performance Administration Corp. ph 800-405-3148 The incremental will run, but it shouldn't back anything up if nothing changed since the last time the job ran. When it runs, it looks to see if anything changed and if not, exits with OK. There is no redundancy there. Check the byte count of the job or list the files backed up by the job to see if it actually moved anything. If the byte count is 0 and/or there are no files in it, then it's working exactly as you expect. However, if your full backup didn't finish before the incremental started, then the incremental could not use the full backup for comparison. It will have used the last completed backup instead. To resolve this, make sure that Maximum Concurrent Jobs in the Job resource to '1'. This is the default so it is already set to '1' unless you've defined it in the Job or a referenced JobDefs resource. Basically, if you're not defining Maximum Concurrent Jobs anywhere, then this second paragraph is not what's happening. Eric -- Special Offer -- Download ArcSight Logger for FREE! Finally, a world-class log management solution at an even better price-free! And you'll get a free Love Thy Logs t-shirt when you download Logger. Secure your free ArcSight Logger TODAY! http://p.sf.net/sfu/arcsisghtdev2dev ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Schedules, Full runs and incremental runs
On Tue, Aug 30, 2011 at 2:02 AM, Christoph Weber we...@mfo.de wrote: Hello Bacula-List, I've got a little problem with the understanding of schedules. I want Bacula to do a full backup once a month on specific dates, which is the monday followed on the last friday of a month, during normal week-days I want to have an incremental run. Schedule { # optional Name = vm-backup-incremental_schedule # required Run = Full jan 31 at 1:00 Run = Full feb 28 at 1:00 Run = Full mar 28 at 1:00 Run = Full may 02 at 1:00 Run = Full may 30 at 1:00 Run = Full jun 27 at 1:00 Run = Full aug 01 at 1:00 Run = Full aug 29 at 1:00 Run = Full oct 02 at 1:00 Run = Full oct 30 at 1:00 Run = Full nov 27 at 1:00 Run = Full jan 01 at 1:00 Run = Incremental mon-fri at 2:00 } The problem is, that I have a full run and an incremental run scheduled on the same day, at 1:00 and 2:00 and the incremental run at 2:00 is not an increment of the 1:00 full backup, but an increment of the last incremental backup preceding the full backup. On the friday before the full tape backup with bacula, I run a VMware full backup which produces several big files, so I have a lot of new data (~1TB) which is now backuped twice with bacula, once with the full backup and once with the incremental backup. I could run the full backup one day before on sunday, but I cannot be sure that it'll be finished before the incremental backup on monday. So how can I achieve that the incremental run is an increment of the full backup and not of the last incremental run? regards, Christoph Weber Try limiting the jobs that can run simultaneously. If your incremental is using the previous incremental as a baseline even though it is scheduled after the full, it is probably because it is kicking off before the full finishes and thus cannot use it as a baseline. If you limit the number of simultaneous jobs to 1, then the full will finish before the incremental and will be used as the baseline. There are a number of ways to limit the number of jobs running or writing to a given storage resource. Use the one that best fits your environment and you should be good to go. Eric -- Special Offer -- Download ArcSight Logger for FREE! Finally, a world-class log management solution at an even better price-free! And you'll get a free Love Thy Logs t-shirt when you download Logger. Secure your free ArcSight Logger TODAY! http://p.sf.net/sfu/arcsisghtdev2dev ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Ghost jobs?
2011/8/23 Carlos André candr...@gmail.com: Running Jobs: Writing: Incremental Backup job SRV10 JobId=61090 Volume= pool=MON device=IBM_TL_LTO3-0 (/dev/IBM_TL_LTO3-0) spooling=0 despooling=0 despool_wait=0 Files=0 Bytes=0 Bytes/sec=0 FDSocket closed --- *cancel jobid=61090 JobId 61090 is not running. Use Job name to cancel inactive jobs. ## BUT these mf... keep showing and f.. with my routines :/ --- Have you tried doing what bacula says to do there and cancel the job by name instead of id? If so, what does it say then? Eric -- EMC VNX: the world's simplest storage, starting under $10K The only unified storage solution that offers unified management Up to 160% more powerful than alternatives and 25% more efficient. Guaranteed. http://p.sf.net/sfu/emc-vnx-dev2dev ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Ignore fileset changes
On Wed, Aug 24, 2011 at 1:50 AM, Olle Romo oller...@gmail.com wrote: On Aug 23, 2011, at 8:43 AM, Adrian Reyer wrote: On Tue, Aug 23, 2011 at 01:38:23AM +0200, Olle Romo wrote: What I mean is that if I have a drive removed, run the job then attach the drive and run the job again, Bacula will do a full backup of the drive even if it previously had done an incremental backup on the same drive. Ideally I want it to just continue the incremental backup. Use 2 filesets and 2 jobs. One like you have now, but exclude the removable drive, The other only has the removable dribe and you only run it when the drive is there, e.g. checked by a pre-script. Regards, Adrian That will be the way to go. Problem is I have quite a few drives that come and go in different combinations. I still wish I could control that particular behavior. Thanks for the tip :) Best, Olle You can. As Adrian says you should be able to use a ClientRunBeforeJob directive. This will tell the client to run a script which should be a shell script that checks for the presence of a drive. If it does not detect the drive, exit with exit code 1 and the job will not run. If it does detect the drive, exit with exit code 0 and the backup job will run. I missed the previous portion of this thread, but I'm assuming this is a Linux client that is mounting a drive. If that's the case, then you're still functionally getting an incremental but because the directory was empty last time (when the drive was not mounted) but now has the entire drive's contents in it. The incremental job is dutifully backing up all changed data from the last incremental which happens to be all of the data on the drive now that Bacula can see it. Merely detecting the drive with a ClientRunBeforeJob script will solve that problem entirely. If you have many drives you interchange in this way, I would recommend making your script take an argument so you only have to maintain a single script that can check any drive you want as dictated by the job. If you do this I don't see why you would need 2 filesets and 2 jobs. It should work just fine with one of each unless that is addressing something I missed from previous emails. -- EMC VNX: the world's simplest storage, starting under $10K The only unified storage solution that offers unified management Up to 160% more powerful than alternatives and 25% more efficient. Guaranteed. http://p.sf.net/sfu/emc-vnx-dev2dev ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] FileStorage strategy
On Tue, Aug 16, 2011 at 12:39 AM, Lyn Amery l...@gaiaresources.com.au wrote: Hi all, I'm just getting started with Bacula and wondering about the best strategy for setting up file storage naming. Is there a best practice, perhaps, or do people have suggestions? I have about 15 systems to back up to about 8 TB of disk. I know that I could create a single volume - labelled say, BIG - or I could go to the other extreme and have a a daily volume for each system, e.g. ServerA_Monday, ServerB_Tuesday. I was thinking of either one for each system (ServerA, ServerB, etc) or by using Use Volume Once, having something like ServerA001, ServerA002, ServerB001, etc. My only reasoning for this is to try and keep things simple. Is there any advantage to having lots of small files as compared to a few large ones? Thanks for any feedback. Cheers, Lyn I'm becoming a big fan of using vchanger to provide a virtual tape library. I was originally using vchanger to make it easier to manage storage on USB drives for offsite rotations, but I'm actually going to start configuring my normal backups to use it too. Basically, I set a maximum volume size and maximum number of volumes in the pool. For instance, if I have a tape library of 100 volumes at 5GB each, then I know I will never exceed 500GB of storage. Of course, you choose volume sizes appropriate to your needs. In your case, with 8TB of backup storage, you might want to consider something like one hundred and sixty 50GB or eighty 100GB volumes. There is always overhead in Bacula for switching from one volume to the other. If you don't have a need for a small file size, then you can reduce the total amount of this overhead by increasing volume size. Make vchanger aware of the max volume count and have it initialize the pre-set number of volumes which later get labeled in bconsole using 'label barcodes'. All normal backup jobs point to the same pool and I don't have to worry about volume naming at all since restores will query the database anyway. The command line tools to restore without database still work on these volumes in a pinch. These become easier to work with if all of the data you're looking to restore is in a single volume It's another good reason for large volume sizes in this configuration. I set volume retention short but job and file retention long. This allows Bacula to automatically recycle the oldest volume when it runs out of space without purging the jobs or files until the oldest volume is actually recycled. If you want to hold onto your data for at least a month and make Bacula prompt you when it can't do that given the pool's capacity, you can set your volume retention to 1 month. That keeps Bacula from overwriting the oldest data too soon. This all amounts to Bacula automatically giving you maximum retention period for the disk space that you have while honoring a minimum retention period as determined by your company's IT policies. This makes the system work a lot more like a tape-based backup system where the tapes have real capacity limits and you don't care about the names of the tapes. It really is a decent and cheap software implementation of a VTL. Now if I could just get some hardware accelerated compression between Bacula and these files, I would be in heaven. Perhaps someone will write a routine to allow Bacula to offload compression to a GPU. Just a wish list item for anyone out there who's looking for something unnecessary (but very cool) to code up . . . Eric -- uberSVN's rich system and user administration capabilities and model configuration take the hassle out of deploying and managing Subversion and the tools developers use with it. Learn more about uberSVN and get a free download at: http://p.sf.net/sfu/wandisco-dev2dev ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Need some help
2011/8/15 Ignacio Cardona ignaciocard...@gmail.com: Dear all, I will be very glad if someone could help me. I have decided after many recommendations to implement Bacula software in the environment of my e-commerce. The version that I'm trying to test is Bacula 5.0.3. I have started reading the docs and all the stuff but I am completly mess up with the information. I have experience in backup tools (TSM, NetBackup, BackupEXEC). But in this case I cannot follow the installation guide so if someone could give me a hand by giving some advices or at least how to begin the instalation (I have downloaded the depackes and the Bacula 5.0.3) but my problem starts when I have to perform de ./configure I don't know how to configure it I mean when I use cat configure it looks like a script I have to edit it or I'm suppose to edit another different file? Could help me with that or give me some kind of guide more organize than the one that is on bacula.com or at least tell me the order in which I am suppose to follow it. Thanks in advance! Don't hesitate in contact me for further questions! Ignacio Cardona. You should check with your operating system's package manager to see if it already has a bacula package. Then you can just install the packages from your package manager and ignore compiling bacula (ie. no need to worry about ./configure.) If your operating system does not have a package, you can always run this command to get help with the configure script: ./configure --help Generally, if you don't think you need any specific compile-time options enabled, you can just type this: ./configure make make install Make sure you are root when you run this. The configure script will automatically scan your system for you. It prepares the software so it can be compiled for your system. Unless you get an error that causes configure to exit, you wont have to do anything more than what's above to get the software configured, compiled, and installed. After that, you can start following the reference manual on configuring your directory, storage, and file daemons. Eric -- uberSVN's rich system and user administration capabilities and model configuration take the hassle out of deploying and managing Subversion and the tools developers use with it. Learn more about uberSVN and get a free download at: http://p.sf.net/sfu/wandisco-dev2dev ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Migrate File-Storage without autochanger to vchanger Skript?
I am in the process of doing exactly this same thing but for a different reason. If only a single job can access your storage at a time, then you should look for a Maximum Concurrent Jobs directive in your configuration files. You can find this in a number of different resources. So just grep for it and track down any resources that might be limiting the number of jobs writing concurrently to the same storage/device. Also be sure to check the manual on Maximum Concurrent Jobs directive for each resource. Even if you're not setting this anywhere, they have default values and there is at least one of these that is '1' by default. From your description, I would check specifically for a Maximum Concurrent Jobs directive set to '1' in the device resource of the storage daemon and/or in the storage resource of the storage daemon. I also thought about the different methods of migrating the data, but decided it wasn't worth doing. I have vchanger auto-create the new volumes with initmag. I then use Bacula's 'label barcodes' command to bulk label the empty volumes vchanger created. I tell the existing Bacula job to start using the new autochanger device but I keep the old volumes on disk until their files and jobs expire. When they do expire, I then remove the old volumes from disk and use 'delete volume' in Bacula. Once you've done that for all old volumes, you're completely migrated to vchanger. The reason I chose this method is that the data in these volumes can remain available for restores until their original job and file retention periods expire. Since you use the same job resource regardless of the storage device, you still get the same file and job retentions automatically applied to the files and jobs stored in the older volumes. This continues to clean your database and allows your backups to continue in the way you want. And of course, you only have minimal work to clean up the old volumes and files. If you have a retention period of a year or less and are not moving large amounts of changed data per night, this is probably a good option for you. If your backups have retention periods in the years and are moving large amounts of changed data per night, this will still work for you, but may not be the best option. Eric On Sun, Aug 7, 2011 at 1:51 AM, Adrian Reyer bacula-li...@lihas.de wrote: Hi, my VirtualFull backups keep failing on me as I have many of them. They run from 2 different File Pools and target 1 autochanger Tape library. No matter how many concurrent jobs I declare and despite I activate Spool Data and it actually does the spooling, only a single job can access a singel File storage at a time. So now one job waits for the autochanger while the other waits for the File and so they are in a deadlock. The idea is now to just use the vchanger and provide multiple File drives to bacula to get rid of the issue. Obviously the File-Volumes are already written and labeled and the naming is incompatibe to the vchanger-names. - Is there a different way to solve this deadlock situation permanently, even if the jobs have to wait some time after being scheduled because of other jobs blocking the devices? - Is there an easy way to make the current File volumes known to the vchanger and Bacula? The current approach would be to stop bacula, move and rename the files and update the occurances of the File names in Job and Media tables. Regards, Adrian -- LiHAS - Adrian Reyer - Hessenwiesenstraße 10 - D-70565 Stuttgart Fon: +49 (7 11) 78 28 50 90 - Fax: +49 (7 11) 78 28 50 91 Mail: li...@lihas.de - Web: http://lihas.de Linux, Netzwerke, Consulting Support - USt-ID: DE 227 816 626 Stuttgart -- BlackBerryreg; DevCon Americas, Oct. 18-20, San Francisco, CA The must-attend event for mobile developers. Connect with experts. Get tools for creating Super Apps. See the latest technologies. Sessions, hands-on labs, demos much more. Register early save! http://p.sf.net/sfu/rim-blackberry-1 ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users -- BlackBerryreg; DevCon Americas, Oct. 18-20, San Francisco, CA The must-attend event for mobile developers. Connect with experts. Get tools for creating Super Apps. See the latest technologies. Sessions, hands-on labs, demos much more. Register early save! http://p.sf.net/sfu/rim-blackberry-1 ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users