Re: bare metal restore, basically
On Wed, Feb 28, 2007 at 11:07:04AM +0100, Paul Bijnens wrote: On 2007-02-27 22:34, Byarlay, Wayne A. wrote: Is there a method I can use to boot from, say, Ubuntu, or knoppix, or some other CD-based OS, run the AMANDA client, connect to the server where the latest backups reside, and basically rebuild this machine with only one /dev/sda? Install a new hard disk. Boot with your bootable CD, partition it, and make the filesystems (mke2fs etc.) Having done these restores numerous times, you will save yourself a lot of consternation if you download the latest Knoppix and use it to do the restore. Before you start, run aptitute update; aptitude install dump amanda-client. The reason you want to do this: a recent change in dump's file format breaks restoring using older copies of restore. If you use the latest Knoppix, you'll save waiting a long time of download upgrades, but you'll still be able to pull down the latest amanda and dump. -- Ross Vandegrift [EMAIL PROTECTED] The good Christian should beware of mathematicians, and all those who make empty prophecies. The danger already exists that the mathematicians have made a covenant with the devil to darken the spirit and to confine man in the bonds of Hell. --St. Augustine, De Genesi ad Litteram, Book II, xviii, 37
Re: Feature suggestion : test restore
On Wed, Jan 31, 2007 at 06:02:58PM -0700, Glenn English wrote: Stefan G. Weichinger wrote: Jon LaBadie schrieb: - assuming the restore is done to the server; the tools used to create the dump may not exist on the server or be in different locations An amanda-admin has to know if he wants to do these tests, and after that he has to provide ressources (diskspace ...) and software (binaries like tar, dump, gzip ...) on the machine he wants to perform these tests. It'd never occurred to me there might be such a thing as a write-only backup server before. And I always restore to the server anyway -- it's in the same room as the tape drive... All of the backup servers that I used to run were write-only. It becomes a major issue with recent changes in RHEL4. Some of our Amanda servers were RHEL3. A new version of dump was integrated into RHEL4 which broke binary compatability from the old version. This split the versions and restore would give really scary errors about unable to find start of tape if you attempted a bad combination. -- Ross Vandegrift [EMAIL PROTECTED] The good Christian should beware of mathematicians, and all those who make empty prophecies. The danger already exists that the mathematicians have made a covenant with the devil to darken the spirit and to confine man in the bonds of Hell. --St. Augustine, De Genesi ad Litteram, Book II, xviii, 37
Slow performance with dump + LVM
Hi everyone, I'm working on setting up some Amanda backups at my house and have run into a serious performance problem. Both the client and the server machine are using LVM2 for their disks. Though there's 320GiB of data to backup, this is a bit ridiculous: sendbackup: time 8008.486: 87: normal(|): DUMP: 8.07% done at 3472 kB/s, finished in 24:40 I've done tons of dumps to LVM volumes on a server with no speed problems, so I'm investigating the client. dumping on the client to /dev/null gives nearly the same performance, so I'm confident that it's an issue with dump + LVM on the client. [1] The client is all but idle, not doing compression/encryption, and only has two PVs in the VG, both of which are fast SATA disks. If I increase the blocksize of dump to 64kiB (the largest the manpage says is smart), performance gets better, in the range of 15-20M/s. Has anyone seen performance this bad from dump before? Is there some tunable to get things running a little faster? 24 hours of a full backup seems just terrible [1] - So yea, technically this the wrong place to ask, the LVM mailing lists seem to be all but completely dead. On the other hand, lots of people know stuff about dump here! -- Ross Vandegrift [EMAIL PROTECTED] The good Christian should beware of mathematicians, and all those who make empty prophecies. The danger already exists that the mathematicians have made a covenant with the devil to darken the spirit and to confine man in the bonds of Hell. --St. Augustine, De Genesi ad Litteram, Book II, xviii, 37
Re: Slow performance with dump + LVM
On Tue, Jan 23, 2007 at 01:01:38AM -0500, Gene Heskett wrote: Well, since dump works at the partition level, it may be that dump and LVM aren't compatible. Switch to tar, which is file oriented see what happens. Looks like I'm hitting nearly the same speed situation. I backed up 7932MiB in 2460 seconds (I got impatient...) which comes out to about 3-4MiB/s. Since tar behaves the same way, I think I'm going to play around with various record sizes when reading from LVM. Something in the blokc layer is shooting my transfer rates. -- Ross Vandegrift [EMAIL PROTECTED] The good Christian should beware of mathematicians, and all those who make empty prophecies. The danger already exists that the mathematicians have made a covenant with the devil to darken the spirit and to confine man in the bonds of Hell. --St. Augustine, De Genesi ad Litteram, Book II, xviii, 37
Re: amandas group membership in FC6?
On Sat, Nov 25, 2006 at 11:22:38PM -0500, Gene Heskett wrote: See what that number maps to in /etc/group. I'm betting it goes to an 'amanda' group and not the 'disk' group. There was not, and still is not, a group named amanda, just the amanda entry in the disk line. Is there any chance you're using ldap/nis/winbind/etc for groups in nsswitch.conf? The amanda group must be coming from somewhere. If it's not listed in /etc/group, I'm wondering if there's another source of groups on your system. -- Ross Vandegrift [EMAIL PROTECTED] The good Christian should beware of mathematicians, and all those who make empty prophecies. The danger already exists that the mathematicians have made a covenant with the devil to darken the spirit and to confine man in the bonds of Hell. --St. Augustine, De Genesi ad Litteram, Book II, xviii, 37
Re: confused and concerned about virtual tape sizes
On Fri, Nov 10, 2006 at 02:43:13PM -0500, René Kanters wrote: I backup a small directory (/home/rene at about 2.1MB) and when I run amdump I get messages with (excerpts): [snip] i.e., the full size of my (compressed) /home/rene You're running up against Amanda being too smart! Since it seems that your data set is very small, it is doing a full backup every time. Why bother to make an incremental backup of 2.1MiB? By default, Amanda must get a 10MiB savings before it will bump to the next incremental level. You can control the bumping with bumpsize, bumppercent, bumpmult, and bumpdays. The short, though, throw some real data at it and you'll see some real performance! -- Ross Vandegrift [EMAIL PROTECTED] The good Christian should beware of mathematicians, and all those who make empty prophecies. The danger already exists that the mathematicians have made a covenant with the devil to darken the spirit and to confine man in the bonds of Hell. --St. Augustine, De Genesi ad Litteram, Book II, xviii, 37
Re: Looooooooong backup
On Wed, Oct 18, 2006 at 01:14:57PM -0400, Steven Settlemyre wrote: I have a monthly (full) backup running for about 22 hrs now. Do you think there is a problem, or is it possible it's just taking a long time? about 150G of data. Is it closer to 150G in on file or 150G in one million files? As the number of files increases, the time it takes to do dumps/restores becomes dominated by the amount of time it takes to deal with individual files. Millions of files add up to extremely long dumps... -- Ross Vandegrift [EMAIL PROTECTED] The good Christian should beware of mathematicians, and all those who make empty prophecies. The danger already exists that the mathematicians have made a covenant with the devil to darken the spirit and to confine man in the bonds of Hell. --St. Augustine, De Genesi ad Litteram, Book II, xviii, 37
Re: bizarre bug fixed but not explained
On Fri, Sep 29, 2006 at 06:47:42PM -0400, Jon LaBadie wrote: On Fri, Sep 29, 2006 at 06:01:46PM -0400, Steve Newcomb wrote: chown root.disk /home/amanda/libexec/runtar All the chown's I've used require a colon (':'), not a period ('.') between the user and group. It's a BSD-ism that's been inherited into the GNU tools. POSIX technically disallows it because a username is allowed to contain dots. Check the documentation for fileutils for more references. Might not be a bad idea to convert that to colon since Amanda can run on many platforms that are both non-BSD and non-GNU. -- Ross Vandegrift [EMAIL PROTECTED] The good Christian should beware of mathematicians, and all those who make empty prophecies. The danger already exists that the mathematicians have made a covenant with the devil to darken the spirit and to confine man in the bonds of Hell. --St. Augustine, De Genesi ad Litteram, Book II, xviii, 37
Re: restoring from DVDs
On Sun, Aug 06, 2006 at 09:41:39AM +0200, Geert Uytterhoeven wrote: I've had quite good luck doing poor man's data recovery. Boot the machine into Knoppix or like ilk and use dd_rescue to copy the disk to an image file or another disk. dd_rescue is smart about skipping areas of the disk it cannot read instead of giving up. It can take a long time, but I've recovered quite a bit of data with that sucker. But this simple methods won't work, as the disk used to be part of a RAID0 setup, and thus contains only half of the data. Then it depends on the stripe size: the larger it is, the more likely you can find useful pieces of data (e.g. a complete password or credit card number). Well, no reason you couldn't apply this principle to the RAID0 data. Suppose your disks are /dev/sda1 and /dev/sdb1, sdb has failed: # dd_rescue /dev/sdb1 some_file_or_device (wait a long time) (if you're using a file # losetup /dev/loop/0 some_file) # mdadm --assemble /dev/md0 /dev/sda1 /dev/whatever If the superblock of the second disk is intact, mdadm should be able to assemble the array with no problems. Of course you'll run into files with corrupt data, as you know one half of the stuff is damaged. If you're unlucky, the filesystem data was damaged and you'll have either lost everything or will need to run fsck... -- Ross Vandegrift [EMAIL PROTECTED] The good Christian should beware of mathematicians, and all those who make empty prophecies. The danger already exists that the mathematicians have made a covenant with the devil to darken the spirit and to confine man in the bonds of Hell. --St. Augustine, De Genesi ad Litteram, Book II, xviii, 37
Re: restoring from DVDs
On Sun, Aug 06, 2006 at 03:13:39PM +0100, Laurence Darby wrote: This is besides the point, but that wont work, and I'm not sure if you completely understand the situation. I *DON'T* want data to be recovered from the drive, and because I'm returning it under warranty, I want the data on it to be destroyed. Aha! I did misunderstand. You shouldn't worry about this - the manufacturers are bound by process requirements that prevent them from recovering data on RMAed drives. My employer is going through the process of becoming certified for various government/banking/etc security processes. Part of it is verifying that sensitive data is not recoverable. According to our compliance manager, sending a drive to RMA places data security liability on the RMA facility as soon as they acceept the package. You'd be able to sue for damages if you did have plaintext passwords that they leaked. This still leaves a hole while the disk is in transit, but you can only do so much... -- Ross Vandegrift [EMAIL PROTECTED] The good Christian should beware of mathematicians, and all those who make empty prophecies. The danger already exists that the mathematicians have made a covenant with the devil to darken the spirit and to confine man in the bonds of Hell. --St. Augustine, De Genesi ad Litteram, Book II, xviii, 37
Re: restoring from DVDs
On Sat, Aug 05, 2006 at 12:29:43AM +0100, Laurence Darby wrote: BTW, it was actually one disk of nice and fast RAID 0 (so I'm restoring to the one good disk). Does anybody know if data recovery from it would be possible? I hope *not*, since I'm sending it back under waranty, and I can't erase it cos its dead, although it sounded like the platter might be all scratched up... As always, it depends on what you want to pay. If you have the money to burn, just about anything besdies physical platter destruction/degaussing can be recovered. I read an article not that long ago about recovering a hard disk that had been burned in a fire. I've had quite good luck doing poor man's data recovery. Boot the machine into Knoppix or like ilk and use dd_rescue to copy the disk to an image file or another disk. dd_rescue is smart about skipping areas of the disk it cannot read instead of giving up. It can take a long time, but I've recovered quite a bit of data with that sucker. Good luck! -- Ross Vandegrift [EMAIL PROTECTED] The good Christian should beware of mathematicians, and all those who make empty prophecies. The danger already exists that the mathematicians have made a covenant with the devil to darken the spirit and to confine man in the bonds of Hell. --St. Augustine, De Genesi ad Litteram, Book II, xviii, 37
Re: Recovering files from DVD-recorded backup
On Mon, Jul 31, 2006 at 06:56:22PM +0100, Anne Wilson wrote: It seems to me that once vtapes are free for overwriting the corresponding index will be lost too. Can either amrestore or amrecover actually access these files? Barring defects in the file/media, amrestore can *always* be used to recover an amanda backup. The procedure isn't quite as automated as amrecover. Here's an outline of what you might need to do: 1) Determine what tape you actually need. You might not know this for real the first time! Trial and error may be required. 2) Mount the DVD, say at /mnt. 3) To figure out what backups are on the tape, run something like this: # amrestore -f 0 -p file:/mnt/tapeX invalid-host /dev/null This says to start at block 0, of the tape file:/mnt/tapeX (path depends on how you named tapes/burned them), search for invalid-host and throw the data to /dev/null. This command will give you a printout of the contents of the tape like this: amrestore: 0: skipping start of tape: date 20050128 label DailySet117 amrestore: 1: skipping blah.sda1.20050128.2 ... This output is formatted like hostname.device.date.dumplevel. Find the dump that you are looking for. Let's assume we want to restore from blah.sda1.20050128.2. 4) Now we run a valid amrestore command to get data: # amrestore -f 1 -p file:/mnt/tapeX blah sda1 | restore -if - This pulls data for blah:/sda1 from file:/mnt/tapeX starting at block 1. You'll get a shell very much like amrecover that you can navigate the filesystem dump, add files, and extract them. If you use tar, you'll want to do something like: # amrestore -f 1 -p file:/mnt/tapeX blah sda1 | tar xpvf - to untar the tarball into your current directory. -- Ross Vandegrift [EMAIL PROTECTED] The good Christian should beware of mathematicians, and all those who make empty prophecies. The danger already exists that the mathematicians have made a covenant with the devil to darken the spirit and to confine man in the bonds of Hell. --St. Augustine, De Genesi ad Litteram, Book II, xviii, 37
Re: What is the highest incremental/differential level performed by amanda?
On Tue, Jul 25, 2006 at 03:19:31PM -0400, Jon LaBadie wrote: Is there such a limit in Amanda anywhere? In planner.c there is code deciding upon level that basically says if I'm thinking of doing level 9, don't even consider anything higher, return level 9 as the decision. And in lots of other code, like reports, cleanup, amadmin, ... there are things like for (level = 0; level = 9; level++). Makes sense - Amanda runs on plenty of platforms where dump levels higher than 9 probably freak out the software. My take from your questions is that you want or expect to get high incremental levels. Your needs may differ of course, but I'm happy that my incrementals seldom even get to level 3. It reduces the number of tapes needed for a recovery and changed data often appears in multiple incrementals at the same level for redundancy. With vtapes, there's not much of a downside to having higher levels. No physical tape switching means that I just have to repoint to a different directory for the next tape. In that sense, having dumps hit levels 5 and 6 is a quite good thing. The space savings from going higher than that are diminishing returns for the servers I run, but I can imagine a highly volatile load where that had good benefits. Perhaps very important for your need might be bumpdays which defaults to 2 days. This is the minimum number of dumps at each incremental level before amanda will bump to the next level. Dropping it to 1 is a good first step. Dropping bumpdays to 1 made a major difference in terms of space usage for me. It helped a lot. The other parameter that helped me a lot was bumpmult. It had bee originaly configured as 4 (not sure if that's a default or not) and I dropped it to 2. This helps Amanda spend less time on larger level 0 and 1 dumps and more quickly gets us to the more space efficient levels. -- Ross Vandegrift [EMAIL PROTECTED] The good Christian should beware of mathematicians, and all those who make empty prophecies. The danger already exists that the mathematicians have made a covenant with the devil to darken the spirit and to confine man in the bonds of Hell. --St. Augustine, De Genesi ad Litteram, Book II, xviii, 37
Re: What is the highest incremental/differential level performed by amanda?
On Tue, Jul 25, 2006 at 11:42:12AM -0400, Joshua Baker-LePain wrote: Look at the various bump* options in amanda.conf -- they control when amanda bumps an incremental up a level. Theoretically it can go to 9 (just like the backup tools), but there's an increasingly high requirement for backup image size savings to do the bumps. Is there such a limit in Amanda anywhere? For example, dump from Linux's e2fstools support arbitrary integers as dump levels. From the manpage: A level number above 0, incremental backup, tells dump to copy all files new or modified since the last dump of a lower level. The default level is 9. Historically only levels 0 to 9 were usable in dump, this version is able to understand any integer as a dump level. My Amandas don't ever get above a 6, and my predecessor's bump settings limited runs to level 3, so I'm not saying level 897125 backups are a good idea ::-) -- Ross Vandegrift [EMAIL PROTECTED] The good Christian should beware of mathematicians, and all those who make empty prophecies. The danger already exists that the mathematicians have made a covenant with the devil to darken the spirit and to confine man in the bonds of Hell. --St. Augustine, De Genesi ad Litteram, Book II, xviii, 37
Re: backing up to DVD-Rs
On Wed, Jul 12, 2006 at 06:31:52PM +0100, Anne Wilson wrote: On Wednesday 12 July 2006 18:17, Ross Vandegrift wrote: Make a holding area, and use dump with a tape size of just slightly less than the DVD-R size. Tell dump to generate manifests and quick file access data. Then, all you have to do is burn each dump file to a DVD-R in pre-prepared mode (check man growisofs for details on this) and burn/store the index/QFA data somewhere. Would you like to amplify this, please? Sure. Assume that you've got /dev/sda1 that you'd like to backup and you'll store the images in /holding. Then it goes a little something like this: # cd /holding # dump -0uy -A MANIFEST -B 4403200 -Q QFA /dev/sda1 -f disc1,disc2,disc3,... ... (wait for dump to dump the filesystem) # growisofs -dvd-compat -Z /dev/dvd=disc1 ... (burn all the images) The dump command generates files named disc1, disc2, etc (you may need more than three depending on your filesystem), as well as a restore index called MANIFEST and quick file access info named QFA. You can do an incremental by changing the -0 to some higher integer. You may need to twiddle with -B, the size param. I don't remember exactly how large it can be. Then, restoring from DVD looks like this: # cd /destination # restore -A MANIFEST -Q QFA -i -f /dev/dvd ... (restore as usual, inserting discX when asked for volume X) That's all there is to my personal backup system. It has one *major* flaw: backups are not automated the way they are with Amanda. I guess if you're burning to DVD, you have to manually interact anyhow. But as a result, I have only ever done two complete backups of my system... On the other hand, I did use such a backup to recover from a catastrophic disk crash and it worked very well. -- Ross Vandegrift [EMAIL PROTECTED] The good Christian should beware of mathematicians, and all those who make empty prophecies. The danger already exists that the mathematicians have made a covenant with the devil to darken the spirit and to confine man in the bonds of Hell. --St. Augustine, De Genesi ad Litteram, Book II, xviii, 37
Re: backing up to DVD-Rs
On Wed, Jul 12, 2006 at 10:29:14AM +0100, Laurence Darby wrote: Bascially I have a 30GB directory (/home) and 50 4.7GB DVD-Rs. Is there any support in Amanda for backing up to DVD-Rs? Actually, the only thing I'm looking for is the capability to divide the 30GB into 4.7GB chunks, or slightly smaller to fit on a 4.7GB ext2 loopback file which can be burned to DVD. Is this the only disklist item you'd have? Also, how frequently will you be running backups? Basically, I found that Amanda with DVD-Rs was too much trouble for backing up my home system. Rather, I run dump manually. Make a holding area, and use dump with a tape size of just slightly less than the DVD-R size. Tell dump to generate manifests and quick file access data. Then, all you have to do is burn each dump file to a DVD-R in pre-prepared mode (check man growisofs for details on this) and burn/store the index/QFA data somewhere. Once you have that burnt, restores are easy. You just tell restore to restore from your DVD device. When it asks for tape number X, you insert DVD X and it does its thing. It works quite well and prevent having to juggle all the information that Amanda is going to keep about your filesystems. -- Ross Vandegrift [EMAIL PROTECTED] The good Christian should beware of mathematicians, and all those who make empty prophecies. The danger already exists that the mathematicians have made a covenant with the devil to darken the spirit and to confine man in the bonds of Hell. --St. Augustine, De Genesi ad Litteram, Book II, xviii, 37
Re: LVM snapshots
On Fri, Jul 07, 2006 at 02:35:23AM -0700, Joe Donner (sent by Nabble.com) wrote: I'm just wondering what happens during the freeze - how freezing all activity to and from the filesystem to reduce the risk of problems affects the system? One would imagine that disk writes are somehow queued up and complete when the file system is unfreezed again? Yes - when you configure an LVM snapshot, you provide another device to which changes are written. LVM creates a new device node that points to the snapshot of the original device. The rest of the system basically sees everything as normal. LVM handles all the changes between the disks until the snapshot is removed. I don't run LVM snapshots with Amanda, but be aware that if you wanted to do this, there'd be a few hoops to jump though. You'd need to specify the snapshot device name in the DLE, not the normal one. As a result, you'll need to somehow ensure that the snapshot was sucessfully created before running amdump. -- Ross Vandegrift [EMAIL PROTECTED] The good Christian should beware of mathematicians, and all those who make empty prophecies. The danger already exists that the mathematicians have made a covenant with the devil to darken the spirit and to confine man in the bonds of Hell. --St. Augustine, De Genesi ad Litteram, Book II, xviii, 37
Re: Amanda on any of the Linux live CDs?
On Fri, Jun 23, 2006 at 01:45:10PM -0500, Frank Smith wrote: I will start my own research this weekend but just in case anyone has done it, I will ask. Do any of the Linux live CDs have amanda installed? It's on the Knoppix live DVD, but not the CD version. This is correct, but remember that Knoppix has a working apt-get configuration. All you need to do is run apt-get update apt-get install amanda-client amanda-server and you've got it. Makes for a great recovery environment. -- Ross Vandegrift [EMAIL PROTECTED] The good Christian should beware of mathematicians, and all those who make empty prophecies. The danger already exists that the mathematicians have made a covenant with the devil to darken the spirit and to confine man in the bonds of Hell. --St. Augustine, De Genesi ad Litteram, Book II, xviii, 37
Re: Disagreements between amplot, amstatus
On Tue, May 30, 2006 at 01:50:21PM -0700, Pavel Pragin wrote: But amplot's Bandwidth Allocated graph shows the line mostly pegged at 0% the entire time. From glancing through the amplot scripts, I'm guessing this should be labeled Bandwidth Free? If you run amstatus again what is the value you have in the field below: network free kps: ? also what is your inparallel and maxdumps set to in amanda.conf? network free kps:189380 inparallel 15 maxdumps 2 I don't think maxdumps will affect much - my DLEs happen to almost always be single spindle boxes. As a result, disks don't happen simultaneously on the same host. That's okay with me though. How is network free kps calulated? I played a little bit yesterday with figuring out how 80 Mbps is related to 189380 but couldn't really come up with anything... -- Ross Vandegrift [EMAIL PROTECTED] The good Christian should beware of mathematicians, and all those who make empty prophecies. The danger already exists that the mathematicians have made a covenant with the devil to darken the spirit and to confine man in the bonds of Hell. --St. Augustine, De Genesi ad Litteram, Book II, xviii, 37
Disagreements between amplot, amstatus
Hello everyone, I'm working on getting one of my Amanda servers to finish dumping a little bit faster. Right now, all but one of my hosts are taking around four hours to finish. Anyhow, I noticed that amstatus says this after the run: 0 dumpers busy : 0:33:39 ( 3.64%)not-idle: 0:33:39 (100.00%) 1 dumper busy : 9:38:18 ( 62.47%)not-idle: 9:37:29 ( 99.86%) no-bandwidth: 0:00:49 ( 0.14%) 2 dumpers busy : 0:28:11 ( 3.05%)not-idle: 0:27:57 ( 99.18%) no-bandwidth: 0:00:13 ( 0.82%) 3 dumpers busy : 0:26:59 ( 2.92%)no-bandwidth: 0:22:59 ( 85.20%) not-idle: 0:03:59 ( 14.80%) 4 dumpers busy : 2:13:20 ( 14.40%)no-bandwidth: 1:56:04 ( 87.06%) not-idle: 0:17:15 ( 12.94%) 5 dumpers busy : 1:12:40 ( 7.85%)no-bandwidth: 1:12:40 (100.00%) 6 dumpers busy : 0:32:23 ( 3.50%)no-bandwidth: 0:32:23 (100.00%) 7 dumpers busy : 0:03:03 ( 0.33%)no-bandwidth: 0:03:03 (100.00%) 8 dumpers busy : 0:00:27 ( 0.05%)no-bandwidth: 0:00:27 (100.00%) 9 dumpers busy : 0:00:48 ( 0.09%)no-bandwidth: 0:00:48 (100.00%) 10 dumpers busy : 0:00:09 ( 0.02%)no-bandwidth: 0:00:09 (100.00%) Looks like I have too many dumpers for my bandwidth, no? FWIW, netusage is set to 80 Mbps. But amplot's Bandwidth Allocated graph shows the line mostly pegged at 0% the entire time. From glancing through the amplot scripts, I'm guessing this should be labeled Bandwidth Free? -- Ross Vandegrift [EMAIL PROTECTED] The good Christian should beware of mathematicians, and all those who make empty prophecies. The danger already exists that the mathematicians have made a covenant with the devil to darken the spirit and to confine man in the bonds of Hell. --St. Augustine, De Genesi ad Litteram, Book II, xviii, 37
Re: Intermittent client failures reading /var/backups/.amandahosts
On Mon, May 29, 2006 at 11:07:29AM +0200, Paul Bijnens wrote: Strange. Is /var/backup a local filesystem? Or mounted via NFS? Is the user backup defined in the local password file or via yp/ldap? Both are local, filesystem and user. Is the DNS consistently giving back the same hostname? (linuxbackup2.smarterlinux.com, always FQDN too?) Has one of the hosts two interface cards, and are they reverse resolved identically? (if not, add lines to .amandahosts) Yep - we have /etc/hosts entries that ensure they are resolved properly, to the correct interface, and have appropriate reverse mappings. That's what's weird about it - it's a pretty simple setup... -- Ross Vandegrift [EMAIL PROTECTED] The good Christian should beware of mathematicians, and all those who make empty prophecies. The danger already exists that the mathematicians have made a covenant with the devil to darken the spirit and to confine man in the bonds of Hell. --St. Augustine, De Genesi ad Litteram, Book II, xviii, 37
Re: Intermittent client failures reading /var/backups/.amandahosts
On Mon, May 29, 2006 at 01:26:08PM +0100, Paul Haldane wrote: As the error reported is a failure to open .amandahosts rather than a problem with the contents of the file (which would result in a amandahostsauth failed error) a likely cause (given that the problem is intermittent) is running out of system resources (file descriptors/memory/?). Ahhh, this is much more likely the cause. These servers are very busy and generally do not get much of a break. You should find more information about the exact cause of the failure in the amanda debug file on the client - that should show the same error along with the system error message. Dang - Debian woody puts those logs in /tmp/amanda, which gets cleaned automatically. Is there any way I can change the log directory on the client programs at runtime? -- Ross Vandegrift [EMAIL PROTECTED] The good Christian should beware of mathematicians, and all those who make empty prophecies. The danger already exists that the mathematicians have made a covenant with the devil to darken the spirit and to confine man in the bonds of Hell. --St. Augustine, De Genesi ad Litteram, Book II, xviii, 37
Intermittent client failures reading /var/backups/.amandahosts
Hello everyone, Approximately once a month, some of my servers fail to make backups with this error message: FAILURE AND STRANGE DUMP SUMMARY: capote sda8 lev 1 FAILED [ [access as backup not allowed from [EMAIL PROTECTED] open of /var/backups/.amandahosts failed] But, there's simply no way this error message is correct. I have two reasons: 1) This host has three other DLEs that worked fine: capote sda1 1 4640547 11.8 0:07 81.6 0:00 93138.1 capote sda7 1 117750 8472 7.2 0:27 313.9 0:00 190309.3 capote sda8 1 FAILED -- capote sda9 31633400 340849 20.9 13:11 431.2 0:1131918.6 2) The permissions on /, /var, /var/backups/.amandahosts, /etc, /etc/amandahosts are all correct and none have been updated in at least a week. I've seen this before on other servers, and the major common fact between them: they are all Debian woody servers running amanda-client 2.4.2p2-4. I'd love to upgrade this, but upgrading woody's amanda-client without upgrading glibc has proven very, very challenging. Is there any chance this is a known issue with the old client software? Maybe there's something else that could be failing? -- Ross Vandegrift [EMAIL PROTECTED] The good Christian should beware of mathematicians, and all those who make empty prophecies. The danger already exists that the mathematicians have made a covenant with the devil to darken the spirit and to confine man in the bonds of Hell. --St. Augustine, De Genesi ad Litteram, Book II, xviii, 37
Re: Parallel dumps of a single file system?
On Tue, May 23, 2006 at 11:04:44AM -0400, Jon LaBadie wrote: Recently I had a look at amplot results for my new vtape setup. One thing it showed was that for 2/3 of the time, only one of the default four dumpers was active. This is a good point. amplot is awesome for checking out what kind of stuff is slowing down your backups! Also check the output at the end of amstatus when a run is finished. It'll give you a summary of the same information. But there's nothing like a cool graph! As far as the original poster's question: I think you should try it out. Whether it's a performance win or loss is going to depend heavily on how the data has ended up across those disks. Your RAID5 performance is always dominated by the time it takes to seek for data. If all n disks can just just stream for a while, you get full streaming performance from the disks. But if even one of them needs to seek to find its blocks, you're going to have to wait until that disk finishes. This makes me think that in most cases, dumping a big RAID5 in parallell would hurt performance. However, if your array is old, it may be highly fragmented. The extra I/O requests might be smoothed over by an elevator algorith somewhere, and you might fit more data into the same time... I'd say it calls for an experiment. -- Ross Vandegrift [EMAIL PROTECTED] The good Christian should beware of mathematicians, and all those who make empty prophecies. The danger already exists that the mathematicians have made a covenant with the devil to darken the spirit and to confine man in the bonds of Hell. --St. Augustine, De Genesi ad Litteram, Book II, xviii, 37
Re: vtape, end of tape waste
On Tue, May 23, 2006 at 04:28:31PM -0400, Jon LaBadie wrote: But running out of disk space caused me to look more closely at the situation and I realized that the failed taping is left on the disk. This of course mimics what happens on physical tape. However with the file:driver if this failed, and useless tape file were deleted, it would free up space for other data. Our setups avoid this situation by having a dedicated chunk of holding space. I cut out 150-200GiB on each Amanda server just for holding space. That way, no dump ever fails because the data in holding was using space in the vtape filesystem. I know throwing more hardware/disk space at the problem isn't a particularly interesting or clever solution, but I can vouch for the fact that it works! -- Ross Vandegrift [EMAIL PROTECTED] The good Christian should beware of mathematicians, and all those who make empty prophecies. The danger already exists that the mathematicians have made a covenant with the devil to darken the spirit and to confine man in the bonds of Hell. --St. Augustine, De Genesi ad Litteram, Book II, xviii, 37
Re: windows vshadow
On Tue, May 16, 2006 at 10:05:22AM -0400, Jon LaBadie wrote: The vshadow.exe is only available for WinXP and 2003 server. So it is not a general solution, but maybe a start? Is this related to their Volume Management Services? IIRC, this interacts with a backgroup service like LVM or Veritas. LEts you create virtual volumes and the like. Does it enable you to take snapshots of existing filesystems without rebuilding them into managed virtual volumes (or whatever the MS verbiage for these things is)? For what it's worth, if it is the same thing that I have heard of, I've heard it works ok. -- Ross Vandegrift [EMAIL PROTECTED] The good Christian should beware of mathematicians, and all those who make empty prophecies. The danger already exists that the mathematicians have made a covenant with the devil to darken the spirit and to confine man in the bonds of Hell. --St. Augustine, De Genesi ad Litteram, Book II, xviii, 37
Re: Would like to get list's impression on (amanda 4TB backup)
On Tue, May 16, 2006 at 11:17:34AM -0400, Jon LaBadie wrote: On Tue, May 16, 2006 at 09:53:13AM -0500, Gordon J. Mills III wrote: here are some of my comments about TSM incremental forever is a feature I've heard of in some other backup systems. I wouldn't expect that to fit the mold of amanda as recovery would then truly demand amanda software and indexes to be present. I've been thinking about a new vtape changer that might make it easy to do this kind of incremental forever backup. One issue that I run into: Most of my system have 18-20 virtual tapes. runtapes is 2, but I really try hard to keep things at one tape. We need to kep to weeks of backups. Suppose some event happens and someone needs to make an emergency backup of what's on a server, right now. They run amdump, it clears the oldest tape. If this happens too many times in a dumpcycle, oops, we don't have our two weeks of backups. I've been thinking about a tape changer that uses a timestamp for the label. No tape is ever reused. When amdump runs, something creates a new tape, labels it, and loads it. The backup is done to this tape. You then have a cronjob that runs nightly to cull old backups. Give it a timeframe, it deletes any tapes that are older than the timeframe. If you had a model like this, where tapes could stick around forever if nothing deleted them, you'd just need a flag that a given timestamp is a full backup and should never be thrown out. With many people doing amanda backups to vtapes, it would be nice to be able to archive desired parts to offsite ptapes. Shouldn't this be easy if you plan to make your vtapes the same size as your ptapes? -- Ross Vandegrift [EMAIL PROTECTED] The good Christian should beware of mathematicians, and all those who make empty prophecies. The danger already exists that the mathematicians have made a covenant with the devil to darken the spirit and to confine man in the bonds of Hell. --St. Augustine, De Genesi ad Litteram, Book II, xviii, 37
Re: Would like to get list's impression on (amanda 4TB backup)
On Tue, May 16, 2006 at 12:31:57PM -0400, Jon LaBadie wrote: On Tue, May 16, 2006 at 11:58:55AM -0400, Ross Vandegrift wrote: I've been thinking about a tape changer that uses a timestamp for the label. No tape is ever reused. When amdump runs, something creates a new tape, labels it, and loads it. The backup is done to this tape. Couldn't this be done now with the autolabel feature and a V.large tapecycle that always called for a new tape? Ooo! That's exciting, I didn't know that there was such a think as autolabel. Ironically, the first google link when searching for amanda autolabel is about TSM. I checked the stuff on www.amanda.org but I didn't find that option. Could I get some more info? Shouldn't this be easy if you plan to make your vtapes the same size as your ptapes? The logs+indexes might be a problem if it was going to a different config. And I was thinking of maybe selecting what I wanted, like all level 0's of a particular host:disk. Or as of a specific date, one level 0 plus incrementals up to that date. Oh that would be a cool tool. It would solve a lot of problems for smaller installations too. Say I want to backup my workstation and have the files accessible most of the time for easy restore. However, I do want occasional archival copies to tape/DVD/whatever. -- Ross Vandegrift [EMAIL PROTECTED] The good Christian should beware of mathematicians, and all those who make empty prophecies. The danger already exists that the mathematicians have made a covenant with the devil to darken the spirit and to confine man in the bonds of Hell. --St. Augustine, De Genesi ad Litteram, Book II, xviii, 37
Subset dumps?
Hello everyone, Been a bad week for my backups. Today, we had a mostly catastrophic loss of a vtape filesystem, lost 14 of 18 vtapes on our busiest Amanda server. I'm just finishing picking up the pieces now and getting things back on track. With amdump I can: 1) Run dumps for an entire config: amdump master 2) Run dumps for a particular host: amdump hostname 3) Run dumps for a specific DLE: amdump hostname partition Is there any way to run dumps for a subset of disklist entries? For example, is there someway I can do amdump [a-h]*? We want to make a lot of full backups to make up for what we've lost, and so we can get back into incrementals ASAP. Forcing everything to full greatly outpaces our holding space. It would have almost been easier to do full dumps that would fill our holding disk until we finished all the hosts, but I couldn't figure out a good way to accomplish that... -- Ross Vandegrift [EMAIL PROTECTED] The good Christian should beware of mathematicians, and all those who make empty prophecies. The danger already exists that the mathematicians have made a covenant with the devil to darken the spirit and to confine man in the bonds of Hell. --St. Augustine, De Genesi ad Litteram, Book II, xviii, 37
Re: Subset dumps?
On Tue, May 09, 2006 at 11:35:45AM +0200, Paul Bijnens wrote: On 2006-05-09 09:34, Ross Vandegrift wrote: Been a bad week for my backups. Today, we had a mostly catastrophic loss of a vtape filesystem, lost 14 of 18 vtapes on our busiest Amanda server. I'm just finishing picking up the pieces now and getting things back on track. [snip] For the future: If the vtapes are really important, make it a RAID system, and/or use RAIT to mirror the Amanda backups to tape + disk or disk + external disk. Heh. The filesystem was a 2.0 TB LUN on an EMC CX300 SAN. RAID, redundant connections, the whole nine yards. Looks like we either hit a bug in the HBA driver or tripped some kind of Linux 2.4.5-esque filesystem damage. It won't even fsck. Note that the host and disk arguments are expressions and that you may specify many of them. Somehow, after you stare at man pages long enough, you start to think you know what's there. For me, at least, I only *think* I know. I've often found that asking questions on a mailing list is a great way to set this straight! Thanks for the tip Paul. Funny how it was staring me in the face this whole time! -- Ross Vandegrift [EMAIL PROTECTED] The good Christian should beware of mathematicians, and all those who make empty prophecies. The danger already exists that the mathematicians have made a covenant with the devil to darken the spirit and to confine man in the bonds of Hell. --St. Augustine, De Genesi ad Litteram, Book II, xviii, 37
Re: Virutal server host
On Sun, May 07, 2006 at 12:07:10PM -0400, Jon LaBadie wrote: If all you are doing is adding a additional name, like amanda_server to an existing hostname/ip_address pairing I can see name confusion easily arising. But would that be true if you ran a second NIC in the server? It should respond with the hostname associated with that IP. The most important thing, as far as Amanda is concerned, is that forward and reverse lookups are sensibly matched. So your hostname should lookup to an address that contains a valid hostname for the client/server. All of our servers provide backups for two physically disparate networks. We manage this currently by creating entries in hosts files for servers that need backups (uggg...), though I want to move it all to DNS. If you ever get confused about what entries should exist, the easiest way to figure it out is to run amcheck and see what it complains about! Usually you'll get something descriptive like 10.0.0.2: No such host, and then you'll know you need to configure a name. No second NIC, then what about additional IP addresses on the same NIC? Those should be associated with their own hostnames also. Solaris, linux, and other unicies I'm familiar with all support multiple IP addresses on a single NIC. If you're concerned about portability, this is probably the easiest way. Bind a second IP to your server, setup name lookups for that second IP, and then float it around to whatever box you need to be migrating to at the time. -- Ross Vandegrift [EMAIL PROTECTED] The good Christian should beware of mathematicians, and all those who make empty prophecies. The danger already exists that the mathematicians have made a covenant with the devil to darken the spirit and to confine man in the bonds of Hell. --St. Augustine, De Genesi ad Litteram, Book II, xviii, 37
Heads up: RHELv4 U3 dump format issues
Hello everyone, Just recovering from a panic attack, where I was unable to perform a restore from dumps done on an RHEL v4 client. Since others may see this as well, I thought I'd drop a heads up. You'll know you've seen this when amrecover complains that root is missing from tape. If you save the dump file from amrecover and look at it with restore, you will get nothing useful. You won't be able to rebuild QFA data, you won't be able to get things off the tape. If you have amrestore look through the tape (something like: amrestore -f 0 -p file:/var/backups/tape01 doesnotexist /dev/null), you'll notice that the tape listing stops before it actually lists everything there (easy for me to see on vtapes, heh). What happened: Some new version of dump must've gone into RHEL v4U3 which produces dumps that make older versions of restore very unhappy. I was able to solve the issue by upgrading the server and the client to the newest dump/restore. In particular, Knoppix v4.0.2 (which I use for baremetal restores) is affected by this issue. I'm not sure if this is a RedHat specific bug, or if it affects all newer versions of dump for ext2/3. -- Ross Vandegrift [EMAIL PROTECTED] The good Christian should beware of mathematicians, and all those who make empty prophecies. The danger already exists that the mathematicians have made a covenant with the devil to darken the spirit and to confine man in the bonds of Hell. --St. Augustine, De Genesi ad Litteram, Book II, xviii, 37
Re: how to tell amanda, that it can use the full bandwith? - some general questions
On Sat, May 06, 2006 at 11:42:57PM +0200, Hans-Christian Armingeon wrote: Look into aspects other than netusage for performance problems. What could be a good place to look at? I have no link problems. Check the output of amstatus when the run is completed. AT the bottom, it includes a breakdown of what Amanda was waiting on when it was waiting. If you're seeing slow transfers of individual dumps, the problem is more than likely related to: 1) Too many dumps at once are fighting for bandwidth 2) The client in question is too loaded/slow 3) The server is too slow In general, I've seen dumping take longer than transferring. -- Ross Vandegrift [EMAIL PROTECTED] The good Christian should beware of mathematicians, and all those who make empty prophecies. The danger already exists that the mathematicians have made a covenant with the devil to darken the spirit and to confine man in the bonds of Hell. --St. Augustine, De Genesi ad Litteram, Book II, xviii, 37
Re: starttime question
On Wed, May 03, 2006 at 11:34:38AM -0400, Jon LaBadie wrote: Are the estimates for delayed DLEs also delayed? Or are they done at the beginning with all other DLEs? I just took a look at the amdump logs for a server with some delayed DLEs. Estimates are performed along with everything else, and an initial schedule is generated. Interestingly, I noticed from reading the logs that estimates take long enough on this server that it just exceeds the starttime. So in my case, it looks likes the dumps aren't actually delayed (heh, in fact, the delayed dump is third in the schedule of ~200 DLEs). -- Ross Vandegrift [EMAIL PROTECTED] The good Christian should beware of mathematicians, and all those who make empty prophecies. The danger already exists that the mathematicians have made a covenant with the devil to darken the spirit and to confine man in the bonds of Hell. --St. Augustine, De Genesi ad Litteram, Book II, xviii, 37
DLEs with large numbers of files
Hello everyone, I recognize that this isn't really related to Amanda, but I thought I'd see if anyone has a good trick... A number of DLEs in my Amanda configuration have a huge number of small files (sometimes hardlinks and symlinks, sometimes just copies) - often times in the millions. Of course this is a class corner case and these DLEs can take a very long time to backup/restore. Currently, they are mostly using dump (which will usually report 1-3MiB/s throughput). Is there a possible performance advantage to using tar instead? On some of our installations I have bumped up the data timeouts. I've got one as high as 5400 seconds. I suspect a reasonable maximum is very installation dependant, but if anyone has thoughts, I'd love to hear them. Thanks for any ideas! -- Ross Vandegrift [EMAIL PROTECTED] The good Christian should beware of mathematicians, and all those who make empty prophecies. The danger already exists that the mathematicians have made a covenant with the devil to darken the spirit and to confine man in the bonds of Hell. --St. Augustine, De Genesi ad Litteram, Book II, xviii, 37
Re: amcheck error : can not execute ...
On Tue, Apr 18, 2006 at 06:39:46PM +0200, Thomas Ginestet wrote: The problem is that I ran apt from this client in order to install amanda. Do you think I can reconfigure Amanda ? (dpkg-reconfigure with special options ?) All you should have to do is apt-get install dump. Debian doesn't install dump by default. Ross
Re: Defining a sensible backup routine
On Mon, Apr 17, 2006 at 12:15:07PM +0100, Anne Wilson wrote: The DLEs are set to a 4-day rotation, but in practice I am seeing level 0 backups almost every other day, for all of them. It would seem reasonable, to me, based on the small amount of changes, to have a level0 weekly, for all of them, I think. Hi Anne, If you want a level0 weekly why not set your dumpcycle to 7 days? This basically tells Amanda to try and do weekly level0 backups and incrementals are okay in between. Another thing you can do that helped me out a lot - change the bump parameters! I work with a fairly large Amanda setup where disk space is usually at a premium. I was able to save a lot of space by tuning the bump parameters to do more higher-level incremental backups. While the system previously wasted a lot of diskspace in the second half of the cycle doing level 1 and 2 backups, now it usually bumps every day, hitting level 5 or 6 by the end of the week. Of course you take a hit on restore time, but restores are pretty decent from virtual tape as it is, so I'm not too concerned about there being five or six steps instead of two or three. -- Ross Vandegrift [EMAIL PROTECTED] The good Christian should beware of mathematicians, and all those who make empty prophecies. The danger already exists that the mathematicians have made a covenant with the devil to darken the spirit and to confine man in the bonds of Hell. --St. Augustine, De Genesi ad Litteram, Book II, xviii, 37
Changing vtape parameters in a live system?
Hello everyone, Is it safe to change the length of a vtape definition on a live system? One of our system was originally defined with 100GiB vtapes, and I'd like to bump that size up to 200+GiB. I guess if the answer is no I can always add a new tapetype and manually rotate the old vtapes out, but I'm wondering if I can just extend the length to make my life easier. Thanks! -- Ross Vandegrift [EMAIL PROTECTED] The good Christian should beware of mathematicians, and all those who make empty prophecies. The danger already exists that the mathematicians have made a covenant with the devil to darken the spirit and to confine man in the bonds of Hell. --St. Augustine, De Genesi ad Litteram, Book II, xviii, 37
Re: asking for real-life examples of Amanda implementations for upcoming book
Hi there, missed the original post, hope this is interesting: --- Name of your organization (optional, but highly desirable). If you can't tell the name of your employer, maybe you can tell what kind of organization that is, e.g. university, manufacturing company, financial services, government, etc I'm not going to disclose my employer without permission, but we are in the managed hosting industry. Any interesting Amanda stories you can tell. For how long have you been using Amanda? (I especially look for examples where people used Amanda for more than a year in production) Amanda has been in use here for number of years. It was well entrenched over a year ago when I arrived. What is your current version of Amanda? Depends on the machine. Debian woody is currently on 2.4.2p2, RHEL is shipping 2.4.5 with RHEL4 (2.4.4 with RHEL3?) We mostly keep to vendor distributed versions on clients, though I am exploring the feasability of upgrading Debian woody machines. On the servers, I build the software from the latest patch release. We're running the latest 2.4.5, but looking to upgrade to 2.5.0 soon. What OS do you use for Amanda server? RHEL 3 and 4. How many Amanda servers do you have? We have four servers. One uses a fibre-channel attached EMC SAN, the other three use fibre-channel attached Apple XRaids. All backups are done to virtual tapes on disk. How many Amanda clients per server do you have? (I especially look for examples with 10+ clients) It varies somewhat; the server on the EMC box manages dumps for 198 disks on about 75 hosts. The Apple Xraids can handle around 100 disks on 50 hosts. What operating systems do you backup? RHEL, Debian, Suse, FreeBSD. What is total amount of data you backup? Around 5TB, give or take some. Do you use dump or tar and why? Both, depending on host. Whenever possible we use dump as it's almost always better at making filesystem images that can be restored under any circumstances. However, we have seen a number of host/filesystem combinations that break dump. Fortunately, on all of these, tar is a happy camper! How often do you do full backups? Our dumpcycle is 7 days. I am very interested in examples of tape-free implementations of Amanda, where backups are done only to disk. Are there any specific details you are interested in? We have FC attached storage for backup/holding space. We create two weeks worth of virtual tapes and use chg-multi for the tape changer. I considered moving to chg-disk after reading the latest documentation, but I seem to recall that restores became more tedious and so I stuck with chg-multi, which works great. How large is your holding disk? We use 150GB. How often do you restore files? (on average and what are thy typical scenarios) Rarely does a week go by and very seldom a month, without a restore. Do you use compression and where (hardware, client or server)? We use fast client compression to get more hosts done in less time. Do you use encryption and which one? Not currently, but the upgrade to 2.5.0 is motivated exclusively by encryption support. How do you backup Windows? (if you have any) Not Amanda. UltraBac is most common setup. Some clients are using their own setup. The inability to backup/restore security labels rules smbclient backups out. -- Ross Vandegrift [EMAIL PROTECTED] The good Christian should beware of mathematicians, and all those who make empty prophecies. The danger already exists that the mathematicians have made a covenant with the devil to darken the spirit and to confine man in the bonds of Hell. --St. Augustine, De Genesi ad Litteram, Book II, xviii, 37
amoverview error codes
Hello everyone, I'm a bit foggy on one case of error codes in the amoverview output. The documentation says that E indicates an error, and that E followed by a number indicated an error flushing that was later corrected. Cool. I sometimes see errors that are of the form 1E and I can't figure out what this would mean. A level 1 dump was flushed, and then the second time there was an error? Doesn't make much sense to flush twice, and I don't see this case in the documentation. Any hints? -- Ross Vandegrift [EMAIL PROTECTED] The good Christian should beware of mathematicians, and all those who make empty prophecies. The danger already exists that the mathematicians have made a covenant with the devil to darken the spirit and to confine man in the bonds of Hell. --St. Augustine, De Genesi ad Litteram, Book II, xviii, 37
Re: amrecover doesn't prompt for tape change?
On Mon, Feb 20, 2006 at 12:42:09PM +0100, Paul Bijnens wrote: In that case, you have already set the tape device, and you just have to physically insert the tape in that device before pressing Y. Or what is exactly the problem here? We use virtual tapes on a RAID array - not really any way I can physically change the tapes. I figured I'd probably need to upgrade the client tools on those boxes - they're running Debian woody, so, yea, prehistoric! Ross
amrecover doesn't prompt for tape change?
Hello everyone, When I do a local restore on a backup server with amrecover, I get prompted to change tapes when it is required: Extracting files using tape drive /dev/null on host linuxbackup1. Load tape DailySet102 now Continue [?/Y/n/t]? I can then press t and set the tape correctly. However, when I do a remote restore with amrecover, I get a different prompt: Load tape DailySet113 now Continue? [Y/n]: This is fine for a restore that requires a single tape, since I can use settape and the extract a second time. However, restores requiring multiple tapes will always fails, since I can't change tapes in the middle of the restore. I've verified that the remote server is correctly configured in .amandahosts for access. The only difference I can see between the two installation is that the backup server runs 2.4.5, while the remote server runs 2.4.2p2. Was this feature added in between these versions? -- Ross Vandegrift [EMAIL PROTECTED] The good Christian should beware of mathematicians, and all those who make empty prophecies. The danger already exists that the mathematicians have made a covenant with the devil to darken the spirit and to confine man in the bonds of Hell. --St. Augustine, De Genesi ad Litteram, Book II, xviii, 37