Re: amrecover with file: driver
amrecover does not change the tape by itself. You have to keep in mind that the chg-disk emulates a robotic with the hard drive. And amrecover (certainly due to security reasons that I won't discuss here) does not automatically search the correct tape in the robot: the tape(s) you want to use when restoring have to be in the tape device when needed. In my case I have 2 xterms, one with amrecover, and the other one where I run amtape to load the right tape when amrecover tells me to do it. On 22 Jan 2004 11:54:11 -0700 Marc Langlois [EMAIL PROTECTED] wrote: Stefan, The settape section of the man page for amrecover says: settape [[server]:][tapedev|default] If you want amrecover to use your changer, the tapedev must be equal to the amrecover_changer setting on the server. I used this setting in amanda.conf: amrecover_changer chg-disk then, in amrecover: settape chg-disk and the extract invokes the chg-disk script and sets the ./data sym-link to the correct slot automatically. Very nice! Marc. It appears that using settape chg-disk in amrecover On Thu, 2004-01-22 at 09:17, Stefan G. Weichinger wrote: Hi, Marc, on Donnerstag, 22. Jänner 2004 at 16:59 you wrote to amanda-users: ML I'm using the chg-disk changer. Does amrecover use the changer script? Good question. There is the parameter amrecover_changer which explicitly sets the changer used. Maybe you try to set this in your conf. ML And would it be possible to modify the script to choose the correct ML slot? As soon as I know it I will tell you ;) Practically YES, it is possible.
RE: Soalris client backup problem resolved
Get thee to alpha.gnu.org with ftp and download 1.13.25. There will be V. recent versions like 1.13.92, but for now ignore them. Get and build .25. When I put it on the system, I name it amgtar to avoid conflict with other tars. I.e. THE gtar intended for amanda. And I use the --with-gnutar ... to point to the proper version. Hi Jon, Just to let you know I followed you instructions last week upgraded gcc to 3.3.2 before compiling gnutar 1.13.25 which I named as amgtar on the Solaris client. Then recompiled the amanda client on Solaris with the following parameters shannon#./configure --with-user=amanda --with-group=amanda --with-libraries=/usr/local/lib --without-server --with-amandahosts --with-gnutar=/usr/local/bin/amgtar --with-index-server=orion.backup.ie --with-tape-server=orion.heanet.ie --with debugging=/var/log/amdebug pointing to the new gnutar version and added --with-amandahosts and sorted out permissions on debugging=/var/log/amdebug. Then I was able to backup all the Solaris partitions and tested a restore which was successful, with no 'Archive contains obsolescent base-64 headers errors' as previously for the Solaris client partitions. I restored with the following command using gnutar on the backup server. orion#/usr/local/amanda/sbin/amrestore -p /dev/nst1 wos-int /ext/isidata432 | tar xvfp - I just thought I let the list know, thanks for all your help. I'm sure I may have some more questions in the future. regards, Stephen smime.p7s Description: S/MIME cryptographic signature
Re: Soalris client backup problem resolved
On Mon, Jan 26, 2004 at 10:22:55AM -, Stephen Yeates wrote: Just to let you know I followed you instructions last week ... Then I was able to backup all the Solaris partitions and tested a restore which was successful, with no 'Archive contains obsolescent base-64 headers errors' as previously for the Solaris client partitions. ... I just thought I let the list know, thanks for all your help. I'm sure I may have some more questions in the future. Successful followup reports always appreciated! That buys you one more answer :)) -- Jon H. LaBadie [EMAIL PROTECTED] JG Computing 4455 Province Line Road(609) 252-0159 Princeton, NJ 08540-4322 (609) 683-7220 (fax)
Re: multiple disklists
Hi, Robert, on Freitag, 23. Jänner 2004 at 19:19 you wrote to amanda-users: DRC I'm already using 2 different dumpcycles. My global config is to cycle 9 DRC days. On the lower priority data, I have the dump cycle increased to 24 DRC days. DRC Isn't there a way to share multiple disklists with a single config? You could copy them in and out via scripts but this would confuse AMANDA and your balance would get way out of order. Personally I would not recommend to do that as it would work against the scheduling-algorithms. Why not use two configs in parallel? -- best regards, Stefan Stefan G. Weichinger mailto:[EMAIL PROTECTED]
Where is the archive?
I am new to this list and wanted to peruse the archives. The welcome message refers to ftp://ftp.amanda.org/pub/amanda/maillist-archives, but maillist-archives seems to be missing. What is wrong? Regards, Erik P. Olsen
Re: Where is the archive?
Hi, Erik, on Montag, 26. Jänner 2004 at 16:48 you wrote to amanda-users: EPO I am new to this list and wanted to peruse the archives. The welcome message EPO refers to ftp://ftp.amanda.org/pub/amanda/maillist-archives, but EPO maillist-archives seems to be missing. Use this link: http://marc.theaimsgroup.com/?l=amanda-usersr=1w=2 -- best regards, Stefan Stefan G. Weichinger mailto:[EMAIL PROTECTED]
Re: Where is the archive?
--On Monday, January 26, 2004 16:48:28 +0100 Erik P. Olsen [EMAIL PROTECTED] wrote: I am new to this list and wanted to peruse the archives. The welcome message refers to ftp://ftp.amanda.org/pub/amanda/maillist-archives, but maillist-archives seems to be missing. What is wrong? Regards, Erik P. Olsen I don't know about the welcome message, but www.amanda.org shows the list archive to be at http://groups.yahoo.com/group/amanda-users/ which does work. Frank -- Frank Smith [EMAIL PROTECTED] Systems Administrator Voice: 512-374-4673 Hoover's Online Fax: 512-374-4501
Re: 3ware 8506-4 raid1 hotswap
Christopher Pulling one out is ok. We had some problems when Christopher inserting a disk that was used as half a mirror Christopher before. This actually killed the machine. The disk had to Christopher be cleaned before reinserting. We've also encountered this problem. We've never had a problem once we clean the first handfull of blocks on the disk. We've never done any testing to see exactly how much needs to be cleaned. What we do is mount the disk in another box, run dd if=/dev/zero of=/dev/DEV count=1024 and then the controller always sees it as a new disk.
Re: amrecover with file: driver
Hi That s a bit strange. my amrecover can change tape and find the good. you have to put the following parameter in amanda.conf amrecover_changer /dev/null and to launch amrecover in that way amrecover -d /dev/null then the tape server will use the changer specified in the amanda.conf file to find the good tape. hope that helps Le lun 26/01/2004 09:42, JC Simonetti a crit : amrecover does not change the tape by itself. You have to keep in mind that the chg-disk emulates a robotic with the hard drive. And amrecover (certainly due to security reasons that I won't discuss here) does not automatically search the correct tape in the robot: the tape(s) you want to use when restoring have to be in the tape device when needed. In my case I have 2 xterms, one with amrecover, and the other one where I run amtape to load the right tape when amrecover tells me to do it. On 22 Jan 2004 11:54:11 -0700 Marc Langlois [EMAIL PROTECTED] wrote: Stefan, The settape section of the man page for amrecover says: settape [[server]:][tapedev|default] If you want amrecover to use your changer, the tapedev must be equal to the amrecover_changer setting on the server. I used this setting in amanda.conf: amrecover_changer chg-disk then, in amrecover: settape chg-disk and the extract invokes the chg-disk script and sets the ./data sym-link to the correct slot automatically. Very nice! Marc. It appears that using settape chg-disk in amrecover On Thu, 2004-01-22 at 09:17, Stefan G. Weichinger wrote: Hi, Marc, on Donnerstag, 22. Jnner 2004 at 16:59 you wrote to amanda-users: ML I'm using the chg-disk changer. Does amrecover use the changer script? Good question. There is the parameter amrecover_changer which explicitly sets the changer used. Maybe you try to set this in your conf. ML And would it be possible to modify the script to choose the correct ML slot? As soon as I know it I will tell you ;) Practically YES, it is possible.
Re: amrecover with file: driver
Hi, Eric, on Montag, 26. Jänner 2004 at 21:11 you wrote to amanda-users: ED Hi ED That s a bit strange. ED my amrecover can change tape and find the good. ED you have to put the following parameter in amanda.conf ED amrecover_changer /dev/null ED and to launch amrecover in that way ED amrecover -d /dev/null ED then the tape server will use the changer specified in the amanda.conf ED file to find the good tape. ED hope that helps Thank you, Eric. My quick greps through the sources did not find that this SHOULD work like this, but maybe someone tells us if this is intended. Marc, could you test Eric's solution in your installation and report your results? -- best regards, Stefan Stefan G. Weichinger mailto:[EMAIL PROTECTED]
Encryption
Amanda Users, I looked though the archives but didn't see anything about encryption and tape, I've seen articles on encrypting the network traffic but that isn't the task we're looking at. We have some potentially sensitive data and this seems like a possible way to make sure that it stays private - vs of course the standard mission of making sure we can get it back incase something happends to the on-disk bits and bytes. Are their hooks for this ? Anyone done it yet ? thanks, Brian --- Brian R Cuttler [EMAIL PROTECTED] Computer Systems Support(v) 518 486-1697 Wadsworth Center(f) 518 473-6384 NYS Department of HealthHelp Desk 518 473-0773
Re: amrecover with file: driver
Stefan G. Weichinger wrote: Hi, Eric, on Montag, 26. Jänner 2004 at 21:11 you wrote to amanda-users: ED That s a bit strange. ED my amrecover can change tape and find the good. ED you have to put the following parameter in amanda.conf ED amrecover_changer /dev/null ED and to launch amrecover in that way ED amrecover -d /dev/null ED then the tape server will use the changer specified in the amanda.conf ED file to find the good tape. ED hope that helps Thank you, Eric. My quick greps through the sources did not find that this SHOULD work like this, but maybe someone tells us if this is intended. Marc, could you test Eric's solution in your installation and report your results? The way it is intended, and the way it works too, is that you specify a string in amanda.conf with the directive amrecover_changer thestring and you instruct amrecover to use the changer with the command settape thestring It is just a way to name the changer. As name thestring you can use anything. If you have set up in amanda.conf: amrecover_changer foobar and then instruct in amrecover settape foobar, then amanda will use the changer. To avoid confusion, it is best to avoid names that seem to indicate something else, like /dev/null, or nocompression, or Legato. Just like: mkdir file; touch file/directory -- Paul @ Home
Re: amrecover with file: driver
Hi, Paul, on Montag, 26. Jänner 2004 at 22:54 you wrote to amanda-users: PB The way it is intended, and the way it works too, is that you PB specify a string in amanda.conf with the directive PB amrecover_changer thestring PB and you instruct amrecover to use the changer with the command PB settape thestring PB It is just a way to name the changer. PB As name thestring you can use anything. PB If you have set up in amanda.conf: amrecover_changer foobar PB and then instruct in amrecover settape foobar, then amanda PB will use the changer. PB To avoid confusion, it is best to avoid names that seem to PB indicate something else, like /dev/null, or nocompression, PB or Legato. PB Just like: mkdir file; touch file/directory Thank you for clarifying this, Paul. I crossed this foobar-mechanism earlier and did not make the connection. Could you point us to the spot where this is explained in the docs? I suppose there is some reason to make this function with ANY string. -- best regards, Stefan Stefan G. Weichinger mailto:[EMAIL PROTECTED]
Re: amrecover with file: driver
Stefan G. Weichinger wrote: Could you point us to the spot where this is explained in the docs? I suppose there is some reason to make this function with ANY string. From man amanda: amrecover_changer string Default: . Amrecover will use the changer if you use settape STRING and that string is the same as the amrecover_changer setting. From man amrecover: settape [[server]:][tapedev|default] [...] If you want amrecover to use your changer, the tapedev must be equal to the amrecover_changer setting on the server. The string is just a name for your changer. Something different than /dev/rmt0. A changer can also be of type chg-multi, grouping a few physical drives together, or chg-disk without any tapedrive involved. With this string you name the group. And you can get silly and name your newborn son Sue and your changer device /dev/null. -- Paul @ Home
Re: amrecover with file: driver
Hi, Paul Bijnens, on Montag, 26. Jänner 2004 at 23:32 you wrote to amanda-users: PB From man amanda: PB From man amrecover: Yeah. Silly me did not get that right. PB And you can get silly and name your newborn son Sue and your PB changer device /dev/null. I will think about that ;) -- best regards, Stefan Stefan G. Weichinger mailto:[EMAIL PROTECTED]
Recover from multiple dumps per day
We would like to be able to recover accidentally deleted files from amanda backups that are made more than once per day, so I'm running amdump from cron every 2 hours during working hours. This seems to work fine with the file driver, and with 20 tape slots a full backup is done at least once every 2 days. The problem I'm having is that when I retrieve a file or directory, amrecover doesn't seem to be able to access the most recent backup. I've tried using the setdate command to both today and tomorrow, but it only displays information about the last backup from yesterday. All the man pages I've seen only list dates to the nearest day, but is it possible to use a finer time resolution? Thanks, Marc Langlois.
Re: Recover from multiple dumps per day
Hi, Marc, on Dienstag, 27. Jänner 2004 at 00:41 you wrote to amanda-users: ML The problem I'm having is that when I retrieve a file or directory, ML amrecover doesn't seem to be able to access the most recent backup. I've ML tried using the setdate command to both today and tomorrow, but it ML only displays information about the last backup from yesterday. All ML the man pages I've seen only list dates to the nearest day, but is it ML possible to use a finer time resolution? AFAIK AMANDA is designed to amdump once per day maximum. Everything else confuses scheduling. -- I provide a similar functionality to the users by doing snapshots, done by a combination of find and tar ... The script tars the files that have been modified in the last X (3) hours to a stamped and zipped archive that can be accessed by the users via standard-packers. This script is run by cron every Y (2) hours. Snapshots are backed up by AMANDA and deleted after a week or so. So the users can access their changed files without having to handle or even understand things like amrecover. -- best regards, Stefan Stefan G. Weichinger mailto:[EMAIL PROTECTED]
Re: Recover from multiple dumps per day
On Mon, Jan 26, 2004 at 04:41:25PM -0700, Marc Langlois wrote: We would like to be able to recover accidentally deleted files from amanda backups that are made more than once per day, so I'm running amdump from cron every 2 hours during working hours. This seems to work fine with the file driver, and with 20 tape slots a full backup is done at least once every 2 days. The problem I'm having is that when I retrieve a file or directory, amrecover doesn't seem to be able to access the most recent backup. I've tried using the setdate command to both today and tomorrow, but it only displays information about the last backup from yesterday. All the man pages I've seen only list dates to the nearest day, but is it possible to use a finer time resolution? I think that could be considered a flaw in design. Amanda was certainly designed around a model of one run a day, not multiple runs. Nothing prevents multiple runs, but I think you are seeing the result of that design model. Maybe someone has a workaround. Though I've not heard of one. -- Jon H. LaBadie [EMAIL PROTECTED] JG Computing 4455 Province Line Road(609) 252-0159 Princeton, NJ 08540-4322 (609) 683-7220 (fax)