Re: Configuration help?
On Wed, 14 Aug 2002, Joshua Baker-LePain wrote: *snip* 1000 x thanks for the help! I have Amanda up and running and backing up our three servers just the way we wanted it now. Now for my last question for this time: what are the pros and cons, if any, for using tar or (ufs)dump? Are there any reasons or situtations for choosing one over the other? /Conny
Can I use amanda without tape backup?
Dear All, I have2 machines running Red Hat 7.3. I want to back up content of these machines intoan othermachines periodically. I plan to use amanda but I don't have tape drive. Can I use amanda to backup into a hard disk? If you can't use amanda, would you please suggest me about backup software. Thanks, Theewara
Re: Can I use amanda without tape backup?
On Friday 16 August 2002 15:41, Theewara Vorakosit wrote: Dear All, I have 2 machines running Red Hat 7.3. I want to back up content of these machines into an other machines periodically. I plan to use amanda but I don't have tape drive. Can I use amanda to backup into a hard disk? If you can't use amanda, would you please suggest me about backup software. Thanks, Theewara From the archived mailing-list, I think the answer is yes you can. However, I've been trying to do that and still can't. Look in archive for the steps, because I can't find any other docs on doing tapeless backup. Good luck!
Re: Sun L20 and amanda
Anne, I have a Sun Storedge L40 (two L20s stacked together). Here are the lines I had to add to the configuration file for the sgen driver (/kernel/drv/sgen.conf): === device-type-config-list=library; inquiry-config-list=HP, C7200-8000; === The device was originally created under /dev/rmt/scsi.. something, but I linked it to /dev/changer to allow mtx to be used with the device designation. The device on my system is located at: /devices/pci@8,70/scsi@5,1/sgen@0,0:changer I am running Solaris 8 and DLT8000 Wayne Richards Hello, I have configured amanda 2.4.3b3 to use the DLT8000 in the library. The next step is to get the changer working. I wasn't able to get mtx working with the L20. If anyone if currently using the L20 library with amanda, or if you have suggestions on how to configure the tape library changer, I'd really appreciate it. Anne Hammond University of Colorado at Boulder
RE: Request for explanation of Index tee error
It does not happen every night, only five or six nights in the past month. I think this rules out file permissions, and wrong paths. I don't think it's full filesystem either cuz my df shows plenty of space. Michael Martinez System Administrator (Contractor) Information Systems and Technology Management CSREES - United States Department of Agriculture (202) 720-6223 -Original Message- From: Jon LaBadie [mailto:[EMAIL PROTECTED]] Sent: Thursday, August 15, 2002 2:19 PM To: [EMAIL PROTECTED] Subject: Re: Request for explanation of Index tee error On Thu, Aug 15, 2002 at 01:54:56PM -0400, Martinez, Michael - CSREES/ISTM wrote: What does the following syslog errors mean? Should I be concerned about it/do something to fix it? Aug 13 00:56:10 ivideo sendbackup[3608]: index tee cannot write [Connection reset by peer] Aug 14 01:00:43 vision sendbackup[1630]: index tee cannot write [Connection reset by peer] Aug 15 00:55:16 vision sendbackup[3042]: index tee cannot write [Connection reset by peer] Don't know why it is happening, but here is what I think is happening. The data from your dumper program, dump/ufsdump/tar/???, passes through a program that duplicates the entire data set and sends each through a separate data stream (pipes/sockets/...). Similar to the unix tee program. One stream goes to the holding disk/tape drive. The duplicate stream goes to the undumper program to generate a table of contents. This TOC is massaged and becomes the index for the dump. For some reason this second stream is unable to write its output (or temporary files). This could be a permission problem, a wrong path with missing directories, a full file system or ??? HTH -- Jon H. LaBadie [EMAIL PROTECTED] JG Computing 4455 Province Line Road(609) 252-0159 Princeton, NJ 08540-4322 (609) 683-7220 (fax)
amrecover
Hi! We have an Amanda 2.4.0 Server running and want to recover some files. I start amrecover -C TwiceAWeek on the Server, I get a timeout. In the daemon.log : Aug 16 13:47:34 calvin amindexd[31263]: warning: can't get client address: Connection timed out Aug 16 13:47:34 calvin amindexd[31263]: connect from unknown Aug 16 13:49:16 calvin amindexd[31264]: warning: can't get client address: Connection timed out Aug 16 13:49:16 calvin amindexd[31264]: connect from unknown In the .amandahosts I set the localhost to root. What can I do? Malte
Re: Configuration help?
On Fri, 16 Aug 2002 at 10:07am, Conny Gyllendahl wrote 1000 x thanks for the help! I have Amanda up and running and backing up our three servers just the way we wanted it now. Glad to hear it. Now for my last question for this time: what are the pros and cons, if any, for using tar or (ufs)dump? Are there any reasons or situtations for choosing one over the other? Tread lightly -- thar be dragons here. :) They both have their pros and cons, and it can be a matter of deep seated religous belief for people. FS specific dump programs can back up things that tar doesn't know about (e.g. ACLs), and can sometimes be faster. But they're limited to partitions only, and require that the recovery machine have them installed. Tar can do subdirectories and doesn't care about OS/FS (and thus you can recover on just about any machine). It's all a matter of choice. -- Joshua Baker-LePain Department of Biomedical Engineering Duke University
Re: Labels and Barcodes
Then how do you set the barcode? Is it a physical label? If my libary reads bar codes, will it also print them? On Thu, 2002-08-15 at 16:25, Stephen Carville wrote: On 15 Aug 2002, Jason Greenberg wrote: - What's the difference between labels and barcodes? Why, when I label - tapes, does the VolumeTag= not show up for that tape? To amanda, a label is the information in the fist part of a tape. It identifies what backup set the tape belongs to, when it was last used, etc. Labels have nothing to do with the 'Volume Tag which are from the barcodes. -- -- Stephen Carville UNIX and Network Administrator DPSI (formerly Ace USA Flood Services) 310-342-3602 [EMAIL PROTECTED] -- Jason Greenberg, CCNP Network Administrator Execulink, Inc. [EMAIL PROTECTED]
Re: Configuration help?
On Fri, 16 Aug 2002, Joshua Baker-LePain wrote: Tread lightly -- thar be dragons here. :) They both have their pros and cons, and it can be a matter of deep seated religous belief for people. FS specific dump programs can back up things that tar doesn't know about (e.g. ACLs), and can sometimes be faster. But they're limited to partitions only, and require that the recovery machine have them installed. Tar can do subdirectories and doesn't care about OS/FS (and thus you can recover on just about any machine). It's all a matter of choice. Basically I agree with Joshua, with one addition. For Linux systems stick with tar. dump on Linux is, well, not know for being robust. -Mitch
Re: Can I use amanda without tape backup?
Hi folks. Yes, there is a way to backup to disk only. By using amanda-2.4.3b3 you can do your backups to a hard drive or some other file-oriented system. The way to do this is to configure your tapedev to be something like file:/var/spool/backup and to make sure you have a subdirectory named data. You label this 'tape' just like you would label a normal tape, using the amlabel command and run your backups as usual. Here are some practical hints for you: 1. Make sure you use 2.4.3b3, otherwise you won't get a file driver. compile and configure as usual (read the manual). 2. I find a good way of doing this is to create subdirectories in the backup directory organized by date, eg. 20020815 or something, and then symlink them to 'data' and label them. This way you can rotate your directories just like tapes. Personally, i run a shell script that does this just before the backups run every night. As the manual suggests, this is also useful if you want to use cd's or dvd's to accomplish the same thing. If you have a SAN or a NAS like i do, its very handy to mount an nfs directory and do the dumps there. You can also use smb (samba), but NFS seems to be faster. 4. Don't forget to do periodic backups *anyway*, esp. of the index directories... never rely on a 'single point of failure'. Good Luck! cheers, john. Stelar wrote: On Friday 16 August 2002 15:41, Theewara Vorakosit wrote: Dear All, I have 2 machines running Red Hat 7.3. I want to back up content of these machines into an other machines periodically. I plan to use amanda but I don't have tape drive. Can I use amanda to backup into a hard disk? If you can't use amanda, would you please suggest me about backup software. Thanks, Theewara From the archived mailing-list, I think the answer is yes you can. However, I've been trying to do that and still can't. Look in archive for the steps, because I can't find any other docs on doing tapeless backup. Good luck!
Re: Can I use amanda without tape backup?
On Fri, 2002-08-16 at 09:41, Theewara Vorakosit wrote: Dear All, I have 2 machines running Red Hat 7.3. I want to back up content of these machines into an other machines periodically. I plan to use amanda but I don't have tape drive. Can I use amanda to backup into a hard disk? If you can't use amanda, would you please suggest me about backup software. Thanks, Theewara I've made it work...on sun solaris ! for this you have to create a folder which will be like the tape device. You have to use amanda-2.4.3b3 b mean BETA and in the amanda.conf file for the entry: tapedev file:/unige/amanda/bckp/ I didn't really look deep in amanda source code...but I feel that the developer guys have made like a wrapper around amanda to make a repertory behave like a tape. so In my case the repertory looks like; atalante:bckp 171 ls -l total 28 lrwxrwxrwx 1 obs_bkp sys 6 Aug 16 14:45 data - sb06_1/ -rw--- 1 obs_bkp sys 11 Aug 16 15:20 info lrwxrwxrwx 1 obs_bkp sys 27 Aug 12 14:41 sb01_1 - /net/obssb1/export/diskB1/1/ lrwxrwxrwx 1 obs_bkp sys 27 Aug 12 14:41 sb01_2 - /net/obssb1/export/diskB1/2/ lrwxrwxrwx 1 obs_bkp sys 27 Aug 12 14:09 sb02_1 - /net/obssb2/export/diskB1/1/ lrwxrwxrwx 1 obs_bkp sys 27 Aug 12 14:42 sb02_2 - /net/obssb2/export/diskB1/2/ lrwxrwxrwx 1 obs_bkp sys 27 Aug 16 00:43 sb03_1 - /net/obssb3/export/diskB1/1/ in this folder the only thing that you really have to create is the data link who has to point a folder on a disk with enough space to keep the backup. The file info is created/used by amanda -- so don't pay attention to it and the other link like sb1_1 sb_1 are linked to HD's spread around the LAN 'cause at each run we backup about 40 Gb. an that's it
Re: Can I use amanda without tape backup?
Thanks for you suggest. Can I use amanda 2.4.2p2-7 that equip with Red Hat 7.3 Thanks, Theewara - Original Message - From: BRINER Cedric [EMAIL PROTECTED] To: Theewara Vorakosit [EMAIL PROTECTED] Cc: _amanda [EMAIL PROTECTED] Sent: Friday, August 16, 2002 8:38 PM Subject: Re: Can I use amanda without tape backup? On Fri, 2002-08-16 at 09:41, Theewara Vorakosit wrote: Dear All, I have 2 machines running Red Hat 7.3. I want to back up content of these machines into an other machines periodically. I plan to use amanda but I don't have tape drive. Can I use amanda to backup into a hard disk? If you can't use amanda, would you please suggest me about backup software. Thanks, Theewara I've made it work...on sun solaris ! for this you have to create a folder which will be like the tape device. You have to use amanda-2.4.3b3 b mean BETA and in the amanda.conf file for the entry: tapedev file:/unige/amanda/bckp/ I didn't really look deep in amanda source code...but I feel that the developer guys have made like a wrapper around amanda to make a repertory behave like a tape. so In my case the repertory looks like; atalante:bckp 171 ls -l total 28 lrwxrwxrwx 1 obs_bkp sys 6 Aug 16 14:45 data - sb06_1/ -rw--- 1 obs_bkp sys 11 Aug 16 15:20 info lrwxrwxrwx 1 obs_bkp sys 27 Aug 12 14:41 sb01_1 - /net/obssb1/export/diskB1/1/ lrwxrwxrwx 1 obs_bkp sys 27 Aug 12 14:41 sb01_2 - /net/obssb1/export/diskB1/2/ lrwxrwxrwx 1 obs_bkp sys 27 Aug 12 14:09 sb02_1 - /net/obssb2/export/diskB1/1/ lrwxrwxrwx 1 obs_bkp sys 27 Aug 12 14:42 sb02_2 - /net/obssb2/export/diskB1/2/ lrwxrwxrwx 1 obs_bkp sys 27 Aug 16 00:43 sb03_1 - /net/obssb3/export/diskB1/1/ in this folder the only thing that you really have to create is the data link who has to point a folder on a disk with enough space to keep the backup. The file info is created/used by amanda -- so don't pay attention to it and the other link like sb1_1 sb_1 are linked to HD's spread around the LAN 'cause at each run we backup about 40 Gb. an that's it
/dev/nst0 : Permission denied
Hi, Hopefully this may be the last stupid question before I get amanda up and running. When I run amlabel I get /dev/nst0 : Permission denied What have I missed. Thanks Kevin
amflush with tapeless configuration
hi, I get problem with amflush with my new tapeless configuration. does someone get such kind of problem also ?? I launch the amflush succesfully right after it i look for some process of amanda ps -ef | grep am don't give me any amexecfile I don't see any new files in the disktape pointed by data and the only log file that I found relevant is: less /export/diskE/amanda-2.4.3b3/DailySet/log : : DISK amflush obssf2 /export/diskB START amflush date 20020816 START driver date 20020816 WARNING driver WARNING: /export/diskF/DailySetDisk: 69206016 KB requested, but only 15932047 KB available. STATS driver startup time 0.018 START taper datestamp 20020816 label sb06_1 tape 0 INFO taper tape sb06_1 kb 0 fm 0 [OK] FINISH driver date 20020816 time 0.465 thanks in advance briner
Re: Can I use amanda without tape backup?
NOP AFAIK the tapeless amanda is with 2.4.3xxx briner Thanks for you suggest. Can I use amanda 2.4.2p2-7 that equip with Red Hat 7.3 Thanks, Theewara - Original Message - From: BRINER Cedric [EMAIL PROTECTED] To: Theewara Vorakosit [EMAIL PROTECTED] Cc: _amanda [EMAIL PROTECTED] Sent: Friday, August 16, 2002 8:38 PM Subject: Re: Can I use amanda without tape backup? On Fri, 2002-08-16 at 09:41, Theewara Vorakosit wrote: Dear All, I have 2 machines running Red Hat 7.3. I want to back up content of these machines into an other machines periodically. I plan to use amanda but I don't have tape drive. Can I use amanda to backup into a hard disk? If you can't use amanda, would you please suggest me about backup software. Thanks, Theewara I've made it work...on sun solaris ! for this you have to create a folder which will be like the tape device. You have to use amanda-2.4.3b3 b mean BETA and in the amanda.conf file for the entry: tapedev file:/unige/amanda/bckp/ I didn't really look deep in amanda source code...but I feel that the developer guys have made like a wrapper around amanda to make a repertory behave like a tape. so In my case the repertory looks like; atalante:bckp 171 ls -l total 28 lrwxrwxrwx 1 obs_bkp sys 6 Aug 16 14:45 data - sb06_1/ -rw--- 1 obs_bkp sys 11 Aug 16 15:20 info lrwxrwxrwx 1 obs_bkp sys 27 Aug 12 14:41 sb01_1 - /net/obssb1/export/diskB1/1/ lrwxrwxrwx 1 obs_bkp sys 27 Aug 12 14:41 sb01_2 - /net/obssb1/export/diskB1/2/ lrwxrwxrwx 1 obs_bkp sys 27 Aug 12 14:09 sb02_1 - /net/obssb2/export/diskB1/1/ lrwxrwxrwx 1 obs_bkp sys 27 Aug 12 14:42 sb02_2 - /net/obssb2/export/diskB1/2/ lrwxrwxrwx 1 obs_bkp sys 27 Aug 16 00:43 sb03_1 - /net/obssb3/export/diskB1/1/ in this folder the only thing that you really have to create is the data link who has to point a folder on a disk with enough space to keep the backup. The file info is created/used by amanda -- so don't pay attention to it and the other link like sb1_1 sb_1 are linked to HD's spread around the LAN 'cause at each run we backup about 40 Gb. an that's it -- * * * -- **--- * :( BRINER Cédric* o + Observatory of Geneva, Switzerland .:oO0Oo:. cedricîbriner@obsóunigeùch* * o ( ^--- hint:replace the odd letters by dot ):*
Re: amrecover
Hi! We have an Amanda 2.4.0 Server running and want to recover some files. I start amrecover -C TwiceAWeek on the Server, I get a timeout. In the daemon.log : Aug 16 13:47:34 calvin amindexd[31263]: warning: can't get client address: Connection timed out Aug 16 13:47:34 calvin amindexd[31263]: connect from unknown Aug 16 13:49:16 calvin amindexd[31264]: warning: can't get client address: Connection timed out Aug 16 13:49:16 calvin amindexd[31264]: connect from unknown In the .amandahosts I set the localhost to root. What can I do? Malte from where are you starting the amrecover... for example if you launch amrecover from calvin.domain.com the put in your .amandahosts calvin.domain.com root calvin root briner -- * * * -- **--- * :( BRINER Cédric* o + Observatory of Geneva, Switzerland .:oO0Oo:. cedricîbriner@obsóunigeùch* * o ( ^--- hint:replace the odd letters by dot ):*
Re: Configuration help?
On Friday 16 August 2002 03:07, Conny Gyllendahl wrote: On Wed, 14 Aug 2002, Joshua Baker-LePain wrote: *snip* 1000 x thanks for the help! I have Amanda up and running and backing up our three servers just the way we wanted it now. Now for my last question for this time: what are the pros and cons, if any, for using tar or (ufs)dump? Are there any reasons or situtations for choosing one over the other? /Conny Thats an almost religious battle that seems to not have a definitive answer and I'd rather not throw any more starter on the fire. Me, I use tar, if for no other reason than it can accept an exclude list of things not to backup, like the holding directory if its not on a seperate partition thats not in the disklist. Someone else will have an equally valid reason to use dump, like its faster or ??? Read the comments said here, read the manpages and decide which collection plate to put your tithe in. :-) -- Cheers, Gene AMD K6-III@500mhz 320M Athlon1600XP@1400mhz 512M 99.11% setiathome rank, not too shabby for a WV hillbilly
Re: Is it advisable to have the holding disk on a separate partition?
On Friday 16 August 2002 05:17, Kevin Passey wrote: Or can I change the conf file and put it as /usr/dumps/amanda. When I then backup /usr/ I would then exclude or include /dumps in my disklist file. Or have I got it completely wrong Thanks anyway Kevin You can use the exclude file if you're using tar, in which case you can put it where the most free space is. Mine's in /usr/dumps as I had 27 gigs free there a year ago when I set this one up, and is so entered in amanda.conf IIRC. The exclude files entries are in relative format, meaning that the entry line for excluding /usr/dumps looks like './dumps/' without the quotes. The leading ./ is very important. AIUI, dump users will need to setup a seperate partition and leave that filesystem out of the disklist in order to achieve that exclusion. -- Cheers, Gene AMD K6-III@500mhz 320M Athlon1600XP@1400mhz 512M 99.11% setiathome rank, not too shabby for a WV hillbilly
Re: /dev/nst0 : Permission denied
On Friday 16 August 2002 10:06, Kevin Passey wrote: Hi, Hopefully this may be the last stupid question before I get amanda up and running. When I run amlabel I get /dev/nst0 : Permission denied What have I missed. Perms maybe? Here's mine: # ls -l /dev/nst0 crw-rw1 root disk 9, 128 Apr 11 10:25 /dev/nst0 You can chown it if you're root. -- Cheers, Gene AMD K6-III@500mhz 320M Athlon1600XP@1400mhz 512M 99.11% setiathome rank, not too shabby for a WV hillbilly
Re: amflush with tapeless configuration
On Friday 16 August 2002 10:19, BRINER Cedric wrote: hi, I get problem with amflush with my new tapeless configuration. does someone get such kind of problem also ?? I launch the amflush succesfully right after it i look for some process of amanda ps -ef | grep am don't give me any amexecfile I don't see any new files in the disktape pointed by data and the only log file that I found relevant is: less /export/diskE/amanda-2.4.3b3/DailySet/log DISK amflush obssf2 /export/diskB START amflush date 20020816 START driver date 20020816 WARNING driver WARNING: /export/diskF/DailySetDisk: 69206016 KB requested, but only 15932047 KB available. STATS driver startup time 0.018 START taper datestamp 20020816 label sb06_1 tape 0 INFO taper tape sb06_1 kb 0 fm 0 [OK] FINISH driver date 20020816 time 0.465 According to that, the flush would be 69gigs, but theres only 15 available. For tapeless backups, I'd skip the holding disk stuff, particularly if its on the same drive as the actual backup as it will use up all the backup space. -- Cheers, Gene AMD K6-III@500mhz 320M Athlon1600XP@1400mhz 512M 99.11% setiathome rank, not too shabby for a WV hillbilly
Host Down on AMCHECK - what are the common error
Hi again, I am getting there - I have installed and configured Amanda and e-mails are coming out and I can label tapes. When I run amcheck I get a Host down error - the host I am trying to dump is my amanda server. I thought I'd start there then roll it out to my other servers and PC. Can anybody point me in the right direction. I know it's vague but where do I start. Thanks Kevin
Re: Configuration help?
On Fri, Aug 16, 2002 at 10:07:54AM +0300, Conny Gyllendahl wrote: On Wed, 14 Aug 2002, Joshua Baker-LePain wrote: Now for my last question for this time: what are the pros and cons, if any, for using tar or (ufs)dump? Are there any reasons or situtations for choosing one over the other? One more point for tar (not better, just a capability) ... As Joshua pointed out tar can backup directory trees. This is sometimes necessary to split a huge partition into pieces that fit on a single tape. -- Jon H. LaBadie [EMAIL PROTECTED] JG Computing 4455 Province Line Road(609) 252-0159 Princeton, NJ 08540-4322 (609) 683-7220 (fax)
amcheck / amdump problem
Hello, during testing, I have noticed that amcheck finds a suitable tape for writing. Then, after invoking amdump, amdump also starts checking from slot 1 all over again ( a lengthly process )., even though -slot current is given on the command line. Does anyone know why it is reverting back to slot 1 again after amcheck runs? Should amcheck not already have the correct tape positioned? Jay
Re: Labels and Barcodes
--On Friday, August 16, 2002 09:08:07 -0400 Jason Greenberg [EMAIL PROTECTED] wrote: Then how do you set the barcode? Is it a physical label? If my libary reads bar codes, will it also print them? It is a physical (usually paper) label normally on the edge of the tape cartridge that is facing out when the tape is in the drive or magazine (similar to the UPC code on most products that the cashier scans when you buy just about anything). I doubt if any libraries print barcode labels (except possibly some of the room-sized robotics systems). You can either buy them from places like colorflex or print your own. Frank -- Frank Smith[EMAIL PROTECTED] Systems Administrator Voice: 512-374-4673 Hoover's Online Fax: 512-374-4501
Re: Host Down on AMCHECK - what are the common error
On Fri, 16 Aug 2002 at 4:12pm, Kevin Passey wrote When I run amcheck I get a Host down error - the host I am trying to dump is my amanda server. I thought I'd start there then roll it out to my other servers and PC. Can anybody point me in the right direction. I know it's vague but where do I start. http://amanda.sourceforge.net/fom-serve/cache/13.html -- Joshua Baker-LePain Department of Biomedical Engineering Duke University
Re: Is it advisable to have the holding disk on a separate partition?
On Fri, Aug 16, 2002 at 10:50:51AM -0400, Gene Heskett wrote: On Friday 16 August 2002 05:17, Kevin Passey wrote: Or can I change the conf file and put it as /usr/dumps/amanda. When I then backup /usr/ I would then exclude or include /dumps in my disklist file. Or have I got it completely wrong Thanks anyway Kevin You can use the exclude file if you're using tar, in which case you can put it where the most free space is. Mine's in /usr/dumps as I had 27 gigs free there a year ago when I set this one up, and is so entered in amanda.conf IIRC. The exclude files entries are in relative format, meaning that the entry line for excluding /usr/dumps looks like './dumps/' without the quotes. The leading ./ is very important. I have dumps on three different partitions. In each case I called the directory dumps and put it directly under the partitions root dir. That way Gene's exclude entry works for all three of them. -- Jon H. LaBadie [EMAIL PROTECTED] JG Computing 4455 Province Line Road(609) 252-0159 Princeton, NJ 08540-4322 (609) 683-7220 (fax)
Re: amrecover
You might want to specify your tape server, index server and tape device when
you run amrecover. Is calvin you tape server or the box you are trying to
recover to, or both? Try running amrecover with the following args:
amrecover -C TwiceAWeek -s {index server} -t { tape server} -d {tape device}
so if calvin is your tape server it would look something like
amrecover -C TwiceAWeek -s calvin -t calvin -d /dev/rmt/0bn (or whatever
your tape device is)
[EMAIL PROTECTED] wrote:
Hi!
We have an Amanda 2.4.0 Server running and want to recover some files.
I start
amrecover -C TwiceAWeek
on the Server, I get a timeout.
In the daemon.log :
Aug 16 13:47:34 calvin amindexd[31263]: warning: can't get client address:
Connection timed out
Aug 16 13:47:34 calvin amindexd[31263]: connect from unknown
Aug 16 13:49:16 calvin amindexd[31264]: warning: can't get client address:
Connection timed out
Aug 16 13:49:16 calvin amindexd[31264]: connect from unknown
In the .amandahosts I set the localhost to root.
What can I do?
Malte
Re: Host Down on AMCHECK - what are the common error
On Friday 16 August 2002 11:12, Kevin Passey wrote: Hi again, I am getting there - I have installed and configured Amanda and e-mails are coming out and I can label tapes. When I run amcheck I get a Host down error - the host I am trying to dump is my amanda server. I thought I'd start there then roll it out to my other servers and PC. Can anybody point me in the right direction. I know it's vague but where do I start. Thats usually because the server stuff isn't being run automaticly by the clients request. Have to read that part about using inetd yet, or does your system use xinetd? Thanks Kevin -- Cheers, Gene AMD K6-III@500mhz 320M Athlon1600XP@1400mhz 512M 99.11% setiathome rank, not too shabby for a WV hillbilly
Re: Host Down on AMCHECK - what are the common error
On Friday 16 August 2002 11:31, Joshua Baker-LePain wrote:
On Fri, 16 Aug 2002 at 4:12pm, Kevin Passey wrote
When I run amcheck I get a Host down error - the host I am
trying to dump is my amanda server.
I thought I'd start there then roll it out to my other servers
and PC.
Can anybody point me in the right direction. I know it's vague
but where do I start.
http://amanda.sourceforge.net/fom-serve/cache/13.html
Unforch, there's no mention of what to do if the system doesn't use
inetd, using xinetd instead.
For xinetd systems, it needs a file, or 3, I use one combined file
and it seems to work just fine. This file is in /etc/xinetd.d as
'amanda', here is mine:
-
# default = off
#
# description: Part of the Amanda server package
# This is the list of daemons such it needs
service amanda
{
disable = no
socket_type = dgram
protocol= udp
wait= yes
user= amanda
group = disk
groups = yes
server = /usr/local/libexec/amandad
}
service amandaidx
{
disable = no
socket_type = stream
protocol= tcp
wait= no
user= amanda
group = disk
groups = yes
server = /usr/local/libexec/amindexd
}
service amidxtape
{
disable = no
socket_type = stream
protocol= tcp
wait= no
user= amanda
group = disk
groups = yes
server = /usr/local/libexec/amidxtaped
}
This really does need to be added to the link above in order to
cover all the bases.
--
Cheers, Gene
AMD K6-III@500mhz 320M
Athlon1600XP@1400mhz 512M
99.11% setiathome rank, not too shabby for a WV hillbilly
restoring a directory structure?
What is the best way to restore an entire directory structure from an amanda dump? I'm using tar for my dumps so all of my disklist entries are directories, not filesystems. It's easy enough to go in through amrecover and restore individual files (which works) and I am sure I can get amrestore to work (once I figure out the correct syntax) but I can't find a simple way of restoring a whole directory tree. Say I want to restore all the files in a users home directory (/home/users/joeuser) and since he has multiple subdirectories (which probably contain subdirectories) I need a recursive restore of some kind. Just saying add /home/users/joeuser/* doesn't work, since that will only restore files in that directory as well as creating any subdirectories that are off there (without restoring the contents of those directories). Any pointers would be appreciated. Nicholas Paufler Systems Administrator The Internet Centre
Re: Host Down on AMCHECK - what are the common error
On Fri, 16 Aug 2002 at 11:58am, Gene Heskett wrote Unforch, there's no mention of what to do if the system doesn't use inetd, using xinetd instead. *snip This really does need to be added to the link above in order to cover all the bases. Go for it -- that's the beauty of F-O-M. -- Joshua Baker-LePain Department of Biomedical Engineering Duke University
Re: Host Down on AMCHECK - what are the common error
On Friday 16 August 2002 12:21, Joshua Baker-LePain wrote: On Fri, 16 Aug 2002 at 11:58am, Gene Heskett wrote Unforch, there's no mention of what to do if the system doesn't use inetd, using xinetd instead. *snip This really does need to be added to the link above in order to cover all the bases. Go for it -- that's the beauty of F-O-M. Not familiar with that at all. Email me with the details please. -- Cheers, Gene AMD K6-III@500mhz 320M Athlon1600XP@1400mhz 512M 99.11% setiathome rank, not too shabby for a WV hillbilly
amrecover - can't talk to tape server
Any idea what could cause this? the same thing happens when connecting from the tape server (local or remote). amrecover add passwd Added /passwd amrecover extract Extracting files using tape drive /dev/nst0 on host 192.168.100.1 The following tapes are needed: E04 Restoring files into directory /home/jg/recover-tmp Continue? [Y/n]: y Load tape E04 now Continue? [Y/n]: y cannot connect to 192.168.100.1: Connection refused amrecover - can't talk to tape server -- Jay
Re: amrecover - can't talk to tape server
Does your inetd.conf (or equivalent) have an entry for amidxtape? You'll need one for amandaidx to if it is also your index server. Frank --On Friday, August 16, 2002 12:33:54 -0400 Jason Greenberg [EMAIL PROTECTED] wrote: Any idea what could cause this? the same thing happens when connecting from the tape server (local or remote). amrecover add passwd Added /passwd amrecover extract Extracting files using tape drive /dev/nst0 on host 192.168.100.1 The following tapes are needed: E04 Restoring files into directory /home/jg/recover-tmp Continue? [Y/n]: y Load tape E04 now Continue? [Y/n]: y cannot connect to 192.168.100.1: Connection refused amrecover - can't talk to tape server -- Jay -- Frank Smith[EMAIL PROTECTED] Systems Administrator Voice: 512-374-4673 Hoover's Online Fax: 512-374-4501
Re: Host Down on AMCHECK - what are the common error
On Fri, Aug 16, 2002 at 12:37:52PM -0400, Gene Heskett wrote: This really does need to be added to the link above in order to cover all the bases. Go for it -- that's the beauty of F-O-M. Not familiar with that at all. Email me with the details please. Go to amanda.org, click on faq-o-matic, click on add a new answer -- Jon H. LaBadie [EMAIL PROTECTED] JG Computing 4455 Province Line Road(609) 252-0159 Princeton, NJ 08540-4322 (609) 683-7220 (fax)
Re: Configuration help?
On Fri, Aug 16, 2002 at 10:07:54AM +0300, Conny Gyllendahl wrote: Now for my last question for this time: what are the pros and cons, if any, for using tar or (ufs)dump? Are there any reasons or situtations for choosing one over the other? You say ufs(dump), so I'm assuming recent Solaris. The calculation is different for Linux. ufsdump plus: - Gets all file system/file attributes, period, no ifs ands or buts, even ones you don't know are there. :-) - Does not touch atime on files. - Does not require running as root. - In my experience, on file systems with large numbers of small files, estimates and incrementals are much faster than GNU tar. ufsdump minus: - No exclude list. - No splitting the file system. - Data is not portable to other OS's. tar plus: - Flexibility. Excludes, splitting. - 100% portable, ubiquitous. (But see also tar minus) tar minus: - Touches atime. - In my experience, on file systems with large numbers of small files, estimates and incrementals can take a long time. - Portable, but: GNU tar output with very long paths/names is only guaranteed to be readable by another GNU tar of similar version. Reading with non-GNU tar (or older GNU tar) may generate errors, or garbled paths/names. Depending on the versions involved, things get interesting at name lengths of 100 or 256, and path lengths of 256 or 1024. This is a very minor minus, you just need to be aware. - Has to run as root (no, I don't lose any sleep worrying about runtar exploits! :-) Speed folklore: In my experience, on full backups, on modern Solaris kit, GNU tar is a bit faster than ufsdump, *not* slower. Disks have gotten a lot faster (helps the more-random seeks that tar has to do), ufs and the processors have gotten faster (so ufs isn't in the way any more), and ufsdumps's initial mapping (wonderful for 'ufsrestore -i') is just time down the drain from amanda's POV. There, no religion here, I think. Look at the plus/minus lists and make up your own mind. As the Perl folk say, TMTOWTDI. -- Jay Lessert [EMAIL PROTECTED] Accelerant Networks Inc. (voice)1.503.439.3461 Beaverton OR, USA(fax)1.503.466.9472
utilities hanging with -slot clean process
Hello, Any help on this matter will be greatly appreciated. During many of my operations, such as amtape, amdump, before anything happens, the process just seems to hang. It must be caught in a loop somewhere. I think it has something to do with the cleaning tape, because during this command: bash-2.05a$ /usr/sbin/amtape Execulink label Execulink02 amtape: scanning for tape with label Execulink02 I see this in the process list, once it has hung: [root@lon-backup root]# ps -feaw | grep zd amanda 19720 19675 0 13:22 pts/000:00:00 /bin/sh /usr/lib/amanda/chg-zd-mtx -search E02 amanda 19780 19720 1 13:22 pts/000:00:06 /bin/sh /usr/lib/amanda/chg-zd-mtx -slot clean amanda 20297 19780 0 13:29 pts/000:00:00 /bin/sh /usr/lib/amanda/chg-zd-mtx -slot clean If I ctrl-c, and try again, it will work. This seems to happen often, but not consistently. Any thoughts? -- Jay
Re: amrecover - can't talk to tape server
That does it, thanks a lot ! On Fri, 2002-08-16 at 13:01, Frank Smith wrote: Does your inetd.conf (or equivalent) have an entry for amidxtape? You'll need one for amandaidx to if it is also your index server. Frank --On Friday, August 16, 2002 12:33:54 -0400 Jason Greenberg [EMAIL PROTECTED] wrote: Any idea what could cause this? the same thing happens when connecting from the tape server (local or remote). amrecover add passwd Added /passwd amrecover extract Extracting files using tape drive /dev/nst0 on host 192.168.100.1 The following tapes are needed: E04 Restoring files into directory /home/jg/recover-tmp Continue? [Y/n]: y Load tape E04 now Continue? [Y/n]: y cannot connect to 192.168.100.1: Connection refused amrecover - can't talk to tape server -- Jay -- Frank Smith[EMAIL PROTECTED] Systems Administrator Voice: 512-374-4673 Hoover's Online Fax: 512-374-4501 -- Jason Greenberg, CCNP Network Administrator Execulink, Inc. [EMAIL PROTECTED]
Re: Labels and Barcodes
The barcode is printed on the physical label on the tape. It is read by a laser and stored in the changer memory. I have no idea if the changer can print barcode labels but there is software that can. I buy my tapes AIT-2 with barcodes labels already on them so that is another option. On 16 Aug 2002, Jason Greenberg wrote: - Then how do you set the barcode? Is it a physical label? If my libary - reads bar codes, will it also print them? - - On Thu, 2002-08-15 at 16:25, Stephen Carville wrote: - On 15 Aug 2002, Jason Greenberg wrote: - - - What's the difference between labels and barcodes? Why, when I label - - tapes, does the VolumeTag= not show up for that tape? - - To amanda, a label is the information in the fist part of a tape. It - identifies what backup set the tape belongs to, when it was last used, - etc. Labels have nothing to do with the 'Volume Tag which are from - the barcodes. - - -- - -- Stephen Carville - UNIX and Network Administrator - DPSI (formerly Ace USA Flood Services) - 310-342-3602 - [EMAIL PROTECTED] - - - -- -- Stephen Carville UNIX and Network Administrator DPSI (formerly Ace USA Flood Services) 310-342-3602 [EMAIL PROTECTED]
Re: Configuration help?
On Fri, Aug 16, 2002 at 10:08:49AM -0700, Jay Lessert wrote:
There, no religion here, I think. Look at the plus/minus lists and
make up your own mind. As the Perl folk say, TMTOWTDI.
For those like me who did not know Jay's acronym:
$ dict TMTOWTDI
*** Source: Jargon File (4.3.0, 30 APR 2001) ***
TMTOWTDI /tim-toh'-dee/There's More Than One Way To Do It. This
abbreviation of the official motto of {Perl} is frequently used on
newsgroups and mailing lists related to that language.
--
Jon H. LaBadie [EMAIL PROTECTED]
JG Computing
4455 Province Line Road(609) 252-0159
Princeton, NJ 08540-4322 (609) 683-7220 (fax)
Confused, irritated. Small dumpcycle+tapecycle.
Howdy, Have been running amanda for a while now, but I'm really battling to get it configured right. Almost every day now I have to manually run amflush because amanda is generating dumps that are too big. I have a very small (relatively speaking) tapecycle: 7 tapes. I understand that my dumpcycle should be such that I end up with two full dumps of each disk per tapecycle. But amanda plans according to days, not runs. Am I being unreasonable here. I have Travan 4/8 GB tapes and am backing up around 20GB of data, but each disk is less than 2GB compressed. Should I be splitting the config? Arrrgggh! Regards, Brian Jonnes -- Init Systems - Linux consulting 031 767-0139082 769-2320[EMAIL PROTECTED]
Re: Confused, irritated. Small dumpcycle+tapecycle.
On Fri, Aug 16, 2002 at 07:59:52PM +0200, Brian Jonnes wrote: Howdy, Have been running amanda for a while now, but I'm really battling to get it configured right. Almost every day now I have to manually run amflush because amanda is generating dumps that are too big. I have a very small (relatively speaking) tapecycle: 7 tapes. I understand that my dumpcycle should be such that I end up with two full dumps of each Well the optimum is 40 full dumps per tapecycle -- JOKE!! No, it is not that you should have 2, but it is recommended that you have at least 2. My config happens to have a minimum of 4 in 24 tapes. disk per tapecycle. But amanda plans according to days, not runs. Have you seen the parameter runspercycle. It is unitless, not days or weeks. Am I being unreasonable here. I have Travan 4/8 GB tapes and am backing up around 20GB of data, but each disk is less than 2GB compressed. Should I be splitting the config? Some info about your config would eliminate guess work. 2 dumpcycles in 7 tapes, looks like 3 runs per dumpcycle plus 1 extra safety tape. OK, your dumpcycle could be 1 week, 3 runs MWF? Or 3 days, dump every day, but in each case 3 runs per dumpcycle. That gives us 3 tapes @ about 3.8GB/tape, about 11.4GB to work with for a set of full dumps plus incrementals. After compression, is there any possibility of your 20GB reliably fitting in this amount of tape along with the incrementals? Only you know your data, but it does not seem likely to me. -- Jon H. LaBadie [EMAIL PROTECTED] JG Computing 4455 Province Line Road(609) 252-0159 Princeton, NJ 08540-4322 (609) 683-7220 (fax)
Restoration of single file via amrestore?
Is it possible to restore one particular file from tape, using amrestore (vs. amrecover from the client)? In the book(O'Reilly) it says, The amrestore command restores whole images from tape. I'm assuming images means entire filesystems. Can you also get individual files...? wab
Mailslot with Amanda
Could anyone please explain the theory behind how the autoloader mailslot works with amanda? If possible, does anyone have a sample dumpcycle, with runs and tape #'s, and an example of how they use the mailslot? I would like to take a tape offsite weekly, but how can I have amanda prepare this for me?
Re: Restoration of single file via amrestore?
On Fri, 16 Aug 2002 at 1:24pm, Wayne Byarlay wrote Is it possible to restore one particular file from tape, using amrestore (vs. amrecover from the client)? In the book(O'Reilly) it says, The amrestore command restores whole images from tape. I'm assuming images means entire filesystems. Can you also get individual files...? By piping the amrestore output to your recovery program, yes. But you can also run amrecover on any machine and 'sethost' to the client for which you want to recover files. -- Joshua Baker-LePain Department of Biomedical Engineering Duke University
file system sizes
Can anybody recommend the best way to estimate whether all my filesystems will fit onto one tape? Right now I'm running a 15-day incremental cycle, but I want to remove friday from that cycle, and do a full backup. The problem is, it'll take me a LONG time to login to each server, df -k, etc. to estimate the sizes. I'm guessing that maybe Amanda might have this info stored away in some secret file somewhere, or I can use some deft command-line switches to audit all my file systems. Any ideas? wab.
Re: utilities hanging with -slot clean process
On Friday 16 August 2002 13:34, Jason Greenberg wrote: Hello, Any help on this matter will be greatly appreciated. During many of my operations, such as amtape, amdump, before anything happens, the process just seems to hang. It must be caught in a loop somewhere. I think it has something to do with the cleaning tape, because during this command: bash-2.05a$ /usr/sbin/amtape Execulink label Execulink02 amtape: scanning for tape with label Execulink02 I see this in the process list, once it has hung: [root@lon-backup root]# ps -feaw | grep zd amanda 19720 19675 0 13:22 pts/000:00:00 /bin/sh /usr/lib/amanda/chg-zd-mtx -search E02 amanda 19780 19720 1 13:22 pts/000:00:06 /bin/sh /usr/lib/amanda/chg-zd-mtx -slot clean amanda 20297 19780 0 13:29 pts/000:00:00 /bin/sh /usr/lib/amanda/chg-zd-mtx -slot clean If I ctrl-c, and try again, it will work. This seems to happen often, but not consistently. Any thoughts? This might be off the wall to some, but this puppy has played at my place before. I had some problems with that earlier, and they all went by-by when I stuffed the scsi card in the trash, ISTR it was an older adaptec 1540/1542 variation. I also finally gave up making a different copy of that card work here at home, seems it doesn't like tape drives with robots or something. Here I bought an advansys ABP-940 and all is now well, and the other machine now has an adaptec 2940 variation in it, working just fine. How close am I? -- Cheers, Gene AMD K6-III@500mhz 320M Athlon1600XP@1400mhz 512M 99.11% setiathome rank, not too shabby for a WV hillbilly
Re: file system sizes
If you've been saving your daily reports, just grep for the level 0's over a dumpcycle and add them up. Frank --On Friday, August 16, 2002 14:16:45 -0500 Wayne Byarlay [EMAIL PROTECTED] wrote: Can anybody recommend the best way to estimate whether all my filesystems will fit onto one tape? Right now I'm running a 15-day incremental cycle, but I want to remove friday from that cycle, and do a full backup. The problem is, it'll take me a LONG time to login to each server, df -k, etc. to estimate the sizes. I'm guessing that maybe Amanda might have this info stored away in some secret file somewhere, or I can use some deft command-line switches to audit all my file systems. Any ideas? wab. -- Frank Smith[EMAIL PROTECTED] Systems Administrator Voice: 512-374-4673 Hoover's Online Fax: 512-374-4501
Re: file system sizes
On Fri, Aug 16, 2002 at 02:16:45PM -0500, Wayne Byarlay wrote: Can anybody recommend the best way to estimate whether all my filesystems will fit onto one tape? Right now I'm running a 15-day incremental cycle, but I want to remove friday from that cycle, and do a full backup. The problem is, it'll take me a LONG time to login to each server, df -k, etc. to estimate the sizes. I'm guessing that maybe Amanda might have this info stored away in some secret file somewhere, or I can use some deft command-line switches to audit all my file systems. Any ideas? amadmin has all sorts of neat reports. Check out the man pages. For example: amanda$ amadmin ds1 balance due-date #fs orig KBout KB balance --- 8/16 Fri1 5407150 3967936+0.9% 8/17 Sat0 0 0 --- 8/18 Sun1 7942010 6118976 +55.6% 8/19 Mon1660340630240 -84.0% 8/20 Tue1 7421420 5370048 +36.6% 8/21 Wed6 4950160 3875904-1.4% 8/22 Thu9 8486540 3624704-7.8% --- TOTAL 19 34867620 23587808 3931301 (estimated 6 runs per dumpcycle) I think the totals are the before and after compression sizes of the planned level 0's for the next dumpcycle. For individual file systems, see your curinfo dir. Under host/filesys there should be a file named info. The line beginning stats: 0, eg. stats: 0 635750 321824 208 1029397183 12 DS1-10 has the last uncompressed (635750KB) and compressed (321824) sizes. You should be able to get them all with grep '^stats: 0 ' curinfo/*/*/info -- Jon H. LaBadie [EMAIL PROTECTED] JG Computing 4455 Province Line Road(609) 252-0159 Princeton, NJ 08540-4322 (609) 683-7220 (fax)
Re: file system sizes
On Fri, 16 Aug 2002, Frank Smith wrote: If you've been saving your daily reports, just grep for the level 0's over a dumpcycle and add them up. If you haven't been saving them, you can easily recreate them: $ amreport DailySet1 -l /var/adm/amanda/DailySet1/log.20020726.0 -f /dev/tty -p /dev/null -- Ulrik Sandberg
