amrecover asking for tape drive on old server
puzzle. If you recall, I switched from Amanda 2.5.1p3 on a Solaris 9 server on an E250 to Amanda 3.3.3 on an Ubuntu 12.04 server on SuperMicro around May 22. The transition seemed to run smoothly. The old E250 is still running mail services, and the new Amanda server is still backing that up. So, now someone wants their mail files recovered from May 26. Cool. Tapes were written by the new server on that date. But, when I run amrecover, I ran into the situation below. It seems to be searching for a tape drive on the old server rather than on the new server. Weird. I've been running Amanda on the new server with the tape library and tape drive on the new server for several weeks and it has been working just fine as far as the Amanda email reports go. I went through the amanda.conf, changer.conf, etc. and find no references to any server. It should just be looking for the devices on the Amanda server that I am running this on. Any idea what's up with this? And how I can fix it? I know I can just read the tape using native facilities, but I really don't want to fall back on that if I don't have to. This should all just work, as it always has in the past. -- attempted amrecover session --- chrisho@supereclogite:/usr/local/adm$ cd /amanda1 chrisho@supereclogite:/amanda1$ sudo amrecover daily [sudo] password for chrisho: AMRECOVER Version 3.3.3. Contacting server on eclogite.geo.umass.edu ... 220 eclogite AMANDA index server (2.5.1p3) ready. Setting restore date to today (2013-06-17) 200 Working date set to 2013-06-17. 200 Config set to daily. 501 Host supereclogite is not in your disklist. Trying host supereclogite.geo.umass.edu ... 200 Dump host set to supereclogite.geo.umass.edu. Use the setdisk command to choose dump disk to recover amrecover sethost eclogite.geo.umass.edu 200 Dump host set to eclogite.geo.umass.edu. amrecover setdisk /var/mail 200 Disk set to /var/mail. amrecover setdate --05-26 200 Working date set to 2013-05-26. amrecover add cooke Added file /cooke amrecover extract Extracting files using tape drive LIB-162A5 on host eclogite.geo.umass.edu. The following tapes are needed: geo-daily-13 Extracting files using tape drive LIB-162A5 on host eclogite.geo.umass.edu. Load tape geo-daily-13 now Continue [?/Y/n/s/d]? y could not stat 'LIB-162A5': No such file or directory Load tape geo-daily-13 now Continue [?/Y/n/d]? ? Enter yes to continue, no to stop or d to change to a new device Continue [?/Y/n/d]? d New device name [?]: /dev/nst0 Using tape /dev/nst0 from server eclogite.geo.umass.edu. Continue [?/Y/n/d]? y could not stat '/dev/nst0': No such file or directory could not stat '/dev/nst0': No such file or directory Load tape geo-daily-13 now Continue [?/Y/n/d]? d New device name [?]: ? Enter a new device name or default New device name [?]: default Using default tape from server eclogite.geo.umass.edu. Continue [?/Y/n/d]? y Got no header and data from server, check in amidxtaped.*.debug and amandad.*.debug files on server amrecover exit 200 Good bye. -- --- Chris Hoogendyk - O__ Systems Administrator c/ /'_ --- Biology Geology Departments (*) \(*) -- 347 Morrill Science Center ~~ - University of Massachusetts, Amherst hoogen...@bio.umass.edu --- Erdös 4
Re: amrecover asking for tape drive on old server
The default index and tape server are set when amanda is compiled, at the configure step. You can overwrite the default different ways: amrecover -s ... -t ...(man amrecover) add them to amanda-client.conf (man amanda-client.conf) you can also change the tape server inside amrecover: setdevice [[-h tape-server] tapedev] (help command inside amrecover) You can't change the index server inside amrecover. jean-Louis On 06/17/2013 01:47 PM, Chris Hoogendyk wrote: puzzle. If you recall, I switched from Amanda 2.5.1p3 on a Solaris 9 server on an E250 to Amanda 3.3.3 on an Ubuntu 12.04 server on SuperMicro around May 22. The transition seemed to run smoothly. The old E250 is still running mail services, and the new Amanda server is still backing that up. So, now someone wants their mail files recovered from May 26. Cool. Tapes were written by the new server on that date. But, when I run amrecover, I ran into the situation below. It seems to be searching for a tape drive on the old server rather than on the new server. Weird. I've been running Amanda on the new server with the tape library and tape drive on the new server for several weeks and it has been working just fine as far as the Amanda email reports go. I went through the amanda.conf, changer.conf, etc. and find no references to any server. It should just be looking for the devices on the Amanda server that I am running this on. Any idea what's up with this? And how I can fix it? I know I can just read the tape using native facilities, but I really don't want to fall back on that if I don't have to. This should all just work, as it always has in the past. -- attempted amrecover session --- chrisho@supereclogite:/usr/local/adm$ cd /amanda1 chrisho@supereclogite:/amanda1$ sudo amrecover daily [sudo] password for chrisho: AMRECOVER Version 3.3.3. Contacting server on eclogite.geo.umass.edu ... 220 eclogite AMANDA index server (2.5.1p3) ready. Setting restore date to today (2013-06-17) 200 Working date set to 2013-06-17. 200 Config set to daily. 501 Host supereclogite is not in your disklist. Trying host supereclogite.geo.umass.edu ... 200 Dump host set to supereclogite.geo.umass.edu. Use the setdisk command to choose dump disk to recover amrecover sethost eclogite.geo.umass.edu 200 Dump host set to eclogite.geo.umass.edu. amrecover setdisk /var/mail 200 Disk set to /var/mail. amrecover setdate --05-26 200 Working date set to 2013-05-26. amrecover add cooke Added file /cooke amrecover extract Extracting files using tape drive LIB-162A5 on host eclogite.geo.umass.edu. The following tapes are needed: geo-daily-13 Extracting files using tape drive LIB-162A5 on host eclogite.geo.umass.edu. Load tape geo-daily-13 now Continue [?/Y/n/s/d]? y could not stat 'LIB-162A5': No such file or directory Load tape geo-daily-13 now Continue [?/Y/n/d]? ? Enter yes to continue, no to stop or d to change to a new device Continue [?/Y/n/d]? d New device name [?]: /dev/nst0 Using tape /dev/nst0 from server eclogite.geo.umass.edu. Continue [?/Y/n/d]? y could not stat '/dev/nst0': No such file or directory could not stat '/dev/nst0': No such file or directory Load tape geo-daily-13 now Continue [?/Y/n/d]? d New device name [?]: ? Enter a new device name or default New device name [?]: default Using default tape from server eclogite.geo.umass.edu. Continue [?/Y/n/d]? y Got no header and data from server, check in amidxtaped.*.debug and amandad.*.debug files on server amrecover exit 200 Good bye.
Re: amrecover asking for tape drive on old server
from the session that I included, supereclogite is the new 3.3.3 Amanda server, and eclogite is the old 2.5.1p3 Amanda server. The old server, eclogite, was added to the disklist on the new server, supereclogite, and has been backed up there since. My session is on the new Amanda server asking to recover a file that was backed up to the new Amanda server and put to tape on the new Amanda server, but the file is from the old server. So I'm doing everything on the new Amanda server. The old one is a client to the new one. hmm. Interesting. As I go over this in my mind, I'm realizing that the client was using ufsdump and then sending the resulting file over to the server. But, the server is Ubuntu. It cannot do a ufsrestore. Interesting. This is the first time I have had an Ubuntu server as the Amanda server (I've had a bunch of them as clients), and it seems that it breaks the work flow that I am used to. Still, I will need to straighten out what it is doing with regard to tape drive and index server. On 6/17/13 2:06 PM, Brian Cuttler wrote: Chris, You are using the amanda client/server and issuing the restore from the E250? Or you are restoring the tar/dump file on the server and (if dump) porting the (decompressed) file back to the client to unpack? If the first option, maybe you are still referencing the config on the E250, that is, the server config, rather than a client config. On Mon, Jun 17, 2013 at 01:47:08PM -0400, Chris Hoogendyk wrote: puzzle. If you recall, I switched from Amanda 2.5.1p3 on a Solaris 9 server on an E250 to Amanda 3.3.3 on an Ubuntu 12.04 server on SuperMicro around May 22. The transition seemed to run smoothly. The old E250 is still running mail services, and the new Amanda server is still backing that up. So, now someone wants their mail files recovered from May 26. Cool. Tapes were written by the new server on that date. But, when I run amrecover, I ran into the situation below. It seems to be searching for a tape drive on the old server rather than on the new server. Weird. I've been running Amanda on the new server with the tape library and tape drive on the new server for several weeks and it has been working just fine as far as the Amanda email reports go. I went through the amanda.conf, changer.conf, etc. and find no references to any server. It should just be looking for the devices on the Amanda server that I am running this on. Any idea what's up with this? And how I can fix it? I know I can just read the tape using native facilities, but I really don't want to fall back on that if I don't have to. This should all just work, as it always has in the past. -- attempted amrecover session --- chrisho@supereclogite:/usr/local/adm$ cd /amanda1 chrisho@supereclogite:/amanda1$ sudo amrecover daily [sudo] password for chrisho: AMRECOVER Version 3.3.3. Contacting server on eclogite.geo.umass.edu ... 220 eclogite AMANDA index server (2.5.1p3) ready. Setting restore date to today (2013-06-17) 200 Working date set to 2013-06-17. 200 Config set to daily. 501 Host supereclogite is not in your disklist. Trying host supereclogite.geo.umass.edu ... 200 Dump host set to supereclogite.geo.umass.edu. Use the setdisk command to choose dump disk to recover amrecover sethost eclogite.geo.umass.edu 200 Dump host set to eclogite.geo.umass.edu. amrecover setdisk /var/mail 200 Disk set to /var/mail. amrecover setdate --05-26 200 Working date set to 2013-05-26. amrecover add cooke Added file /cooke amrecover extract Extracting files using tape drive LIB-162A5 on host eclogite.geo.umass.edu. The following tapes are needed: geo-daily-13 Extracting files using tape drive LIB-162A5 on host eclogite.geo.umass.edu. Load tape geo-daily-13 now Continue [?/Y/n/s/d]? y could not stat 'LIB-162A5': No such file or directory Load tape geo-daily-13 now Continue [?/Y/n/d]? ? Enter yes to continue, no to stop or d to change to a new device Continue [?/Y/n/d]? d New device name [?]: /dev/nst0 Using tape /dev/nst0 from server eclogite.geo.umass.edu. Continue [?/Y/n/d]? y could not stat '/dev/nst0': No such file or directory could not stat '/dev/nst0': No such file or directory Load tape geo-daily-13 now Continue [?/Y/n/d]? d New device name [?]: ? Enter a new device name or default New device name [?]: default Using default tape from server eclogite.geo.umass.edu. Continue [?/Y/n/d]? y Got no header and data from server, check in amidxtaped.*.debug and amandad.*.debug files on server amrecover exit 200 Good bye. -- --- Chris Hoogendyk - O__ Systems Administrator c/ /'_ --- Biology Geology Departments (*) \(*) -- 347 Morrill Science Center ~~ - University of Massachusetts, Amherst hoogen...@bio.umass.edu --- Erdös 4 --- Brian R Cuttler brian.cutt...@wadsworth.org Computer Systems Support(v) 518 486-1697 Wadsworth Center
Re: amrecover asking for tape drive on old server
Chris, I think Jean-Louis nailed the problem with the server and tape unit. I think amrecover will pass the data stream back to the client and ufsrestore on the client side, I don't think you will have any issues after taking JML's advise. On Mon, Jun 17, 2013 at 02:30:50PM -0400, Chris Hoogendyk wrote: from the session that I included, supereclogite is the new 3.3.3 Amanda server, and eclogite is the old 2.5.1p3 Amanda server. The old server, eclogite, was added to the disklist on the new server, supereclogite, and has been backed up there since. My session is on the new Amanda server asking to recover a file that was backed up to the new Amanda server and put to tape on the new Amanda server, but the file is from the old server. So I'm doing everything on the new Amanda server. The old one is a client to the new one. hmm. Interesting. As I go over this in my mind, I'm realizing that the client was using ufsdump and then sending the resulting file over to the server. But, the server is Ubuntu. It cannot do a ufsrestore. Interesting. This is the first time I have had an Ubuntu server as the Amanda server (I've had a bunch of them as clients), and it seems that it breaks the work flow that I am used to. Still, I will need to straighten out what it is doing with regard to tape drive and index server. On 6/17/13 2:06 PM, Brian Cuttler wrote: Chris, You are using the amanda client/server and issuing the restore from the E250? Or you are restoring the tar/dump file on the server and (if dump) porting the (decompressed) file back to the client to unpack? If the first option, maybe you are still referencing the config on the E250, that is, the server config, rather than a client config. On Mon, Jun 17, 2013 at 01:47:08PM -0400, Chris Hoogendyk wrote: puzzle. If you recall, I switched from Amanda 2.5.1p3 on a Solaris 9 server on an E250 to Amanda 3.3.3 on an Ubuntu 12.04 server on SuperMicro around May 22. The transition seemed to run smoothly. The old E250 is still running mail services, and the new Amanda server is still backing that up. So, now someone wants their mail files recovered from May 26. Cool. Tapes were written by the new server on that date. But, when I run amrecover, I ran into the situation below. It seems to be searching for a tape drive on the old server rather than on the new server. Weird. I've been running Amanda on the new server with the tape library and tape drive on the new server for several weeks and it has been working just fine as far as the Amanda email reports go. I went through the amanda.conf, changer.conf, etc. and find no references to any server. It should just be looking for the devices on the Amanda server that I am running this on. Any idea what's up with this? And how I can fix it? I know I can just read the tape using native facilities, but I really don't want to fall back on that if I don't have to. This should all just work, as it always has in the past. -- attempted amrecover session --- chrisho@supereclogite:/usr/local/adm$ cd /amanda1 chrisho@supereclogite:/amanda1$ sudo amrecover daily [sudo] password for chrisho: AMRECOVER Version 3.3.3. Contacting server on eclogite.geo.umass.edu ... 220 eclogite AMANDA index server (2.5.1p3) ready. Setting restore date to today (2013-06-17) 200 Working date set to 2013-06-17. 200 Config set to daily. 501 Host supereclogite is not in your disklist. Trying host supereclogite.geo.umass.edu ... 200 Dump host set to supereclogite.geo.umass.edu. Use the setdisk command to choose dump disk to recover amrecover sethost eclogite.geo.umass.edu 200 Dump host set to eclogite.geo.umass.edu. amrecover setdisk /var/mail 200 Disk set to /var/mail. amrecover setdate --05-26 200 Working date set to 2013-05-26. amrecover add cooke Added file /cooke amrecover extract Extracting files using tape drive LIB-162A5 on host eclogite.geo.umass.edu. The following tapes are needed: geo-daily-13 Extracting files using tape drive LIB-162A5 on host eclogite.geo.umass.edu. Load tape geo-daily-13 now Continue [?/Y/n/s/d]? y could not stat 'LIB-162A5': No such file or directory Load tape geo-daily-13 now Continue [?/Y/n/d]? ? Enter yes to continue, no to stop or d to change to a new device Continue [?/Y/n/d]? d New device name [?]: /dev/nst0 Using tape /dev/nst0 from server eclogite.geo.umass.edu. Continue [?/Y/n/d]? y could not stat '/dev/nst0': No such file or directory could not stat '/dev/nst0': No such file or directory Load tape geo-daily-13 now Continue [?/Y/n/d]? d New device name [?]: ? Enter a new device name or default New device name [?]: default Using default tape from server eclogite.geo.umass.edu. Continue [?/Y/n/d]? y Got no header and data from server, check in amidxtaped.*.debug and amandad.*.debug files
Re: amrecover asking for tape drive on old server
On Mon, Jun 17, 2013 at 03:00:44PM -0400, Chris Hoogendyk wrote: hmm, didn't specify either of those in configure when I built Amanda 3.3.3. However, now that you mention amanda-client.conf, I went and looked for that. Turns out they were both specified there (I originally rsynced /usr/local/etc/amanda/ to the new server and then went about modifying stuff to set it up). However, changing the entries there and then running amrecover didn't seem to change anything. So, I tried `setdevice -h localhost /dev/nst0`, but it tells me it cannot talk to the tape server [request failed: timeout waiting for ACK]. I also tried chrisho@supereclogite:/amanda1$ sudo amrecover -s localhost -t localhost -d /dev/nst0 -h eclogite.geo.umass.edu -C daily AMRECOVER Version 3.3.3. Contacting server on localhost ... [request failed: timeout waiting for ACK] Totally puzzled, because the Amanda backups seems to have been running flawlessly. I can access the tape drive with mt and the changer with mtx, and Amanda has been writing stuff to tapes. As I mentioned in my reply to Brian, I'm going to have to change my workflow to recover stuff for the Solaris client. But, I'll still need to resolve some of these issues, because the old Amanda server might have stuff in its configuration and files that make it think it is still the server. Chris, If you are restoring on the server, you can just use amrestore, rather than amrecover, and pull the dump set back, amrestore will automatically decompress, if necessary, but will not unpack it if you leave off the -p option and do not pipe it to ufsrestore. Then you can scp the dump file to the E250. I believe that amrecover running on the client (once those issues are corrected) will decompress and pipe the stream to the client and run ufsrestore for you on the client. There shouldn't be an issue with trying to run ufsrestore on the linux amanda server, amanda was programmed smarter than that. On 6/17/13 2:04 PM, Jean-Louis Martineau wrote: The default index and tape server are set when amanda is compiled, at the configure step. You can overwrite the default different ways: amrecover -s ... -t ...(man amrecover) add them to amanda-client.conf (man amanda-client.conf) you can also change the tape server inside amrecover: setdevice [[-h tape-server] tapedev] (help command inside amrecover) You can't change the index server inside amrecover. jean-Louis On 06/17/2013 01:47 PM, Chris Hoogendyk wrote: puzzle. If you recall, I switched from Amanda 2.5.1p3 on a Solaris 9 server on an E250 to Amanda 3.3.3 on an Ubuntu 12.04 server on SuperMicro around May 22. The transition seemed to run smoothly. The old E250 is still running mail services, and the new Amanda server is still backing that up. So, now someone wants their mail files recovered from May 26. Cool. Tapes were written by the new server on that date. But, when I run amrecover, I ran into the situation below. It seems to be searching for a tape drive on the old server rather than on the new server. Weird. I've been running Amanda on the new server with the tape library and tape drive on the new server for several weeks and it has been working just fine as far as the Amanda email reports go. I went through the amanda.conf, changer.conf, etc. and find no references to any server. It should just be looking for the devices on the Amanda server that I am running this on. Any idea what's up with this? And how I can fix it? I know I can just read the tape using native facilities, but I really don't want to fall back on that if I don't have to. This should all just work, as it always has in the past. -- attempted amrecover session --- chrisho@supereclogite:/usr/local/adm$ cd /amanda1 chrisho@supereclogite:/amanda1$ sudo amrecover daily [sudo] password for chrisho: AMRECOVER Version 3.3.3. Contacting server on eclogite.geo.umass.edu ... 220 eclogite AMANDA index server (2.5.1p3) ready. Setting restore date to today (2013-06-17) 200 Working date set to 2013-06-17. 200 Config set to daily. 501 Host supereclogite is not in your disklist. Trying host supereclogite.geo.umass.edu ... 200 Dump host set to supereclogite.geo.umass.edu. Use the setdisk command to choose dump disk to recover amrecover sethost eclogite.geo.umass.edu 200 Dump host set to eclogite.geo.umass.edu. amrecover setdisk /var/mail 200 Disk set to /var/mail. amrecover setdate --05-26 200 Working date set to 2013-05-26. amrecover add cooke Added file /cooke amrecover extract Extracting files using tape drive LIB-162A5 on host eclogite.geo.umass.edu. The following tapes are needed: geo-daily-13 Extracting files using tape drive LIB-162A5 on host eclogite.geo.umass.edu. Load tape geo-daily-13 now Continue [?/Y/n/s/d]? y could not stat 'LIB-162A5':
Re: amrecover asking for tape drive on old server
hmm, didn't specify either of those in configure when I built Amanda 3.3.3. However, now that you mention amanda-client.conf, I went and looked for that. Turns out they were both specified there (I originally rsynced /usr/local/etc/amanda/ to the new server and then went about modifying stuff to set it up). However, changing the entries there and then running amrecover didn't seem to change anything. So, I tried `setdevice -h localhost /dev/nst0`, but it tells me it cannot talk to the tape server [request failed: timeout waiting for ACK]. I also tried chrisho@supereclogite:/amanda1$ sudo amrecover -s localhost -t localhost -d /dev/nst0 -h eclogite.geo.umass.edu -C daily AMRECOVER Version 3.3.3. Contacting server on localhost ... [request failed: timeout waiting for ACK] Totally puzzled, because the Amanda backups seems to have been running flawlessly. I can access the tape drive with mt and the changer with mtx, and Amanda has been writing stuff to tapes. As I mentioned in my reply to Brian, I'm going to have to change my workflow to recover stuff for the Solaris client. But, I'll still need to resolve some of these issues, because the old Amanda server might have stuff in its configuration and files that make it think it is still the server. On 6/17/13 2:04 PM, Jean-Louis Martineau wrote: The default index and tape server are set when amanda is compiled, at the configure step. You can overwrite the default different ways: amrecover -s ... -t ...(man amrecover) add them to amanda-client.conf (man amanda-client.conf) you can also change the tape server inside amrecover: setdevice [[-h tape-server] tapedev] (help command inside amrecover) You can't change the index server inside amrecover. jean-Louis On 06/17/2013 01:47 PM, Chris Hoogendyk wrote: puzzle. If you recall, I switched from Amanda 2.5.1p3 on a Solaris 9 server on an E250 to Amanda 3.3.3 on an Ubuntu 12.04 server on SuperMicro around May 22. The transition seemed to run smoothly. The old E250 is still running mail services, and the new Amanda server is still backing that up. So, now someone wants their mail files recovered from May 26. Cool. Tapes were written by the new server on that date. But, when I run amrecover, I ran into the situation below. It seems to be searching for a tape drive on the old server rather than on the new server. Weird. I've been running Amanda on the new server with the tape library and tape drive on the new server for several weeks and it has been working just fine as far as the Amanda email reports go. I went through the amanda.conf, changer.conf, etc. and find no references to any server. It should just be looking for the devices on the Amanda server that I am running this on. Any idea what's up with this? And how I can fix it? I know I can just read the tape using native facilities, but I really don't want to fall back on that if I don't have to. This should all just work, as it always has in the past. -- attempted amrecover session --- chrisho@supereclogite:/usr/local/adm$ cd /amanda1 chrisho@supereclogite:/amanda1$ sudo amrecover daily [sudo] password for chrisho: AMRECOVER Version 3.3.3. Contacting server on eclogite.geo.umass.edu ... 220 eclogite AMANDA index server (2.5.1p3) ready. Setting restore date to today (2013-06-17) 200 Working date set to 2013-06-17. 200 Config set to daily. 501 Host supereclogite is not in your disklist. Trying host supereclogite.geo.umass.edu ... 200 Dump host set to supereclogite.geo.umass.edu. Use the setdisk command to choose dump disk to recover amrecover sethost eclogite.geo.umass.edu 200 Dump host set to eclogite.geo.umass.edu. amrecover setdisk /var/mail 200 Disk set to /var/mail. amrecover setdate --05-26 200 Working date set to 2013-05-26. amrecover add cooke Added file /cooke amrecover extract Extracting files using tape drive LIB-162A5 on host eclogite.geo.umass.edu. The following tapes are needed: geo-daily-13 Extracting files using tape drive LIB-162A5 on host eclogite.geo.umass.edu. Load tape geo-daily-13 now Continue [?/Y/n/s/d]? y could not stat 'LIB-162A5': No such file or directory Load tape geo-daily-13 now Continue [?/Y/n/d]? ? Enter yes to continue, no to stop or d to change to a new device Continue [?/Y/n/d]? d New device name [?]: /dev/nst0 Using tape /dev/nst0 from server eclogite.geo.umass.edu. Continue [?/Y/n/d]? y could not stat '/dev/nst0': No such file or directory could not stat '/dev/nst0': No such file or directory Load tape geo-daily-13 now Continue [?/Y/n/d]? d New device name [?]: ? Enter a new device name or default New device name [?]: default Using default tape from server eclogite.geo.umass.edu. Continue [?/Y/n/d]? y Got no header and data from server, check in amidxtaped.*.debug and amandad.*.debug files on server amrecover exit 200 Good bye. -- --- Chris Hoogendyk - O__
Re: amrecover asking for tape drive on old server
OK. I've got amrestore pulling files off the tape. Then I'll transfer those over to the Solaris client and ufsrestore what I need locally there. That's less crude than going totally back to manual basics, but I would still like to work out amrecover. I'll see if I can get that set up to run on the client. I'll also need to test and see that it will work on the server when I'm dealing with recovering files for the server or for clients that are Ubuntu. On 6/17/13 3:12 PM, Brian Cuttler wrote: On Mon, Jun 17, 2013 at 03:00:44PM -0400, Chris Hoogendyk wrote: hmm, didn't specify either of those in configure when I built Amanda 3.3.3. However, now that you mention amanda-client.conf, I went and looked for that. Turns out they were both specified there (I originally rsynced /usr/local/etc/amanda/ to the new server and then went about modifying stuff to set it up). However, changing the entries there and then running amrecover didn't seem to change anything. So, I tried `setdevice -h localhost /dev/nst0`, but it tells me it cannot talk to the tape server [request failed: timeout waiting for ACK]. I also tried chrisho@supereclogite:/amanda1$ sudo amrecover -s localhost -t localhost -d /dev/nst0 -h eclogite.geo.umass.edu -C daily AMRECOVER Version 3.3.3. Contacting server on localhost ... [request failed: timeout waiting for ACK] Totally puzzled, because the Amanda backups seems to have been running flawlessly. I can access the tape drive with mt and the changer with mtx, and Amanda has been writing stuff to tapes. As I mentioned in my reply to Brian, I'm going to have to change my workflow to recover stuff for the Solaris client. But, I'll still need to resolve some of these issues, because the old Amanda server might have stuff in its configuration and files that make it think it is still the server. Chris, If you are restoring on the server, you can just use amrestore, rather than amrecover, and pull the dump set back, amrestore will automatically decompress, if necessary, but will not unpack it if you leave off the -p option and do not pipe it to ufsrestore. Then you can scp the dump file to the E250. I believe that amrecover running on the client (once those issues are corrected) will decompress and pipe the stream to the client and run ufsrestore for you on the client. There shouldn't be an issue with trying to run ufsrestore on the linux amanda server, amanda was programmed smarter than that. On 6/17/13 2:04 PM, Jean-Louis Martineau wrote: The default index and tape server are set when amanda is compiled, at the configure step. You can overwrite the default different ways: amrecover -s ... -t ...(man amrecover) add them to amanda-client.conf (man amanda-client.conf) you can also change the tape server inside amrecover: setdevice [[-h tape-server] tapedev] (help command inside amrecover) You can't change the index server inside amrecover. jean-Louis On 06/17/2013 01:47 PM, Chris Hoogendyk wrote: puzzle. If you recall, I switched from Amanda 2.5.1p3 on a Solaris 9 server on an E250 to Amanda 3.3.3 on an Ubuntu 12.04 server on SuperMicro around May 22. The transition seemed to run smoothly. The old E250 is still running mail services, and the new Amanda server is still backing that up. So, now someone wants their mail files recovered from May 26. Cool. Tapes were written by the new server on that date. But, when I run amrecover, I ran into the situation below. It seems to be searching for a tape drive on the old server rather than on the new server. Weird. I've been running Amanda on the new server with the tape library and tape drive on the new server for several weeks and it has been working just fine as far as the Amanda email reports go. I went through the amanda.conf, changer.conf, etc. and find no references to any server. It should just be looking for the devices on the Amanda server that I am running this on. Any idea what's up with this? And how I can fix it? I know I can just read the tape using native facilities, but I really don't want to fall back on that if I don't have to. This should all just work, as it always has in the past. -- attempted amrecover session --- chrisho@supereclogite:/usr/local/adm$ cd /amanda1 chrisho@supereclogite:/amanda1$ sudo amrecover daily [sudo] password for chrisho: AMRECOVER Version 3.3.3. Contacting server on eclogite.geo.umass.edu ... 220 eclogite AMANDA index server (2.5.1p3) ready. Setting restore date to today (2013-06-17) 200 Working date set to 2013-06-17. 200 Config set to daily. 501 Host supereclogite is not in your disklist. Trying host supereclogite.geo.umass.edu ... 200 Dump host set to supereclogite.geo.umass.edu. Use the setdisk command to choose dump disk to recover amrecover sethost eclogite.geo.umass.edu 200 Dump host set to eclogite.geo.umass.edu. amrecover setdisk /var/mail 200 Disk set to /var/mail. amrecover setdate --05-26 200 Working
