Re: missing backup data
James Williamson wrote: We've got Amanda set up to backup to a Onstream tape driver (SC-30), Are you aware there are two generations of Onstream tape drives? The first generation had firmware problems, which made it unusable. The company went down since, and you cannot get a firware upgrade. The european part of Onstream continued business, and, it seems, their hardware is much more reliable. (personally I have no experience with any Onstream device -- this is just what I heard.) If you're unlucky and have the first generation hardware, it could be the cause of your problems. when we run amdump it reports that it's successully backed up the contents of the disklist. However, when it comes to try and restore something on the disklist using amrecover it reports this error: EOF, check amidxtapid.debug file on backup. amrecover: short block 0 bytes UNKNOWN file amrecover: Can't read file header I've gone through the docs and a likely cause seems to be the wrong tape's in the drive and / or the tape needs rewinding. I've tried both these without any joy. In fact, I've got the dumpcycle set to 0 and the number of tapecycle set to 1. I'm under the impression that with this configuration Amanda will dump everything in my disklist daily. I've gone through the docs and tried this to see what's on the disk: amrestore -p /dev/nosst0 no-such-host /dev/null which returns this: amrestore:0: skipping start of tape: date 20030703 label nameon1 amrestore:1: skipping ..backup file1... Yet if I run it again I get this: amrestore: WARNING: not at start of tape file numbers will be offset amrestore:1: skipping ..backup file2... and so on. I'm assuming that amrestore invoked in this manner will iterate through each file without having to be manually restarted. Does this mean when I use amrecover I must wind the tape to the correct place? Please excuse me if my knowledge of tape driver is woefully inadequate. Yes you have to rewind yourself. Before amanda 2.4.4, it was usually faster to rewind and then fsf manually to the correct position on tape, and than start amrecover. If you didn't position the tape, amanda had to read the tape up that position, which is much slower than fsf. Since amanda 2.4.4, amanda can do that for you, if you add the directive amrecover_do_fsf on in your amanda.conf. -- Paul Bijnens, XplanationTel +32 16 397.511 Technologielaan 21 bus 2, B-3001 Leuven, BELGIUMFax +32 16 397.512 http://www.xplanation.com/ email: [EMAIL PROTECTED] *** * I think I've got the hang of it now: exit, ^D, ^C, ^\, ^Z, ^Q, F6, * * quit, ZZ, :q, :q!, M-Z, ^X^C, logoff, logout, close, bye, /bye, * * stop, end, F3, ~., ^]c, +++ ATH, disconnect, halt, abort, hangup, * * PF4, F20, ^X^X, :D::D, KJOB, F14-f-e, F8-e, kill -1 $$, shutdown, * * kill -9 1, Alt-F4, Ctrl-Alt-Del, AltGr-NumLock, Stop-A, ...* * ... Are you sure? ... YES ... Phew ... I'm out * ***
Re: missing backup data
Thanks for your help, James Williamson wrote: We've got Amanda set up to backup to a Onstream tape driver (SC-30), Are you aware there are two generations of Onstream tape drives? Yes, I've got the second generation. when we run amdump it reports that it's successully backed up the contents of the disklist. However, when it comes to try and restore something on the disklist using amrecover it reports this error: EOF, check amidxtapid.debug file on backup. amrecover: short block 0 bytes UNKNOWN file amrecover: Can't read file header I've gone through the docs and a likely cause seems to be the wrong tape's in the drive and / or the tape needs rewinding. I've tried both these without any joy. In fact, I've got the dumpcycle set to 0 and the number of tapecycle set to 1. I'm under the impression that with this configuration Amanda will dump everything in my disklist daily. I've gone through the docs and tried this to see what's on the disk: amrestore -p /dev/nosst0 no-such-host /dev/null which returns this: amrestore:0: skipping start of tape: date 20030703 label nameon1 amrestore:1: skipping ..backup file1... Yet if I run it again I get this: amrestore: WARNING: not at start of tape file numbers will be offset amrestore:1: skipping ..backup file2... and so on. I've been doing some more experimenting / research and it looks like it's a blocksize issue. If I check the status of the drive it reports that the blocksize is 512 bytes. 'man amanda' (2.4.3) says the default block size is 32KB with the minimum being 32KB and the maximum being 32KB so I've arrived at the conclusion that this is something to do with the blocksize. I've attempted to change it with defsetblk to 32KB and re-ran the dump, yet when I attempt amrestore -p /dev/nosst0 no-such-host /dev/null it still always returns just one file, surely it should return all the files on the tape? I've seen from doing some searching that people have had similar problems with the default 512 byte block size which Onstream tape drivers ship with. I've experimented with tar to ensure the tape drive will backup and restore files which works. I don't really understand the what the block size means / does, there doesn't appear to be much info in the man pages / docs. I'd apprecate any help. Regards, James Williamson Name On The Net Ltd www.nameonthe.net I'm assuming that amrestore invoked in this manner will iterate through each file without having to be manually restarted. Does this mean when I use amrecover I must wind the tape to the correct place? Please excuse me if my knowledge of tape driver is woefully inadequate. Yes you have to rewind yourself. Before amanda 2.4.4, it was usually faster to rewind and then fsf manually to the correct position on tape, and than start amrecover. If you didn't position the tape, amanda had to read the tape up that position, which is much slower than fsf. Since amanda 2.4.4, amanda can do that for you, if you add the directive amrecover_do_fsf on in your amanda.conf. -- Paul Bijnens, XplanationTel +32 16 397.511 Technologielaan 21 bus 2, B-3001 Leuven, BELGIUMFax +32 16 397.512 http://www.xplanation.com/ email: [EMAIL PROTECTED] *** * I think I've got the hang of it now: exit, ^D, ^C, ^\, ^Z, ^Q, F6, * * quit, ZZ, :q, :q!, M-Z, ^X^C, logoff, logout, close, bye, /bye, * * stop, end, F3, ~., ^]c, +++ ATH, disconnect, halt, abort, hangup, * * PF4, F20, ^X^X, :D::D, KJOB, F14-f-e, F8-e, kill -1 $$, shutdown, * * kill -9 1, Alt-F4, Ctrl-Alt-Del, AltGr-NumLock, Stop-A, ...* * ... Are you sure? ... YES ... Phew ... I'm out * ***
Re: missing backup data
James Williamson wrote: I've been doing some more experimenting / research and it looks like it's a blocksize issue. If I check the status of the drive it reports that the blocksize is 512 bytes. That is a fixed blocksize (as opposed to variable blocksize). 'man amanda' (2.4.3) says the default block size is 32KB with the minimum being 32KB and the maximum being 32KB so I've arrived at the This means that amanda will do write()'s with a blocksize of 32k. If your tapedrive were set to variable blocksize, that would be blocksize on the tape. But you are using a fixed blocksize of 512 bytes. Now the kernel will reblock automatically the 32k amanda writes in 512 byte blocks on tape. I don't really understand the what the block size means / does, there doesn't appear to be much info in the man pages / docs. I'd apprecate any help. The nice thing about open software is that you can read the source, or, if that's too complicated (like for me), you can at least read the comments about in it. Read the comments in: /usr/src/linux/drivers/scsi/README.st I had to read it several times before I understood it, but it helped a lot. -- Paul Bijnens, XplanationTel +32 16 397.511 Technologielaan 21 bus 2, B-3001 Leuven, BELGIUMFax +32 16 397.512 http://www.xplanation.com/ email: [EMAIL PROTECTED] *** * I think I've got the hang of it now: exit, ^D, ^C, ^\, ^Z, ^Q, F6, * * quit, ZZ, :q, :q!, M-Z, ^X^C, logoff, logout, close, bye, /bye, * * stop, end, F3, ~., ^]c, +++ ATH, disconnect, halt, abort, hangup, * * PF4, F20, ^X^X, :D::D, KJOB, F14-f-e, F8-e, kill -1 $$, shutdown, * * kill -9 1, Alt-F4, Ctrl-Alt-Del, AltGr-NumLock, Stop-A, ...* * ... Are you sure? ... YES ... Phew ... I'm out * ***
missing backup data
Hi All, We've got Amanda set up to backup to a Onstream tape driver (SC-30), when we run amdump it reports that it's successully backed up the contents of the disklist. However, when it comes to try and restore something on the disklist using amrecover it reports this error: EOF, check amidxtapid.debug file on backup. amrecover: short block 0 bytes UNKNOWN file amrecover: Can't read file header I've gone through the docs and a likely cause seems to be the wrong tape's in the drive and / or the tape needs rewinding. I've tried both these without any joy. In fact, I've got the dumpcycle set to 0 and the number of tapecycle set to 1. I'm under the impression that with this configuration Amanda will dump everything in my disklist daily. I've gone through the docs and tried this to see what's on the disk: amrestore -p /dev/nosst0 no-such-host /dev/null which returns this: amrestore:0: skipping start of tape: date 20030703 label nameon1 amrestore:1: skipping ..backup file1... Yet if I run it again I get this: amrestore: WARNING: not at start of tape file numbers will be offset amrestore:1: skipping ..backup file2... and so on. I'm assuming that amrestore invoked in this manner will iterate through each file without having to be manually restarted. Does this mean when I use amrecover I must wind the tape to the correct place? Please excuse me if my knowledge of tape driver is woefully inadequate. I would appreciate anyone's help/thoughts. Regards, James Williamson Name On The Net Ltd www.nameonthe.net +44 208 7415453
