fbackup Misery

[Edward Wynn]

Hi,

I am trying to get fbackup working with an external SCSI DDS4 DAT drive - the drive works fine with cpio but I cannot get it to work with fbackup - see below:

fbackup -0 -g /var/adm/fbackupfiles/graphs/weekly -f /dev/rmt/0m
fbackup(1004): session begins on Wed Jan 3 15:55:33 2001
fbackup(3203): volume 1 has been used -1 time(s)
fbackup(3024): writing volume 1 to the output file /dev/rmt/0m
fbackup(3044): ioctl error, can't query outfile
at media record 12
fbackup(3102): attempting to make this volume salvagable
fbackup(3123): could not read the previous checkpoint record
fbackup(3105): writing 2 EOFs and rewinding the tape
fbackup(3106): please mount a good tape
fbackup(3316): enter '^[yY]' when volume 1 is ready on /dev/rmt/0m,
or '^[nN]' to discontinue:

I have tried this same command but using device /dev/rmt/c0todoBEST but this does not help - I have also tried a range of different tapes - from DDS2 - DDS4 capacities used and new tapes.

Interestingly though if fbackup fails as above and I do the following something gets written to the tape (that I can verify with frecover)

fbackup(3316): enter '^[yY]' when volume 1 is ready on /dev/rmt/0m,
or '^[nN]' to discontinue:
n
fbackup(3019): would you like to enter a new output file?
y
fbackup(3020): please enter the new file's name:
c0t0d0BEST
fbackup(3024): writing volume 1 to the output file c0t0d0BEST
fbackup(1301): resuming at file 4
fbackup(3055): total file blocks read for backup: 33268
fbackup(3056): total blocks written to output file c0t0d0BEST: 39293

Has anyone any ideas what I am doing wrong please ?

Many thanks in advance, Eddie

[Alex Glennie]

I'd of said this was a tape or h/w error except cpio works ... next best culprit is patches wrt fbackup,frecover,SCSI Pass-thru Pseudo
Drivers etc :

fyi : a similar problem at 10.20 was fixed by installing :

[PHCO_14990/PACHRDME/English] : s700_800 10.20 LVM commands cumulative
patch
[PHKL_15200/PACHRDME/English] : s800 10.20
LVM/UFS/NDDB/flkmgr/VxFS/PCI cumulative patch
[PHCO_14842/PACHRDME/English] : s700_800 10.20 fsck_vxfs(1M)
cumulative patch
[PHNE_14770/PACHRDME/English] : s800 10.20 cumulative ARPA Transport
patch
[PHKL_12340/PACHRDME/English] : s800 10.20 VxFS (JFS) fsck(1M)
cumulative patch
[PHCO_14176/PACHRDME/English] : s700_800 10.20
fbackup(1M)/frecover(1M) cumulative patch
[PHKL_13044/PACHRDME/English] : s800 10.20 SCSI Pass-thru Pseudo
Driver cumulative patch
[PHNE_13416/PACHRDME/English] : s800 10.20 HP-PB 100 BaseT LAN driver
patch
[PHSS_14405/PACHRDME/English] : s800 10.20 Cumulative Online
Diagnostics Tools Patch
[PHCO_14888/PACHRDME/English] : s700_800 10.20 Cumulative SAM/ObAM
Patch
[PHCO_14176/PACHRDME/English] : s700_800 10.20
fbackup(1M)/frecover(1M) cumulative patch
[PHKL_13044/PACHRDME/


NB this is an old list

[Edward Wynn]

Thanks for the note on patches - on reviewing them I am not sure what to do as they are all for 10.20 so I guess they should all be in HP-UX 11 ? If not where can I get a list of equivalent patches for HP-UX 11 from ?

There is also some more information (I used the -v switch on fbackup) and you can now see that it does actually write some files to tape before it gets the ioctl error - which seems to happen at a random point in the backup.

351: /home/malcolmh/.netscape/xover-cache/host-news/hostinfo.dat 1
352: /home/malcolmh/.profile 1
353: /home/malcolmh/.sh_history 7
fbackup(3044): ioctl error, can't query outfile
at media record 137
fbackup(3102): attempting to make this volume salvagable
fbackup(3123): could not read the previous checkpoint record
fbackup(3105): writing 2 EOFs and rewinding the tape
fbackup(3106): please mount a good tape
fbackup(3316): enter '^[yY]' when volume 1 is ready on /dev/rmt/0m,
or '^[nN]' to discontinue:

Has anyone any further pointers ?

Thanks, Eddie

[Darrel Louis]

Hi,

When using a new tape is this accuring aswell?

Check what the maxdsiz kernel parameter is set to, if this is the default try increasing this, triple it.

Check man - 5 ioctl
For writes, non-blocking I/O prevents all write requests to that device file from blocking, whether the requests succeed or fail. Such a write request completes in one of three ways:

+ If there is enough space available in the system to buffer all the data, the write completes successfully,having written out all of the data, and returns the number of bytes written;

+ If there is not enough space in the buffer to write out the entire request, the write completes successfully,
having written as much data as possible, and returns the number of bytes it was able to write;
+ If there is no space in the buffer, the write fails and errno is set to EWOULDBLOCK.

Good Luck

[Edward Wynn]

Thanks for the replies to date - unfortunately none has solved the problem. Whilst at the same time the problem has moved on - cpio will not work on large saves but it will on small saves - I intend to start a new thread with this problem as the heading of this one is no longer appropriate.