[Rear-users] Relax and Recover 1.7.18 is released

Werner Flamme werner.flamme at ufz.de
Thu Apr 23 17:13:44 CEST 2009


Hi Schlomo,

no, I did not think that bash was generally unable to hold bigger
variables ;-) I thought there might be something like ulimit -
restricting sizes of variables. BTW, ulimit -a shows max memory size
(-m) 20933885 on the box where ReaR does not run, and 10071990 on a box
where ReaR works fine. So ulimit may not be the problem.

Tracing - yes, it will be that. I will find the part of the code where
COPY_AS_IS is copied and look what happens to the array elements there.
/etc/ssh is in this array, so I only have to see why it isn't copied.

Regards,
Werner

Schlomo Schapiro [23.04.2009 12:01]:
> Hi Werner,
> 
> man bash:
>    Arrays
>        Bash provides one-dimensional array variables.  Any variable may
> be used as an array; the declare  builtin
>        will  explicitly declare an array.  There is no maximum limit on
> the size of an array, nor any requirement
>        that members be indexed or assigned contiguously.  Arrays are
> indexed using integers and are zero-based.
> 
> I did some tests which make it seem that bash itself has no problem with
> very large command lines:
> $ wc <<<$(seq 1 300000)
>       1  300000 1988895
> $ wc <<<"$(seq 1 300000)"
>  300000  300000 1988895
> 
> (replace wc with cat to see what is going on :-) )
> 
> I am afraid you will have to trace rear on both machines and compare in
> detail what is the difference and why. This is something which is really
> hard to diagnose remotely.
> 
> Schlomo
> 
> Werner Flamme wrote:
>> Schlomo Schapiro [22.04.2009 18:24]:
>>> Hi Werner,
>>>
>>> so did we change something or did the issue disappear "magically"?
>> Well, the problem was the sshd, and the file that caused the problem -
>> /usr/share/rear/build/default/50_patch_sshd_config.sh - looks modified
>> ;-) So the script does not "hang" any more.
>>
>> I think the magic was you :-)
>>
>>> Nice to hear that you are happy.
>> Yes, I am happy that "rear mkrescue" succeeds. But I'm not too glad...
>>
>> On one box (after "rear -d -D mkrescue") I enter "ls -l rootfs/etc" and get:
>> ---snip---
>> drwxr-xr-x   8 root root   4096 Apr 23 09:19 .
>> drwxr-xr-x  14 root root   4096 Apr 23 09:19 ..
>> -rw-r--r--   1 root root      7 Apr 23 09:19 HOSTNAME
>> lrwxrwxrwx   1 root root      7 Apr 23 09:19 bash.bashrc -> profile
>> -rw-r--r--   1 root root 109067 Apr 23 09:19 dumpkeys.out
>> -rw-r--r--   1 root root      1 Mar 14 21:24 ethers
>> drwxr-xr-x   2 root root   4096 Apr 22 16:16 event.d
>> -rw-r--r--   1 root root    316 Mar 14 21:24 fstab
>> -rw-r--r--   1 root root     19 Mar 14 21:24 group
>> -rw-r--r--   1 root root    370 Jun 30  2004 host.conf
>> -rw-r--r--   1 root root   1550 Apr  9  2008 hosts
>> -rw-r--r--   1 root root    567 Apr 23 09:19 inittab
>> drwxr-xr-x   2 root root   4096 Apr  5  2007 iproute2
>> -rw-r--r--   1 root root    459 Apr 23 09:19 issue
>> -rw-r--r--   1 root root   3555 Apr 23 09:19 ld.so.cache
>> -rw-r--r--   1 root root     12 Mar 14 21:24 ld.so.conf
>> -rw-r--r--   1 root root     63 Mar 14 21:24 lilo.conf
>> -rw-r--r--   1 root root  12645 Nov 17  2005 modprobe.conf
>> -rw-r--r--   1 root root    115 Feb 19  2008 modprobe.conf.local
>> drwxr-xr-x   2 root root   4096 Apr  6 11:57 modprobe.d
>> -rw-r--r--   1 root root     50 Apr 23 09:19 modules
>> -rw-r--r--   1 root root 111892 Mar 14 21:24 moduli
>> -rw-r--r--   1 root root     74 Mar 14 21:24 motd
>> -rw-r--r--   1 root root      0 Apr 23 09:19 mtab
>> -rw-r--r--   1 root root    287 Apr 23 09:19 network.sh
>> -rw-r--r--   1 root root    160 Mar 14 21:24 nsswitch.conf
>> -rw-r--r--   1 root root   2068 Dec 12  2007 ntp.conf
>> -rw-r--r--   1 root root     81 Apr 23 09:19 passwd
>> -rw-r--r--   1 root root    754 Mar 14 21:24 profile
>> -rw-r--r--   1 root root   6561 Jun 30  2004 protocols
>> drwxr-xr-x   3 root root   4096 Apr 23 09:19 rear
>> -rw-r--r--   1 root root    117 Jun  7  2007 resolv.conf
>> -rw-r--r--   1 root root   1615 Oct 19  2007 rpc
>> drwxr-xr-x   2 root root   4096 Apr 23 09:19 scripts
>> -rw-r--r--   1 root root 596871 Oct  9  2007 services
>> drwxr-xr-x   2 root root   4096 Apr 23 09:19 ssh
>> -rw-r--r--   1 root root    319 Mar 14 21:24 syslog-ng.conf
>> -rw-r--r--   1 root root     69 Mar 14 21:24 syslog.conf
>> lrwxrwxrwx   1 root root     23 Apr 23 09:19 termcap ->
>> /usr/share/misc/termcap
>> ---pins---
>>
>> on the problem box, I only see:
>> ---snip---
>> total 320
>> drwxr-xr-x   5 root root   4096 Apr 23 09:10 .
>> drwxr-xr-x  14 root root   4096 Apr 23 09:10 ..
>> -rw-r--r--   1 root root      7 Apr 23 09:10 HOSTNAME
>> lrwxrwxrwx   1 root root      7 Apr 23 09:10 bash.bashrc -> profile
>> -rw-r--r--   1 root root 109076 Apr 23 09:10 dumpkeys.out
>> -rw-r--r--   1 root root      1 Mar 14 21:24 ethers
>> drwxr-xr-x   2 root root   4096 Apr 22 16:14 event.d
>> -rw-r--r--   1 root root    316 Mar 14 21:24 fstab
>> -rw-r--r--   1 root root     19 Mar 14 21:24 group
>> -rw-r--r--   1 root root     26 Mar 14 21:24 hosts
>> -rw-r--r--   1 root root    567 Apr 23 09:10 inittab
>> -rw-r--r--   1 root root    459 Apr 23 09:10 issue
>> -rw-r--r--   1 root root   3569 Apr 23 09:10 ld.so.cache
>> -rw-r--r--   1 root root     12 Mar 14 21:24 ld.so.conf
>> -rw-r--r--   1 root root     63 Mar 14 21:24 lilo.conf
>> -rw-r--r--   1 root root     27 Apr 23 09:10 modules
>> -rw-r--r--   1 root root 111892 Mar 14 21:24 moduli
>> -rw-r--r--   1 root root     74 Mar 14 21:24 motd
>> -rw-r--r--   1 root root      0 Apr 23 09:10 mtab
>> -rw-r--r--   1 root root    290 Apr 23 09:10 network.sh
>> -rw-r--r--   1 root root    160 Mar 14 21:24 nsswitch.conf
>> -rw-r--r--   1 root root     81 Apr 23 09:10 passwd
>> -rw-r--r--   1 root root    754 Mar 14 21:24 profile
>> drwxr-xr-x   2 root root   4096 Apr 23 09:10 rear
>> drwxr-xr-x   2 root root   4096 Apr 23 09:10 scripts
>> -rw-r--r--   1 root root    319 Mar 14 21:24 syslog-ng.conf
>> -rw-r--r--   1 root root     69 Mar 14 21:24 syslog.conf
>> ---pins---
>>
>> This is not very equal :-(, and among others the ssh subdir is not
>> copied on the problem box. OK, it's only ;-) the box with the productive
>> SAP ERP on it, and it has KVM-over-IP, but in the case of a recovery I
>> may have a problem... I cannot test it since I don't have similar
>> hardware at hand, and restoring to different hardware (or a virtual
>> environment) may fail because of missing drivers or missing hardware (FC
>> cards).
>>
>> All 3 boxes are SLES9 SP4 freshly patched, and there shouldn't be any
>> different configuration (except of the running SAP software, due to the
>> different hardware). But in the logfile I find
>> ++ Log 'Warning: ssh configuration files missing'
>> In the '20_copy_as_is.sh' part of the log, /etc/ssh is in the list of
>> files and directories. It looks as if the string is too long since
>> /etc/modprobe.conf and other files mentioned there (near the end) aren't
>> present, either. But I did not find info about a maximum string length
>> in bash.
>>
>> Regards,
>> Werner
>>
>>> Werner Flamme wrote:
>>>> Schlomo Schapiro [15.03.2009 21:20]:
>>>>> Hi,
>>>>>
>>>>> we just released Relax and Recover 1.7.18.
>>>>>
>>>>> It is a minor bugfix/compatibility release without any new feature:
>>>>>
>>>>> * openSUSE 11.1 supported/tested (and possibly other new distros using rpcbind
>>>>> instead of portmap)
>>>>> * SATA works now with openSUSE (and probably other distros as well), we did not
>>>>> load the drivers/ata modules :-(
>>>> Just a little notice: all of my SLES 9 machines now run "rear mkrescue"
>>>> with ISO output without any problem. The sshd_config problem I had in
>>>> January seems to be solved (for now...)
>>>>
>>>> Best regards, and many thanks for this software!





More information about the rear-users mailing list