Re: BIND module doesn't create secondary zones remotely

[Jamie C. <jca...@we...>]

The octal 4000 part is supposed to indicate a directory, which is why the total
mask should be 4755 . It's very odd that maketemp.pl fails, because the
same code is used in web-lib.pl in the tempname function. On own solaris 7 box,
I wasn't able to replicate this problem..

  - Jamie

G.J. Moed wrote:

> Okay, deleting the files didn't help...
> 
> Checking out the setup.sh, it refers to 'maketemp.pl' so I took a look at that:
> 
> if the dir doesn't exist, it creates it, sets owner and group, and sets 755.
> if it exists, it checks owner and group, checks for 4000 and then for 755, returns 0 if all goes well, and otherwize 1.
> 
> I'm wondering about the 4755 when it creates with 755 and of course my /tmp/.webmin exists and has 755 and not 4755 and hence the check for 4000 fails and thus returns 1 to the setup.sh.
> 
> Simply deleting the tmp/.webmin dir will allow me to upgrade I figure, but again, maybe you may have wanted to know about this....
> 
> Best regards,
> 
> G.J.
> 
> 
> On 8/29/2002 12:44 PM, G.J. Moed <gj...@xe...> wrote:
> 
>>I was hoping for a bug cuz that's something we can resolve ;-)
>>
>>Trying to update the slave first now, to 0.990, and now we stumble 
>>upon more funny stuff:
>>
>>during setup, after hitting return for the location to install into:
>>
>>ERROR: Failed to create or check temp files directory /tmp/.webmin
>>
>>Looking at the /tmp/.webmin directory I find a bunch of rpc.cgi files 
>>with hash structures in them (looks very much like what I captured 
>>with snoop), probably had to do with the bug then? Just thought you 
>>may wanted to know about them. I'll just remove them and hopefully 
>>that'll allow me to upgrade without any probs...
>>
>>Thanks for your quick help...
>>
>>Best regards,
>>
>>G.J.
>>
>>
>>G.J. Moed
>>XenoSite
>>
>>
>>On 8/29/2002 12:00 PM, Jamie Cameron <jca...@we...> wrote:
>>
>>>G.J. Moed wrote:
>>>
>>>
>>>>Hi All,
>>>>
>>>>stumbling upon the following problem:
>>>>
>>>>Using webmin 0.970 on a cluster of Solaris8 x86 machines, setup 
>>>>
>>>everything correctly such as allowing RPC etc. etc.
>>>
>>>>I can login to one machine and get to any other machine from there. 
>>>>
>>>I can also login to the machine with the secondary DNS and get back 
>>>
>>>to the primary.
>>>
>>>>I can create master zones on the primary just fine, slaves I can 
>>>>
>>>do on the secondary. Nothing seems wrong.
>>>
>>>>When I create a master and ask to also create the slave on the 
>>>>
>>>secondary, it takes a while and I get the screen back where I can 
>>>
>>>edit all records and such. Seems everything went fine. However, 
>>>
>>looking 
>>
>>>at the secondary, nothing got added to the named.conf.
>>>
>>>>Turned on logging for everything on both machines and tried numerous 
>>>>
>>>times. Also did some packet captures with snoop.
>>>
>>>>This is what I found but I'm still puzzled:
>>>>
>>>>Primary: logging shows me accessing the module and it reports 
>>>>
>>the 
>>
>>>actions of creating a master zone.
>>>
>>>>Secondary: logging shows the primary accessing the rpc.cgi but 
>>>>
>>>doesn't show any actions taken after that, which it probably should?
>>>
>>>>This is all that gets logged on the secondary:
>>>>.... "POST /rpc.cgi HTTP/1.0" 200 265
>>>>.... "POST /rpc.cgi HTTP/1.0" 200 2242154
>>>>.... "POST /rpc.cgi HTTP/1.0" 200 0
>>>>.... "POST /rpc.cgi HTTP/1.0" 200 0
>>>>
>>>>I find the last 2 lines kinda strange and maybe that's the problem.
>>>>
>>>>What am I doing wrong or am I stumbling upon some bug?
>>>>
>>>
>>>I seem to recalling finding a bug related to secondary zone creation 
>>>
>>>not long ago, so it may be
>>>fixed only in the latest release. Have you tried installing version 
>>>
>>>0.990 on both the master and
>>>slave servers? It is quite easy to upgrade from within the Webmin 
>>>
>>>Configuration module ..
>>>
>>>Those "POST /rpc.cgi HTTP/1.0" 200 0 lines strongly suggest that 
>>>
>>>something is going wrong
>>>on the slave server side ..
>>>
>>> - Jamie