>> For simplicity's sake I'd recommend using the same username everywhere, and
>> specifying that username at compile time. You may still need to set
>> permissions in a few places, but having one username minimizes this.
>>
>> And -- just checking -- you have a properly configured .amandahosts in the
>> home directory of your amanda user, right? If not, "man amanda" and search
>> on .amandahosts.
>
> I configured all host the same using amanda as user name and group name at
> compile time. I checked and rechecked. When I run amcheck on 11.00 server
> to connect to itself there is a debug entry in /tmp/amanda for amcheck but
> nothing for amandad. If I do this on a working machine there is an entry
> in /tmp/amanda for amcheck and anandad.
Yes, after running amcheck you should see at least four files in
/tmp/amanda: two from amandad, one from amcheck, and one from selfcheck.
It sounds like this is indeed some problem with amandad not running on the
11.00 box. From the packets below, it doesn't look like amandad ever gets
invoked.
I *really* don't think inetd is the problem, though; many, many other users
would have the same problem if this were true.
You might want to check that the amanda user has execute permission on all
amanda executables -- on my FreeBSD boxes, these are in /usr/local/sbin and
/usr/local/libexec/amanda, but they may be someplace else on HP/UX.
Sorry I couldn't be of further help.
dn
>
>>
>>>
>>>>
>>>>> I have amanda running on several servers running HP/UX 10.20 and one
>>>>> running HP/UX 11.11 with no problems but amcheck always fails with "Host
>>>>> down?" on the 11.00 server.
>>>>
>>>> On the tape server, try running "amcheck -m <configname>" and see what
>>>> errors it returns.
>>>>
>>>> I had a similar problem between two FreeBSD boxes. In my case, amcheck told
>>>> me that the tape user didn't have permissions set correctly on the amanda
>>>> client in various places, including /tmp, /usr/local/var/amanda, and of
>>>> course the filesystems I wanted to back up. The problem cleared once I
>>>> corrected this.
>>>>
>>>> Another thing to verify is that the client machine lists the tape server
>>>> (and user, if it's not the default called from inetd.conf) in its
>>>> .amandahosts file. I had a CNAME instead of the canonical FQDN, which may
>>>> also have been a problem.
>>>>
>>>> dn
>>>
>>> I've used nettl, HP's network tracing and logging app, to spy on the UDP
>>> traffic. On the 11.00 machine when I run amcheck I see 3 UDP packets from
>>> amcheck and that's all.
>>
>> what's in those three packets?
>>
>> dn
>
>
> vvvvvvvvvvvvvvvvvvvvvvvvvvvvvvARPA/9000
> NETWORKINGvvvvvvvvvvvvvvvvvvvvvvvvvv@#%
> Timestamp : Sat Jan 08 CST 2005 16:16:38.561022
> Process ID : 7319 Subsystem : NS_LS_UDP
> User ID ( UID ) : 800 Trace Kind : PDU OUT
> TRACE
> Device ID : -1 Path ID : 0
> Connection ID : 0
> Location : 00123
> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> Transmitted 145 bytes via UDP Sat Jan 08 16:16:38.056102 CST 2005
> pid=7319
> 0: 02 21 30 39 00 7d 00 7d 41 6d 61 6e 64 61 20 32 .!09.}.}Amanda 2
> 16: 2e 34 20 52 45 51 20 48 41 4e 44 4c 45 20 30 30 .4 REQ HANDLE 00
> 32: 30 2d 34 30 30 31 44 45 45 30 20 53 45 51 20 31 0-4001DEE0 SEQ 1
> 48: 31 30 35 32 32 32 35 39 38 0a 53 45 43 55 52 49 105222598.SECURI
> 64: 54 59 20 55 53 45 52 20 61 6d 61 6e 64 61 0a 53 TY USER amanda.S
> 80: 45 52 56 49 43 45 20 6e 6f 6f 70 0a 4f 50 54 49 ERVICE noop.OPTI
> 96: 4f 4e 53 20 66 65 61 74 75 72 65 73 3d 66 66 66 ONS features=fff
> 112: 66 66 65 66 66 39 66 66 65 30 66 3b 0a -- -- -- ffeff9ffe0f;....
>
>
> vvvvvvvvvvvvvvvvvvvvvvvvvvvvvvARPA/9000
> NETWORKINGvvvvvvvvvvvvvvvvvvvvvvvvvv@#%
> Timestamp : Sat Jan 08 CST 2005 16:16:48.552955
> Process ID : 7319 Subsystem : NS_LS_UDP
> User ID ( UID ) : 800 Trace Kind : PDU OUT
> TRACE
> Device ID : -1 Path ID : 0
> Connection ID : 0
> Location : 00123
> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> Transmitted 145 bytes via UDP Sat Jan 08 16:16:48.055295 CST 2005
> pid=7319
> 0: 02 21 30 39 00 7d 00 7d 41 6d 61 6e 64 61 20 32 .!09.}.}Amanda 2
> 16: 2e 34 20 52 45 51 20 48 41 4e 44 4c 45 20 30 30 .4 REQ HANDLE 00
> 32: 30 2d 34 30 30 31 44 45 45 30 20 53 45 51 20 31 0-4001DEE0 SEQ 1
> 48: 31 30 35 32 32 32 35 39 38 0a 53 45 43 55 52 49 105222598.SECURI
> 64: 54 59 20 55 53 45 52 20 61 6d 61 6e 64 61 0a 53 TY USER amanda.S
> 80: 45 52 56 49 43 45 20 6e 6f 6f 70 0a 4f 50 54 49 ERVICE noop.OPTI
> 96: 4f 4e 53 20 66 65 61 74 75 72 65 73 3d 66 66 66 ONS features=fff
> 112: 66 66 65 66 66 39 66 66 65 30 66 3b 0a -- -- -- ffeff9ffe0f;....
>
>
> vvvvvvvvvvvvvvvvvvvvvvvvvvvvvvARPA/9000
> NETWORKINGvvvvvvvvvvvvvvvvvvvvvvvvvv@#%
> Timestamp : Sat Jan 08 CST 2005 16:16:58.552938
> Process ID : 7319 Subsystem : NS_LS_UDP
> User ID ( UID ) : 800 Trace Kind : PDU OUT
> TRACE
> Device ID : -1 Path ID : 0
> Connection ID : 0
> Location : 00123
> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> Transmitted 145 bytes via UDP Sat Jan 08 16:16:58.055293 CST 2005
> pid=7319
> 0: 02 21 30 39 00 7d 00 7d 41 6d 61 6e 64 61 20 32 .!09.}.}Amanda 2
> 16: 2e 34 20 52 45 51 20 48 41 4e 44 4c 45 20 30 30 .4 REQ HANDLE 00
> 32: 30 2d 34 30 30 31 44 45 45 30 20 53 45 51 20 31 0-4001DEE0 SEQ 1
> 48: 31 30 35 32 32 32 35 39 38 0a 53 45 43 55 52 49 105222598.SECURI
> 64: 54 59 20 55 53 45 52 20 61 6d 61 6e 64 61 0a 53 TY USER amanda.S
> 80: 45 52 56 49 43 45 20 6e 6f 6f 70 0a 4f 50 54 49 ERVICE noop.OPTI
> 96: 4f 4e 53 20 66 65 61 74 75 72 65 73 3d 66 66 66 ONS features=fff
> 112: 66 66 65 66 66 39 66 66 65 30 66 3b 0a -- -- -- ffeff9ffe0f;....
>
>
>
>
>>
>>
>> On one of the machines that are OK if I perform the
>>> same procedure I see packets from amcheck and reply packets from amandad.
>>> If I run netstat -a | grep amanda it shows the amanda service listening.
>>> It just never gets launched. I'm a bit stumped.
>>>
>>> Thanks
>>> Richard Ray
>>>
>>>>
>>>>
>>>> I'm trying to determine if the problem is with
>>>>> inetd or amanda. I feel certain it's with HP but I don't know where else
>>>>> to look. I'm running amanda 2.4.4p4.
>>>>>
>>>>> Thanks
>>>>> Richard Ray
>>>>>
>>>>
>>>>
>>>
>>
>>
>
|
