Re: hdup not exiting on remote clients?

Robert Haskins <rhaskins-X6DOoJyBSnRWk0Htik3J/[email protected]> Wed, 02 Feb 2005 17:43:26 -0500
Newsgroups gmane.comp.sysutils.backup.hdup.general
Message-ID <[email protected]>
I'm not sure, but I think I have two similar problems. On some clients, 
hdup leaves only an ssh session hanging around. I believe this is a 
separate and distinct issue from the "hdup never exiting" problem. Here 
is what I have on a client that only has ssh sessions that never seem to 
die:

[rhaskins@efrat rhaskins]$ ps -ef | grep hdup
root     14931     1  0  2004 ?        00:00:00 /usr/bin/ssh 
[email protected] /usr/sbin/hdup remote efrat
root     26615     1  0 Jan14 ?        00:00:00 /usr/bin/ssh 
[email protected] /usr/sbin/hdup remote efrat

I'm not sure if I ever got a backup from this host on this day, as we 
recently began a policy of purging our daily backups so I can't tell if 
this backup was successful or not.

I'm still watching out for the hdup never exiting issue too.

Thanks!

Robert Haskins wrote:
> Sorry, I neglected to say these are Fedora Core 1/2 boxes, running stock 
> Openssh. "rpm -qa" reports:
> 
> [root@server cron.daily]# rpm -qa | grep -i openssh
> openssh-server-3.6.1p2-34
> openssh-clients-3.6.1p2-34
> openssh-3.6.1p2-34
> 
> I will run an strace the next time I see this problem. I'd offer to run 
> a debug version, but with ~25 boxes and having to re-install a special 
> debug version I'd have to think long and hard about doing that 
> especially where this is essentially non-reproduceable. However, a new 
> hdup version with a command line option would be worth doing.
> 
> I'll keep you posted. Thanks!
> 
> Miek Gieben wrote:
> 
>> [On 01 Feb, @ 21:33, Robert wrote in "[hdup-user] hdup not exiting o 
>> ..."]
>>
>>> In the past 30 days or so, we've noticed 3 instances where remote 
>>> hdup clients never exit and take up cpu time. However, backups 
>>> complete successfully.
>>
>>
>>
>> Someone reported that hdup sometimes crashes when used in conjunction
>> with other ssh implementations (like the original ssh). I have only
>> access to openssh so I can't and couldn't test/debug this. It might be
>> related, but it doesn't really help you :P
>>
>>
>>> I haven't been able to nail down any exact circumstances around what 
>>> might be causing this. It seems to always happen to clients running 
>>> in remote data centers, though that may be because we have only 3 
>>> machines (out of ~25) local to the hdup server. We are running 
>>> hdup-2.0.5 everywhere.
>>
>>
>>
>> can strace such a running hdup? And send me the trace when this
>> happens?
>>
>>
>>> Has anyone seen this issue? Any ideas on what might be causing this? 
>>> Thanks for any help you can provide.
>>
>>
>>
>> No clue :( I do plan to release a 2.0.6 with some further fixes,
>> alltough nothing in the remote backup code. Also a -D (for debugging)
>> switch is added to give debugging information without the need to 
>> recompile hdup. This might help in getting to the bottom of this.
>>
>> -- 
>> grtz,
>>   - Miek
>>
>> http://www.miek.nl                   http://www.nlnetlabs.nl
>>
>