[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [HTCondor-users] Collector not reachable on localhost?

Hi all,

just to close this issue : scraping and respawning the collector 'fixed'
the problem and the daemons can communicate with each other again.
No idea what the actual problem was, maybe some networking problem
within OpenStack...?


On 2017-06-30 16:54, Thomas Hartmann wrote:
> Hi all,
> I am struggling somewhat to spawn a test cluster (v8.6.4 on OpenStack,
> IPv4 only) where a Master, Collector, Negotiator and Scheduler running
> on one host (plus a few workers).
> Thing seems to be, that the collector cannot connect to itself(?). The
> master is restarting the collector several times but cannot connect to
> it [1].
> As for the Collector log, the DemonCore also complaints about not being
> able to connect to the Collector (on the same host/IP).
> Network-wise I was able to communicate via ncat between a worker and the
> collector host on the shared port 9618 (& 9620). And -AFAIS- the
> collector is actually listening on 9618 [5].
> (Sched and Negotiator seem to be happy  and are listening to the
> DemonCore on p9620)
> Maybe somebody has an idea what could be jamming the collector?
> (being bound also to IPv6 link-local should be no problem, or??)
> Cheers and thanks for ideas,
>   Thomas
> btw: is it actually necessary to set POOL_HISTORY_DIR [6] ~~>
> /var/ViewHist/ ? I had to create the directory manually but I do not
> remember that setting explicitly the dir had been necessary it before?

Attachment: smime.p7s
Description: S/MIME Cryptographic Signature