Hi, thank you very much for all the answers and ideas. We have found out that after the server was moved to different switch in the co-location centre the network interface and the switch auto-negotiated at the 10 Mbit Full Duplex mode. After setting it to GBit manually, everything seems to be working normally, but I am going to check it for some more time. SNMP connects to the local, isolated network from public and we have control over all the devices in the network.
I will post if anything new happens, but for now it seems this throughput limitation was causing those issues (although I am still wondering why there is nothing in log files, but due to network "overload" every service we were trying to connect to through network was not working any more). Thank you for your time. BR, Matej On Fri, Dec 17, 2010 at 2:48 PM, krad <kra...@gmail.com> wrote: > > > On 17 December 2010 13:47, krad <kra...@gmail.com> wrote: > >> >> >> On 16 December 2010 17:42, Matej Šerc <matej.s...@gmail.com> wrote: >> >>> Hi, >>> >>> I am experiencing a strange issue that has never occurred to me in all >>> the >>> years of using different versions of FreeBSD. >>> >>> One of our servers, which was running without any issues until yesterday, >>> stopped responding for two times now - yesterday and today. About three >>> days >>> ago another process of pulling out SNMP data from devices was added, but >>> I >>> was looking the system load and the system was working normally and also >>> processes were cmpleting successfully within the timeframe of 5 minutes >>> (much faster, they completed in about 2 minutes). I also want to mention >>> that those SNMP pulling processes were already working about a month or >>> so >>> on the same server (no hardware was changed in the meantime) and I am >>> pretty >>> sure that it should work normally as it did. >>> >>> My main problem is, that there is abcolutely nothing in log files - no >>> errors, no warnings, nothing. No strange messages, every process just >>> stops >>> logging at one time and then continues after the reboot. Another >>> interesting >>> issue is that both hangs occured at approximately the same time, but >>> there >>> was nobody in the server room and also no one was logged into the server >>> at >>> that time except me. About 10 minutes before hang I was investigating >>> processes and everything was very normal - no large CPU eating or memory >>> eating processes. This might be interesting, even after every process >>> stops >>> responding, I was still able to ping the network interfaces and receive >>> ICMP >>> replies back. >>> >>> Of course my idea about it is that it must be connected to some hardware >>> problems - my suggestion was to make some memory tests. But I would like >>> to >>> hear some your oppinions about the entire situation. Could some power >>> supply >>> issues be doing it? The server is about a year old and has, as I already >>> mentioned, worked like a charm until now. How come there is no kernel >>> panic >>> since no daemon seems to be working? Why is network interface still up >>> and >>> working? >>> >>> I was unable to go to the co-location facility so I can't say what was on >>> the screen at both times, but I suppose there was nothing else than >>> messages >>> I can read from log files. >>> >>> I know that 7.2 is pretty old version, but it was working until now on >>> the >>> same hardware and we had no reason to change that. Now the system is >>> after >>> reboot again running smoothly and without any issues at all. >>> >>> Thank you very much for any information regarding the issue. >>> >>> BR, Matej >>> _______________________________________________ >>> freebsd-questions@freebsd.org mailing list >>> http://lists.freebsd.org/mailman/listinfo/freebsd-questions >>> To unsubscribe, send any mail to " >>> freebsd-questions-unsubscr...@freebsd.org" >>> >> >> I'm not a huge fan of letting snmp spawn heavy weight scripts and >> processes as it is to easy for a remote machine to effectively dos the >> machine. I realise you are fairly sure the scripts arent an issue, but try >> croning them every 5 minutes, and writing the results to a file. SNMP can >> then simply retrieve the results from the file. This safeguard to to a >> certain extent, in that it stops many processes being spawned. All you have >> to watch after that is the job run time >> >> >> > Also lets stops resources being tied up on the monitoring machine, as it > doent have to hang around for x minutes for the results for its query > _______________________________________________ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "freebsd-questions-unsubscr...@freebsd.org"
