Dear all,
I have tracked down my performance problems to be mainly caused by
dns. I have tried pointing to various machines, one of them not
overloaded at all and also setting up a cache-only bind in the same
machine with no results. I am worried about the following traces (snoop
and tcpdump) run on the bind servers I was pointing to:
09:35:42.510000 belinda.rediris.es.domain > proxy.rediris.es.47059:
1710* q: pages.prodigy 1/2/2 . (154) (ttl 64, id 15931)
09:35:42.510000 proxy.rediris.es > belinda.rediris.es: icmp:
proxy.rediris.es udp port 47059 unreachable (DF) (ttl 255, id 43637)
--- proxy.rediris.es -> tais.rediris.es DNS C port=53883 tais.rediris.es -> proxy.rediris.es DNS R port=53833 proxy.rediris.es -> tais.rediris.es ICMP Destination unreachable (Bad port) they repeat very often (one out of ten requests) Any hint on what might be the cause of the ICMP unreachable one milisecond after requesting from that port ?? Unluckily enough from time to time I keep having the storeSwapOutFileOpened problem: (2) No such file or directory 1999/03/04 10:52:18| storeSwapOutFileOpened: Unable to open swapfile: /cache7®/00/01/0000010D this time /cache7 does not even exist neither is referenced anywhere. ok, I'll tell if I find out anything. Please report anyone having the same problems. Best regards, Javier Puche.Received on Thu Mar 04 1999 - 03:59:20 MST
This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 16:45:08 MST