pingu Posted February 9, 2004 Report Share Posted February 9, 2004 We are having a very unusal problem since upgrading to Server version 6. Retro is running on it's own machine (OS 10.3.2 client) with an Asante 1000TA NIC and an ATTO UL3S-66 (all drivers and firmware up to date). Our backup server scripts seem to run just fine for short backups from clients and servers, but we have one very large (300+ Gig) subvolume on an Xserve/Xraid server that begins backing up okay but after 5-20Gb reports an error 519 (network communication failed). At this point we loose all tcp/ip connection to the Retro server (ARD, ping, ssh etc). Going down to the retro machine we usually find it alive and responsive, but all communications on the Asante NIC are dead (internal interface is not plugged in), we can't ping out from it or access any other network services. I cannot for the life of me figure out how Retrospect could be shutting down an entire interface. No logs show any indication of a problem except the retro log (showing the 519) Can anyone shed any light on this? Cheers Dan Link to comment Share on other sites More sharing options...
natew Posted February 10, 2004 Report Share Posted February 10, 2004 Hi Can you try running the backup via file sharing rather than the client? If that fails too we can assume there are other problems at work as well. The current driver for the Asante nic is 2.4. What are you running? Any single files over 2GB in size? Nate Link to comment Share on other sites More sharing options...
CallMeDave Posted February 11, 2004 Report Share Posted February 11, 2004 Quote: Going down to the retro machine we usually find it alive and responsive, but all communications on the Asante NIC are dead (internal interface is not plugged in), we can't ping out from it or access any other network services. I'd be curious to see what "ifconfig -a" shows when entered in Terminal (with some reference to know which interface is the NIC). Dave Link to comment Share on other sites More sharing options...
pingu Posted February 16, 2004 Author Report Share Posted February 16, 2004 Thanks for input there guys. Strangely enought though, the problem seemed to disappear as mysteriously as it appeared. I'll post if I find out any more. Still bit worrying anyway Cheers Dan Link to comment Share on other sites More sharing options...
pingu Posted March 9, 2004 Author Report Share Posted March 9, 2004 Damn! Spoke too soon... the problem has returned. I have configured a second IP on the built in interface to see whether it manages to kill both if's. I have also set the client speed threshold to a specific value rather than zero to see if this helps and set prefs to log all SCSI activity just in case this is something to do with it (doubt it somehow). We did attempt to run a backup via file-sharing, but due (i thinks) to the permissions on the volume the script reported that no files were selected for backup, so that was a bit of a non starter. As for the Asante driver, yes we are running the latest version (2.4) We did track the problem with some machines down to the 4.3 version of the client, but this doesn't apply to the server which is 10.3.2 Any further input much appreciated; this is driving me nuts! cheers Dan Link to comment Share on other sites More sharing options...
pingu Posted April 5, 2004 Author Report Share Posted April 5, 2004 So we have finally tracked down the cause of this problem. It seems that the Asanté GigaNix 1000TA drivers (v2.4) are not compatible with Panther, only 10.2. As soon as we removed the Gigabit card and started using a 100BT card, problem disappeared. What is really odd, is that the card seems to operate fine on a day-to-day basis, it only seems to fail during long sustained data transfers. Is anyone else using an Asanté Gig card under 10.3? Any workarounds? Link to comment Share on other sites More sharing options...
Recommended Posts
Archived
This topic is now archived and is closed to further replies.