Strange file read problem 2.4.18 SuSE
Subsystems (2)
ALright guys, we've been troubleshooting this problem at the office, and now
I need a little developer help with this. lol First, the setup is 2 SuSE SLES 7 machines running SuSE's patched 2.4.18 kernel. We have an app which listens on a socket and accepts file through a proprietary protocol. It writes the file to the NFS server, then it calls an import thread which then imports the file into an Oracle database. The app is written purely in java, so the improt is a seperate class that is instantiated and handed it's own thread. Now, this works fine the vast majority of the time, but occasionally, when dealing with a 10+M file, the import will barf reading the file (it's a line-by-line read, so it's not like a copy, which block-by-blocks as fast as possible), with the error message "No such file or directory". ANyone can ls and see that indeed, the file does exist, and the next time the import runs (after a new file comes down, the importer will process all pending files before grabbing the new one), it'll import the 10M file just fine (problem being that part of the file was already imported, so data can get duplicated).
I was able to replicate an error similar to this one under different
conditions: when the cluster that advertises teh NFS export to clients mvoes from one node to another, then reads can fail with the same error message, at which point a file close and open solves things. Howver, in this case the cluster isn't moving, so I'm kinda at a loss.
SO, questions are:
What kinds of situations can lead to this error being given by the client,
and:
what can I do to troubleshoot?
Thanks!
Donavan Pantke
-------------------------------------------------------
Enterprise Linux Forum Conference & Expo, June 4-6, 2003, Santa Clara The only event dedicated to issues related to Linux enterprise solutions www.enterpriselinuxforum.com
_______________________________________________
NFS maillist - [email protected]
https://lists.sourceforge.net/lists/listinfo/nfs