We have an NFS server and our custom board which is our NFS Client. We are generating a log file as part of our custom application. This log file is getting generated in the root file system of the board(NFS client) immediately after the application execution. But however its taking about 30 seconds to get updated on the NFS server which is a high delay in our use case. In the NFS server I ran the command tail -f log.txt and only after 30 seconds the log is updated in the log file. Is there a way we can eliminate this delay. I performed an analysis on the NFS port using the tcpdump and observed that the packets are arriving at the NFS Server side after this 30 s delay.
Asked
Active
Viewed 688 times
0
-
Sounds like a write-back behavior on the client side. Which os your client runs? – kofemann Sep 02 '21 at 08:11
-
Client runs on UBUNTU 18.04.2. – Aparna Sep 03 '21 at 12:25
-
try to lower the `vm.dirty_writeback_centisecs` with sysctl https://www.kernel.org/doc/html/latest/admin-guide/sysctl/vm.html#dirty-writeback-centisecs – kofemann Sep 03 '21 at 12:46