Fixed
Created: Apr 15, 2019
Updated: May 16, 2019
Resolved Date: Apr 19, 2019
Found In Version: 8.0.0.29
Fix Version: 8.0.0.30
Severity: Severe
Applicable for: Wind River Linux 8
Component/s: Kernel
While using NFSv4 to save some logs, NFSv4 hangs during operation.
/etc/exports on server:
/export/Core 192.168.122.0/255.255.255.0(sync,rw,no_root_squash,no_all_squash,no_wdelay,no_subtree_check)
/etc/fstab on client:
master:/export/Core /Core nfs4 async,intr,noatime,nodev,port=2049,timeo=5 0 0
On client run this sequence:
1) Start script once-a-second.py that opens a file on nfs share and write a value each second to file /Core/log/once-a-second.txt
2) Reboot server (simulates the fail over that customer have)
3) Wait until grace period has finished (90 seconds default)
4) Start 'ndWorker-mimic.py /Core/log /Core/res' - this script will read all files from /Core/log, compress them and write to /Core/res
5) Wait a couple of seconds to let once-a-second.py time to write to file
6) Run 'ndWorker-mimic.py /Core/log /Core/res' a second time and the hung task occurs
I have created a script test-seq.sh that runs the sequence automatically. It have hard-coded IP addresses in it and assumes that client can run command through ssh without password. Scripts ndWorker-mimic.py and once-a-second.py are also attached to this ticket.
