What needs to get the thread list / process stack list when customer hit any hang symptom on CFS/VxFS/Process issue before forcing system dump?
[ Collecting evidence live ]
# q4 /stand/vmunix /dev/kmem
After q4 is started
Q4>load kthread_t from kthread_list max nkthread next kt_factp
Q4>trace -a -u pile > /tmp/thread.out
Q4>load proc_t from proc_list max nproc next p_factp
Q4> trace -a -u pile > /tmp/proc.out
Also the customer can repeat it against system dump file which is provided to Veritas Support Team.
[ Collecting evidence from the core ]
1. Please go to the same dir where the system dump file which is provided to us.
2. # q4 –p
After q4 is started
Q4>load kthread_t from kthread_list max nkthread next kt_factp
Q4>trace -a -u pile > /tmp/thread1.out
Q4>load proc_t from proc_list max nproc next p_factp
Q4> trace -a -u pile > /tmp/proc1.out
This will help Veritas Support Team to get the thread of the running system when issue happened & get the thread from core file.
Please send evidence_case_number.tar file including thread.out, thread1.out, proc.out and proc1.out files to Veritas Support.
This article contains information about providing data to Technical Support
https://www.veritas.com/docs/000097935
#cd /tmp
# tar cvf evidence_case_number.tar thread.out thread1.out proc.out proc1.out
Applies To
- SF/SFHA/SFCFS/SFORARAC/VCS for HP-UXv1/v2/v3