On one of the Linux jenkins agent node, there are some PHP jobs that get stuck for days and I tried to troubleshoot that using strace command and I see the following output
epoll_wait(8, [{EPOLLHUP, {u32=2914329680, u64=94792842552400}}], 1, 5000) = 1
This doesn't provide much information when I dig into the corresponding fd which points to an anon_inode
8 -> anon_inode:[eventpoll]
and lsof gives me the following info
COMMAND PID USER FD TYPE DEVICE SIZE/OFF NODE NAME
php 16634 jenkins 8u a_inode 0,10 0 6379 [eventpoll]
^ Last line in the lsof -p 16634 output.
How to troubleshoot this further? This php process is running on a jenkins node (worker node).