On Thu, Mar 28, 2019 at 5:06 PM Tetsuo Handa <penguin-kernel@xxxxxxxxxxxxxxxxxxx> wrote: > > > On Thu, Mar 28, 2019 at 3:46 PM Tetsuo Handa > > <penguin-kernel@xxxxxxxxxxxxxxxxxxx> wrote: > > > > > > On 2019/03/29 6:37, Vincent Li wrote: > > > > Hi, > > > > > > > > not sure if this is the right place, I tried to use echo f > > > > > /proc/sysrq-trigger to manually trigger OOM, the OOM killer is > > > > triggered to kill a process, does it make sense to trigger OOM killer > > > > manually but not actually kill the process, this could be useful to > > > > diagnosis problem without actually killing a process in production > > > > box. > > > > > > Why not use "/usr/bin/top -o %MEM" etc. ? > > > Reading from /proc will give you more information than from SysRq. > > > > I am interested to see OOM output including swap entries per process > > in swap partition and all > > the other kernel internal virtual memory stats, I find it useful than > > top or free or /proc/meminfo > > > > Please read http://man7.org/linux/man-pages/man5/proc.5.html and/or > https://www.kernel.org/doc/Documentation/filesystems/proc.txt . Thanks, I know I can get lots of information from /proc/<pid>/, but I like the format/style that OOM dumps since I don't have to use some kind of script to parse out the /proc/<pid>, plus, OOM also dumps node/zone/page orders....information all together, lots of useful information so I can have big picture on the memory usage of the system.
