On Mon, Sep 2, 2019 at 5:39 PM Toby Darling <toby@mrc-lmb.cam.ac.uk> wrote:
Hi
We have a couple of RHEL 7.6 (3.10.0-957.21.3.el7.x86_64) clients that have a number of uninterruptible threads and I'm wondering if we're looking at the issue fixed by https://www.spinics.net/lists/ceph-devel/msg45467.html (the fix hasn't made it into RHEL 7.7 3.10.0-1062).
Stack traces of the hung threads are at http://p.ip.fi/9pQA
There are a number of entries listed in /sys/kernel/debug/ceph/*/{osdc,mdsc} at http://p.ip.fi/VVzx
Unfortunately, the issue isn't consistently reproducible
Hi Toby, Looks very similar. If you see it happen again, dump all sleeping tasks with "echo w > /proc/sysrq-trigger". The "iput" task in that email goes to sleep voluntarily, so it is not considered hung by the watchdog. Thanks, Ilya