Hung processes

[troyd1]

we are on jbase 5.8/6.3, rhel 8.3.  We occasionally have jbase_agent processes that are hung when doing top in linux.  Sometimes several a day.  They will be using 100% cpu.  They could possibly be in a loop, but don't know.  The are not tied to any port in a listu.  How can I determine what those processes are doing?  Is there like a where command for process instead of port?

[Peter Falson]

In a nut shell, some (usually network based) request is coming into your jbase_agent listener which results in a child process (worker thread). 

There is/was a bug with jbase_agent worker threads not being cleaned up under certain circumstances. I was under the impression this was fixed in 5.8.7.

Sent from my iPhone

On Jan 26, 2023, at 12:49 PM, troyd1 <tr...@mninter.net> wrote:

we are on jbase 5.8/6.3, rhel 8.3.  We occasionally have jbase_agent processes that are hung when doing top in linux.  Sometimes several a day.  They will be using 100% cpu.  They could possibly be in a loop, but don't know.  The are not tied to any port in a listu.  How can I determine what those processes are doing?  Is there like a where command for process instead of port?

--
--
IMPORTANT: T24/Globus posts are no longer accepted on this forum.
 
To post, send email to jB...@googlegroups.com
To unsubscribe, send email to jBASE-un...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/jBASE?hl=en

---
You received this message because you are subscribed to the Google Groups "jBASE" group.
To unsubscribe from this group and stop receiving emails from it, send an email to jbase+un...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/jbase/eaeb3a8f-822e-41a2-8513-af324813aa64n%40googlegroups.com.

[troyd1]

Just to confirm, this is still happening on 5.8/6.3

Typically we will see them and they go away quickly there are currently 6 hung.  Is there a way to find out what initiated them?

Here is our top:

   2853 dm_jbas+  20   0 4123772    476    460 R  94.7   0.0   4284:05 jbase_a+
  54208 dm_jbas+  20   0 4192380    740    720 R  89.5   0.0   4206:28 jbase_a+
1034016 dm_jbas+  20   0 4127500   1620   1604 R  89.5   0.0   2849:39 jbase_a+
3377532 dm_jbas+  20   0 4119560   1.4g 616668 R  89.5   9.1  16:46.51 jbase_a+
4139747 dm_jbas+  20   0 4148796    504    488 R  89.5   0.0   4297:11 jbase_a+
3292549 dm_jbas+  20   0 4121908 995216 133128 R  84.2   6.1 105:44.34 jbase_a+

[Peter Falson]

You could stop and restart your jbase_agent with logging (to a file) and try to find the request(s). This may be needles in haystacks if you have a lot going on.

If when you find the culprit you could run a separate jbase_agent in foreground (using a different port) and even potentially debug if you have a version that supports that (and assuming it’s a REST type client app).

jbase_agent -d -JD {remaining options….}

With the above you can also do remote debugging with -JR

Having said that it might not be 

Sent from my iPhone

On Mar 9, 2023, at 9:01 AM, troyd1 <tr...@mninter.net> wrote:

Just to confirm, this is still happening on 5.8/6.3

To view this discussion on the web visit https://groups.google.com/d/msgid/jbase/e19abfe1-609f-4cd2-83a9-beeff5be8eddn%40googlegroups.com.