Unstable openSSI 1.9.x ...

Hi,

after some days uptime, my cluster "freezes" somehow in its cluster
functions, this means, e.g. the call aof the "cluster" command hangs,
I cannot boot a dependent node, a clusternode_shutdown hangs, even the
normal "halt" on the initnode hangs, so that I have to do a hard reset.

Non-cluster functions (e.g. mail server, web server) work.
There is sometimes the problem, that a remote login via ssh into the cluster
hangs, but I don't know, whether these problems are related.


In the error the last lines of the output of
"strace -f -r cluster -v 4" are:

89499      0.000442 close(3)            = 0
89499      0.000318 open("//lib/tls/libc.so.6", O_RDONLY) = 3
89499      0.000337 read(3,
....
89499      0.000377 close(3)            = 0
89499      0.001056 set_thread_area({entry_number:-1 -> 6,
base_addr:0xb7fe2800, limit:1048575, seg_32bit:1, contents:0,
read_exec_only:0, limit_in_pages:1, seg_not_present:0, useable:1}) = 0
89499      0.000481 ssisys(0xbffffd70, 0x14) = 125
89499      0.000263 ssisys(0xbffffd80, 0x14) = 125
89499      0.000396 brk(0)              = 0x804b000
89499      0.000214 brk(0x806c000)      = 0x806c000
89499      0.000226 brk(0)              = 0x806c000
89499      0.000322 ssisys(0xbffffd20, 0x14) = 125
89499      0.000230 ssisys(0xbffffd20, 0x14


So "cluster" hangs in a "ssisys"-(system-)call.

(How could I redirect the output of kdb into a file???)

in kdb the command btp <pid> yields:
schedule + 0x399 (0xca450e78, 0x1a10f28, 0x3cdfa97, 0x1, 0xdead4end)
schedule_timeout + 0x82 (0x4, 0x38, 0xbffffd60)
__clms_api_user_ + 0x92
clusternode_info + 0xbc

(copied manually, don't know how to write kdb-output into a file :-( )


My system is a small cluster (1 to 3 nodes normally, up to 10 nodes
maximum).
In the minimum case (where this error also occurs)
	one initnode, which should run 24x7
	a "file server", which will be automatically switched on, if
		some of my "terminal nodes" (windows PCs or my notebook),
		is up in the local network
	e "terminal node" (my desktop, which runs X11, is member in the
		cluster, and is switched on/off manually)

The system is a debian 3.1 based openSSI 1.9.1, with a (I hope, actual)
kernel from CVS..

Any ideas? Tips?

Regards, Robert



-------------------------------------------------------
Using Tomcat but need to do more? Need to support web services, security?
Get stuff done quickly with pre-integrated technology to make your job easier
Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642