Hi all,

I just got a third encounter of the lock's kind - I think.
I'm not at work for the next week, but there is this tiny voice in my
head asking for an answer.

We're running IDS 10.00.FC8 on Solaris Sparc. One of our applications
always got trapped on locks even when setting "lock mode to wait"
without any timeout. The table is fragmented by round robin and has
lock mode set to row.

So I tried to track the problem. First checking "onstat -u". Here's my
beastie waiting for the lock 'til the world is coming to an end:

address          flags   sessid   user     tty      wait
tout locks nreads   nwrites
266350ef8        L--PR-- 654369   intsys   -        10de23550
-1   2     152501   10

Now take a look at "onstat -k" and find the "wait userthread". Got it:

address          wtlist            owner             lklist
type     tblsnum  rowid    key#/bsiz
10de23550        266350ef8        (22ce57438)        10bc2f9f8
HDR+X    10000f2  328cf03     0  

Grep the owner of the lock in "onstat -x":
address          flags userthread       locks  beginlg curlog
logposit   isol    retrys coord
22ce57438        -L--G 0                4      0       0       0x0
DIRTY   0    

... and identify the session_id from userthread.
No way - there is this 0 looking at me and waiting for an explanation.
Am I missing the point or did our instance catch some internal flu?
If this is an internal function - shouldn't the lock be released at
least after some hours;-)
Can someone enlighten me?

And now for something completely different:
"onstat -k" offers a rowid - but the table is fragmented by round
robin so I can't do a select using this value. Is there any workaround
to find the locked column?

Thanks
Axel