Subject | Re: [firebird-support] page type 5 lock conversion denied (215) |
---|---|
Author | Ann W. Harrison |
Post date | 2008-06-18T19:51:42Z |
derryck.welas wrote:
hash slots, the system does get slow, and you should get a lock
print and see if that's part of the problem. Look a the min
max and avg hash chains and increase the number of slots if
those numbers are higher than 10, 30, and 15 respectively.
It might be that the conversion failed because something timed
out, and increasing the size of hash table will make things
better.
The most likely reason for a conversion (e.g. from a read lock
to a write lock) to fail is that it would cause a deadlock with
another thread that's asked to convert a lock from read to write
on a page that your thread currently holds.
That shouldn't happen, but can under some rare circumstances.
Regards,
Ann
> Hi ,Unfortunately, probably no. When the lock manager runs out of
>
> Anybody know what this means ?
>
> "page 6726152, page type 5 lock conversion denied (215)"
>
> Should i increase "LockSemCount" or "LockHashSlots"
> Found it in the firebird.log after queries were running verry long on
> the server (I had to stop/Start some client applications).
>
> FB 2.0.3/Classic on Linux/redHat
hash slots, the system does get slow, and you should get a lock
print and see if that's part of the problem. Look a the min
max and avg hash chains and increase the number of slots if
those numbers are higher than 10, 30, and 15 respectively.
It might be that the conversion failed because something timed
out, and increasing the size of hash table will make things
better.
The most likely reason for a conversion (e.g. from a read lock
to a write lock) to fail is that it would cause a deadlock with
another thread that's asked to convert a lock from read to write
on a page that your thread currently holds.
That shouldn't happen, but can under some rare circumstances.
Regards,
Ann