Possible contamination of full lock table between ECP application servers or via remote databases ?

Question

Question

If several application servers are connected using ECP, and one of them create many locks, so many that lock table became full :

Should we expect only that application server to be impacted ? (usually, when lock table is full the system became instable). Or, are other application servers going to be impacted the same way ? (because that lock table is synchronized and maintained in sync across all servers)

I have another question (which is related) : is this the same with remote databases ?
Let's say 2 application servers access the same remote database and one application server create many locks. Is the other application server going to be impacted as well ?

Product version: IRIS 2021.1

$ZV: IRIS for Windows (x86-64) 2021.1 (Build 215U) Wed Jun 9 2021 09:39:22 EDT

Discussion (3)1

Log in or sign up to continue

score 0 · Answer 1 · 2024-10-02T10:30:05-04:00

Hello,
When several instances are connected to remote DB, then all their locks are managed in the IRIS instance where this DB is local (lets call this "the source")
I recommend increasing the "gmheap" parameter at "the source" to have more space on the lock table.

There are utilities to check the lock table free space, or full.

score 0 · Answer 2 · 2024-10-02T12:32:47-04:00

In other words, once the instance where the remote DBs are located has it's lock table full, any other server requiring a lock on a database hosted by this instance will be in trouble, is this right ?

Eg:
FOO and BAR database are located on an instance where the lock table is full
BAZ database is located on an instance where the lock table is almost empty

Application Server A	lock on FOO.X	denied
Application Server B	lock on FOO.X	denied
Application Server C	lock on BAR.X	denied
Application Server D	lock on BAZ.X	OK

Increasing gmheap : yes this might help but it you have some dummy process that enter a loop and create many many lock in a short amount of time, it's only delaying the issue (it will occur at some point no matter what)

score 0 · Answer 3 · 2024-10-03T02:54:28-04:00

When several instances are connected to remote DB, then all their locks are managed in the IRIS instance where this DB

That's only partially true: they are kept in both LOCKTABs, on data server and on app server as well.
This behavior can be easily checked in Management Portal. So if one of app servers makes great trouble with LOCKs, it can affect other servers.