Hello friends,
in console i see lkmgr by one user (iNotes). How can i solve this problem? This user yesterday deleted many documents. Helps fixup?
Thanks for help and best regards.
"0A0C:0004-0AA8] rm_lkmgr_cpp:2308
[0A0C:0004-0AA8] rm_lkmgr_cpp:1401
[0A0C:0004-0AA8] rm_lkmgr_cpp:1670
[0A0C:0004-0AA8] dbutil_dbcompac_c:581
[0A0C:0004-0AA8] BEGIN ERR_SEMTIMEOUT:
[0A0C:0004-0AA8] caller - rm\lkmgr.cpp:1730"
JP
Usually lkmgr is generated when there is a contention between task/processes. Example, a compact is running on a database for an extended amount of time and a fixup is trying to run on the same database.
Check the database size, it may take some time for process to finish if the size is very large.
Then check if any task is running on the database mentioned on the lkmgr. Wait for it to finish or stop it and see if that helps.
Hello Jan, how long are the lkmgr messages been showing in the console?
Yes, you can definitely run a full maintenance to see if that help stop the messages. But if the lkmgr messages re-occurs, I recommend running a manual NSD and then open a case with us so we can check the logs. We will have to check why the server is holding the database.
Thanks!
Dear Joanne,
i see problem from today morning, but i think, that is from yesterday. User deleted many documents in mailbox.
Thanks and best,
JP
Hi Jan
Hope you are doing well.
The LKMGR messages is not error, this message is informational messages,
The LKMGR (Lock manager) Long Held Lock Dump events is a database access control process that runs in a transaction logging-enabled Domino environment.
Normally, Domino uses semaphores to control access to data and shared memory, but when transaction logging is enabled, the Long-Held Lock Dump protects the data in the process of opening all databases.
A simple scenario of a deadlock:
Thread A has resource R1 locked and is waiting for resource R2, while thread B has R2 locked and is waiting on R1. In this case both Thread A and B would keep waiting for each other until some intervention.
The deadlock detection feature would resolve this and prevent an hang as the threads may begin to wait on these resources.
In case if these messages are reporting continually then we need to identify which Process is locking the Database.
For the investigation of cause of the Database lock long held dump, you must perform following process at the time when these messages are reporting
Run Back to back three Manual NSDs with two to three minutes of interval between each two NSDs and collect NSD logs
How to run a manual NSD for Notes/Domino
https://support.hcl-software.com/csm?id=kb_article&sysparm_article=KB0032823
Please attach the NSD logs collected via Manual NSD along with Domino console logs and SEMDEBUG logs at the time of issue and raise Service ticket with HCL support team for further investigation.
I hope the above information will help in answering your concerns.
Thank you
Regards
Nishant Shendre
Just FYI - I am running 4 clusters of servers ranging from 12.0.2 FP1 to 12.0.2 FP3 HF70 and I continually have "LkMgr Long Held Lock Dump" on all my servers. It happens randomly across system databases, mail files, custom databases, there doesn't seem to be anything safe from the locks. I understand HCL is saying "this is an information message" however, the problem is when this happens to an active database/active users connected - all users are prohibited from functioning in the database and the notes client freezes and only recourse is to force shutdown the client. I think this is the part HCL doesn't get. I have 100's sometimes 1000's of users whose clients freeze until the system releases the database. I've been working with Domino since V4 and have not seen this random behavior. Nothing is safe from Lock Holds, even system db's on a server that has almost no user activity it will occur. Seems to be a 12.0.2 issue. I have a ticket open with HCL but wanted to post here.