on 04-08-2010 12:10 AM
After a system restart it is not possible to log into one of the two databases on the server to restart it.
Trying to log in via dbmcli never returns.
The dbmcli process takes 100% cpu
100 0.1 7:22.77 dbmcli
After using ctrl-C we get:
Error! Connection failed to node (local) for database <dbname>:
Reading header from pipe:(32:Broken pipe
We have a second database on the same server - there are no problems.
Nothing is written to knldiag or any other of the log files there.
MaxDB version is 7.6.06.03
What can we do ?
Regards
Bernhard
Hello ,
our database is now down for 1 day, because we cannot start it.
Has no one an idea how to log into an instance if dbmcli hangs completely ?
Is there any way to find out the reason for this?
Is there a tool that you can use to do a backup of the database - wihout connecting via dbmcli?
Any help is appreciated, thanks in advance!
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
Dear Mr. Bernhard Schiefer,
1) If you are SAP customer, I recommend to create the SAP message on this issue.
2) Please post additional information:
The output of the following commands on the database server:
sdbregview u2013l
What is the OS of the database server?
If you run the database on UNIX or LINUX please run & update with output:
uname u2013a
df u2013k /sapdb
ps u2013efe | grep sdb
ps u2013efe | grep kernel
ps u2013efe | grep dbmsrv
ps u2013efe | grep dbmcli
Do you have OS truss tool install on the database server?
If u2018YESu2019 => run
truss -f -o /tmp/truss.txt /sapdb/programs/bin/dbmcli
< wait 2-3min >
<Ctrl+C>
Post the truss.txt
Please check the errors in dbmsrv*.err file
3) Check if the database is running.< Is MAXDB kernel process active? >
If not, you could follow the document to copy the database to another server
http://wiki.sdn.sap.com/wiki/display/MaxDB/SAPMaxDBHowTo
->HowTo - Creating a clone of a SAP MaxDB database
->Go to section u201CCreating a clone manually via reusing volumes and parametersu201D
Thank you and best regards, Natalia Khlopina
Just to close this case:
We solved it by creating a new instance with exactly the same geometry as the old one
(same number and size of database- and log-files),
stopped it and coppied the database files from the non-working db to the new one.
Restart was successfull and everything works fine since then.
User | Count |
---|---|
86 | |
23 | |
11 | |
9 | |
8 | |
5 | |
5 | |
5 | |
5 | |
4 |
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.