Ryszard Musielak
Member
Hello,
I would like have some advice, maybe others have come across this problem. We have a Red Hat Linux server with a Progress database. I have a user running a report using SQL that links data from the Progress DB and and external Oracle DB. When this runs there is a process that creates an "unlinked" file, this can grow exponentially and can be seen using the lsof command:
# lsof | grep deleted | grep "_sqlsrv2" | nl
<file size>
1 _sqlsrv2 12661 root 27u REG 8,1 5371854848 297918 /usr/wrk/DBI1082624320rSmtLE (deleted)
1 _sqlsrv2 12661 root 27u REG 8,1 5407113216 297918 /usr/wrk/DBI1082624320rSmtLE (deleted)
1 _sqlsrv2 12661 root 27u REG 8,1 5425004544 297918 /usr/wrk/DBI1082624320rSmtLE (deleted)
1 _sqlsrv2 12661 root 27u REG 8,1 5439160320 297918 /usr/wrk/DBI1082624320rSmtLE (deleted)
1 _sqlsrv2 12661 root 27u REG 8,1 5473959936 297918 /usr/wrk/DBI1082624320rSmtLE (deleted)
1 _sqlsrv2 12661 root 27u REG 8,1 5481365504 297918 /usr/wrk/DBI1082624320rSmtLE (deleted)
1 _sqlsrv2 12661 root 27u REG 8,1 5490343936 297918 /usr/wrk/DBI1082624320rSmtLE (deleted)
1 _sqlsrv2 12661 root 27u REG 8,1 5538775040 297918 /usr/wrk/DBI1082624320rSmtLE (deleted)
1 _sqlsrv2 12661 root 27u REG 8,1 5597954048 297918 /usr/wrk/DBI1082624320rSmtLE (deleted)
1 _sqlsrv2 12661 root 27u REG 8,1 5657001984 297918 /usr/wrk/DBI1082624320rSmtLE (deleted)
You can see it's using over 5GB of space. The problem I've got is that it uses the /usr/wrk folder which is on the Root file system that is filling up and the space sometimes does not get released. After 3 months space usage has gone from 25% to now 67%. We had to reboot the server 3 months ago as we finally reached a point when / was filled up. That that cleared the space.
When I kill the process space goes back from 90% to 67%, so it doesn't seem like a problem, but we are sure this is the root cause of our problems. I suspect the process may sometimes crash and leave a zombie file in the system that is held up by Progress. Maybe there is a way to identify such a process?
The "top" command shows "0 zombie". The "promon" program doesn't offer any help.
Can anyone offer any advice or confirm that the _sqlsrv2 process is causing this issue? We will see if there is another way of running those reports to eliminate this problem.
This doesn't look like an OS problem, so Red Hat won't help.
Thanks,
Richard
I would like have some advice, maybe others have come across this problem. We have a Red Hat Linux server with a Progress database. I have a user running a report using SQL that links data from the Progress DB and and external Oracle DB. When this runs there is a process that creates an "unlinked" file, this can grow exponentially and can be seen using the lsof command:
# lsof | grep deleted | grep "_sqlsrv2" | nl
<file size>
1 _sqlsrv2 12661 root 27u REG 8,1 5371854848 297918 /usr/wrk/DBI1082624320rSmtLE (deleted)
1 _sqlsrv2 12661 root 27u REG 8,1 5407113216 297918 /usr/wrk/DBI1082624320rSmtLE (deleted)
1 _sqlsrv2 12661 root 27u REG 8,1 5425004544 297918 /usr/wrk/DBI1082624320rSmtLE (deleted)
1 _sqlsrv2 12661 root 27u REG 8,1 5439160320 297918 /usr/wrk/DBI1082624320rSmtLE (deleted)
1 _sqlsrv2 12661 root 27u REG 8,1 5473959936 297918 /usr/wrk/DBI1082624320rSmtLE (deleted)
1 _sqlsrv2 12661 root 27u REG 8,1 5481365504 297918 /usr/wrk/DBI1082624320rSmtLE (deleted)
1 _sqlsrv2 12661 root 27u REG 8,1 5490343936 297918 /usr/wrk/DBI1082624320rSmtLE (deleted)
1 _sqlsrv2 12661 root 27u REG 8,1 5538775040 297918 /usr/wrk/DBI1082624320rSmtLE (deleted)
1 _sqlsrv2 12661 root 27u REG 8,1 5597954048 297918 /usr/wrk/DBI1082624320rSmtLE (deleted)
1 _sqlsrv2 12661 root 27u REG 8,1 5657001984 297918 /usr/wrk/DBI1082624320rSmtLE (deleted)
You can see it's using over 5GB of space. The problem I've got is that it uses the /usr/wrk folder which is on the Root file system that is filling up and the space sometimes does not get released. After 3 months space usage has gone from 25% to now 67%. We had to reboot the server 3 months ago as we finally reached a point when / was filled up. That that cleared the space.
When I kill the process space goes back from 90% to 67%, so it doesn't seem like a problem, but we are sure this is the root cause of our problems. I suspect the process may sometimes crash and leave a zombie file in the system that is held up by Progress. Maybe there is a way to identify such a process?
The "top" command shows "0 zombie". The "promon" program doesn't offer any help.
Can anyone offer any advice or confirm that the _sqlsrv2 process is causing this issue? We will see if there is another way of running those reports to eliminate this problem.
This doesn't look like an OS problem, so Red Hat won't help.
Thanks,
Richard