On 06/18/2012 04:25 AM, Robert Middleswarth wrote:
> 2012-06-17 05:21:24,682 ERROR
> [org.ovirt.engine.core.engineencryptutils.EncryptionUtils]
> (ajp--0.0.0.0-8009-9) Failed to decryptjava.io.FileNotFoundException:
> /etc/pki/ovirt-engine/.keystore (Too many open files)
> 2012-06-17 05:21:26,121 ERROR
> [org.ovirt.engine.core.engineencryptutils.EncryptionUtils]
> (QuartzScheduler_Worker-88) Failed to
> decryptjava.io.FileNotFoundException: /etc/pki/ovirt-engine/.keystore
> (Too many open files)
> 2012-06-17 05:21:26,130 ERROR
> [org.ovirt.engine.core.vdsbroker.vdsbroker.VdsBrokerCommand]
> (QuartzScheduler_Worker-88) XML RPC error in command GetCapabilitiesVDS
> ( Vds: ann100ovirt01 ), the error was:
> java.util.concurrent.ExecutionException:
> java.lang.reflect.InvocationTargetException, SocketException: Too many
> open files
> 2012-06-17 15:49:08,577 ERROR
> [org.ovirt.engine.core.engineencryptutils.EncryptionUtils]
> (QuartzScheduler_Worker-28) Failed to
> decryptjava.io.FileNotFoundException: /etc/pki/ovirt-engine/.keystore
> (Too many open files)
> 2012-06-17 20:21:13,943 ERROR
> [org.ovirt.engine.core.engineencryptutils.EncryptionUtils]
> (QuartzScheduler_Worker-64) Failed to
> decryptjava.io.FileNotFoundException: /etc/pki/ovirt-engine/.keystore
> (Too many open files)
> 2012-06-17 20:21:15,949 ERROR
> [org.ovirt.engine.core.engineencryptutils.EncryptionUtils]
> (QuartzScheduler_Worker-65) Failed to
> decryptjava.io.FileNotFoundException: /etc/pki/ovirt-engine/.keystore
> (Too many open files)
> 2012-06-17 21:00:15,133 ERROR
> [org.ovirt.engine.core.engineencryptutils.EncryptionUtils]
> (QuartzScheduler_Worker-98) Failed to
> decryptjava.io.FileNotFoundException: /etc/pki/ovirt-engine/.keystore
> (Too many open files)
> 2012-06-17 21:07:00,037 WARN [org.ovirt.engine.core.ServletUtils]
> (ajp--0.0.0.0-8009-5) File "/usr/share/vdsm-bootstrap/deployUtil.py is
> 47867 bytes long. Please reconsider using this servlet for files larger
> than 8192 bytes.
> 2012-06-17 21:07:00,115 WARN [org.ovirt.engine.core.ServletUtils]
> (ajp--0.0.0.0-8009-6) File "/usr/share/vdsm-bootstrap/vds_bootstrap.py
> is 33884 bytes long. Please reconsider using this servlet for files
> larger than 8192 bytes.
> 2012-06-17 21:32:10,686 WARN [org.ovirt.engine.core.ServletUtils]
> (ajp--0.0.0.0-8009-6) File "/usr/share/vdsm-bootstrap/deployUtil.py is
> 47867 bytes long. Please reconsider using this servlet for files larger
> than 8192 bytes.
> 2012-06-17 21:32:10,752 WARN [org.ovirt.engine.core.ServletUtils]
> (ajp--0.0.0.0-8009-1) File "/usr/share/vdsm-bootstrap/vds_bootstrap.py
> is 33884 bytes long. Please reconsider using this servlet for files
> larger than 8192 bytes.
> 2012-06-17 21:44:34,028 ERROR
> [org.ovirt.engine.core.vdsbroker.vdsbroker.VdsBrokerCommand]
> (QuartzScheduler_Worker-23) XML RPC error in command ListVDS ( Vds:
> ann100ovirt01 ), the error was: java.util.concurrent.ExecutionException:
> java.lang.reflect.InvocationTargetException, SocketException: Too many
> open files
> 2012-06-17 21:44:36,034 ERROR
> [org.ovirt.engine.core.engineencryptutils.EncryptionUtils]
> (QuartzScheduler_Worker-26) Failed to
> decryptjava.io.FileNotFoundException: /etc/pki/ovirt-engine/.keystore
> (Too many open files)
> 2012-06-17 21:44:36,045 ERROR
> [org.ovirt.engine.core.vdsbroker.vdsbroker.VdsBrokerCommand]
> (QuartzScheduler_Worker-26) XML RPC error in command GetCapabilitiesVDS
> ( Vds: ann100ovirt01 ), the error was:
> java.util.concurrent.ExecutionException:
> java.lang.reflect.InvocationTargetException, SocketException: Too many
> open files
>
>
> I restarted ovirt-engine and ran lsof -p 3557
>
> I see hundreds of the same open file.
>
> java 3557 ovirt 839r REG 253,0 3230 7865608
> /etc/pki/ovirt-engine/.keystore
> java 3557 ovirt 840r REG 253,0 3230 7865608
> /etc/pki/ovirt-engine/.keystore
> java 3557 ovirt 841r REG 253,0 3230 7865608
> /etc/pki/ovirt-engine/.keystore
> java 3557 ovirt 842r REG 253,0 3230 7865608
> /etc/pki/ovirt-engine/.keystore
> java 3557 ovirt 843r REG 253,0 3230 7865608
> /etc/pki/ovirt-engine/.keystore
> java 3557 ovirt 844r REG 253,0 3230 7865608
> /etc/pki/ovirt-engine/.keystore
> java 3557 ovirt 845r REG 253,0 3230 7865608
> /etc/pki/ovirt-engine/.keystore
> java 3557 ovirt 846r REG 253,0 3230 7865608
> /etc/pki/ovirt-engine/.keystore
> java 3557 ovirt 847r REG 253,0 3230 7865608
> /etc/pki/ovirt-engine/.keystore
> java 3557 ovirt 848r REG 253,0 3230 7865608
> /etc/pki/ovirt-engine/.keystore
> java 3557 ovirt 849r REG 253,0 3230 7865608
> /etc/pki/ovirt-engine/.keystore
> java 3557 ovirt 850r REG 253,0 3230 7865608
> /etc/pki/ovirt-engine/.keystore
> java 3557 ovirt 851r REG 253,0 3230 7865608
> /etc/pki/ovirt-engine/.keystore
> java 3557 ovirt 852r REG 253,0 3230 7865608
> /etc/pki/ovirt-engine/.keystore
> java 3557 ovirt 853r REG 253,0 3230 7865608
> /etc/pki/ovirt-engine/.keystore
> java 3557 ovirt 854r REG 253,0 3230 7865608
> /etc/pki/ovirt-engine/.keystore
>
> Is anyone else seeing this?
There are two issues that might be affecting you here:
1. We may have a leak of file descriptors on the logic that opens the
keystore. Can you take a look at the following change? Can you apply
it and let us know if it fixes your problem?
http://gerrit.ovirt.org/#/c/5423 I have been using rpm's but am attempting to
compile the code.
2. You may be having a problem with the file descriptors limit due to
other issues. During setup we should increase the file descriptor
limit from 1024 (the default) to 65535, but we do that for the "jboss"
user instead of for "ovirt". The fix for this issue is in review:
http://gerrit.ovirt.org/#/c/5375
The workaround for this is to edit the /etc/security/limits.conf file
and replace "jboss" with "ovirt", then restart the engine.
I
made that change but the web interface still dies after about 15 min
or so with the same number of open files and the same errror. I
rebooted the system so I am sure the change should have take place.
Thanks
Robert