CS4.0 build failed while setting up a dev environment on windows
Hi All: I have been setting up cloudstack4 dev environment on widows following the instruction of https://cwiki.apache.org/confluence/display/CLOUDSTACK/Setting+up+a+CloudStack+dev+environment+on+Windows There are some problem while deploying the database with command of mvn -P developer -pl developer -Ddeploydb on cygwin (Step 11 Build), Log as follows: $ mvn -P developer -pl developer -Ddeploydb [INFO] Scanning for projects... [ERROR] Could not find the selected project in the reactor: developer - [Help 1] [ERROR] [ERROR] To see the full stack trace of the errors, re-run Maven with the -e switch. [ERROR] Re-run Maven using the -X switch to enable full debug logging. [ERROR] [ERROR] For more information about the errors and possible solutions, please read the following articles: [ERROR] [Help 1] http://cwiki.apache.org/confluence/display/MAVEN/MavenExecutionException Does anywhere have the same failure?
Re: Review Request: Added unit test cases for api/commands
--- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/8355/#review14047 --- Suppose they should follow the same flavor, so I only took a look at AddTrafficTypeCmdTest.java 1) indent: should be 4 spaces 2) naming: method testCreateSuccess actually tests a failure scenario, and testCreateFailure actually tests a successful case. 3) test coverage: do you think both *create* and *execute* should be unit-tested in AddTrafficTypeCmd? The former method is called when a DB entry is persisted, the later is called when the async job is executed. - mice xia On Dec. 5, 2012, 11:53 a.m., Meghna Kale wrote: --- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/8355/ --- (Updated Dec. 5, 2012, 11:53 a.m.) Review request for cloudstack, Chip Childers, Prasanna Santhanam, Chiradeep Vittal, and Alex Huang. Description --- Added unit test cases for api/commands Diffs - api/src/com/cloud/api/commands/AssignToLoadBalancerRuleCmd.java 2a88e87 api/src/com/cloud/api/commands/AssignVMCmd.java be28cc0 api/src/com/cloud/api/commands/AssociateIPAddrCmd.java 7aaa5b5 api/src/com/cloud/api/commands/AuthorizeSecurityGroupEgressCmd.java a6088d0 api/src/com/cloud/api/commands/AuthorizeSecurityGroupIngressCmd.java e8f8b98 api/src/com/cloud/api/commands/CreateAutoScalePolicyCmd.java 4d93747 api/src/com/cloud/api/commands/CreateAutoScaleVmGroupCmd.java 83d7607 api/src/com/cloud/api/commands/CreateAutoScaleVmProfileCmd.java 68c85d0 api/src/com/cloud/api/commands/CreateConditionCmd.java eafd8a0 api/test/src/com/cloud/api/commands/test/AddTrafficTypeCmdTest.java PRE-CREATION api/test/src/com/cloud/api/commands/test/AssignToLoadBalancerRuleCmdTest.java PRE-CREATION api/test/src/com/cloud/api/commands/test/AssignVMCmdTest.java PRE-CREATION api/test/src/com/cloud/api/commands/test/AttachIsoCmdTest.java PRE-CREATION api/test/src/com/cloud/api/commands/test/AttachVolumeCmdTest.java PRE-CREATION api/test/src/com/cloud/api/commands/test/AuthorizeSecurityGroupEgressCmdTest.java PRE-CREATION api/test/src/com/cloud/api/commands/test/AuthorizeSecurityGroupIngressCmdTest.java PRE-CREATION api/test/src/com/cloud/api/commands/test/CancelMaintenanceCmdTest.java PRE-CREATION api/test/src/com/cloud/api/commands/test/CancelPrimaryStorageMaintenanceCmdTest.java PRE-CREATION api/test/src/com/cloud/api/commands/test/CreateAutoScalePolicyCmdTest.java PRE-CREATION api/test/src/com/cloud/api/commands/test/CreateAutoScaleVmGroupCmdTest.java PRE-CREATION api/test/src/com/cloud/api/commands/test/CreateAutoScaleVmProfileCmdTest.java PRE-CREATION api/test/src/com/cloud/api/commands/test/CreateConditionCmdTest.java PRE-CREATION api/test/src/com/cloud/api/commands/test/CreateCounterCmdTest.java PRE-CREATION Diff: https://reviews.apache.org/r/8355/diff/ Testing --- Thanks, Meghna Kale
RE: Impact of tomcat CVE-2012-4534
This reminds me of CVE-2011-3192, It's also a denial of service exploit existing in Apache HTTP 2.2.x to 2.2.19, our Apache on system VM is 2.2.16 if my system VM template is the latest one. http://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2011-3192 http://www.securityfocus.com/bid/49303 These two are probably not very critical, but if possible, it's valuable to add a regular security scan job. Regards Mice -Original Message- From: Gavin Lee [mailto:gavin@gmail.com] Sent: Wednesday, December 05, 2012 10:20 PM To: cloudstack Subject: Impact of tomcat CVE-2012-4534 This vulnerability possibly causes denial of service. See below link: http://mail-archives.apache.org/mod_mbox/www-announce/201212.mbox/%3c50be535a.9000...@apache.org%3E It was fixed in tomcat 6.0.36, but we recommand to use 6.0.33. Should we test a higher version and change the guide? -- Gavin
[jira] [Created] (CLOUDSTACK-585) DHCP entry provisioning is broken in the KVM agent
Wolfram Schlich created CLOUDSTACK-585: -- Summary: DHCP entry provisioning is broken in the KVM agent Key: CLOUDSTACK-585 URL: https://issues.apache.org/jira/browse/CLOUDSTACK-585 Project: CloudStack Issue Type: Bug Security Level: Public (Anyone can view this level - this is the default.) Components: Hypervisor Controller, KVM Environment: CloudStack 3.0.2 KVM agent running on Fedora 14 Reporter: Wolfram Schlich When adding an instance to a routerVM DHCP configuration, it seems that the KVM agent calls /usr/lib64/cloud/agent/scripts/network/domr/dhcp_entry.sh with wrongly constructed command line arguments, making the script fail to add correct entries (specifically default router, DNS servers and static routes) for that instance to the routerVM's /etc/dhcphosts.txt + /etc/dhcpopts.txt. Especially adding a specific default gateway fails, so the routerVM will always announce itself as the default router, because the correct entry in /etc/dhcpopts specifying the gateway of the instance's default network as the gateway is missing. This is especially nasty for non-default/additional networks of an instance, messing up the default routing. Examples: Management server log entry: 2012-11-29 14:36:37,764 DEBUG [resource.virtualnetwork.VirtualRoutingResource] (agentRequest-Handler-3:null) Executing: /usr/lib64/cloud/agent/./scripts/network/domr/dhcp_entry.sh -r 169.254.0.122 -v 172.31.2.233 -m 06:b5:88:00:02:30 -n vmname -d 172.31.2.1 -N 172.31.2.201 END-OF-LINE Notice the double spaces before -d and -N (and the extra space at the EOL). After patching /usr/lib64/cloud/agent/scripts/network/domr/dhcp_entry.sh to do meaningful logging, it's clear that the script does not get called with -d, but -d instead (same for -N), so with an extra space before the dash. Thus, getopts fails to parse/recognize these two arguments correctly and passed empty values for $dfltrt and $dns to the /root/edithosts.sh being called on the routerVM. It's also clear that the CloudStack KVM Java agent calls it with the wrongly constructed command line, because if a shell would interpret this command line, it would just ignore the extra spaces itself. I've not been able to dig it down, but I somehow suspect that one of ./utils/src/com/cloud/utils/script/Script.java:protected String buildCommandLine(String[] command) { } ./utils/src/com/cloud/utils/script/Script.java:public String execute(OutputInterpreter interpreter) { } might mess up building the command line of the command that had been built by ./core/src/com/cloud/agent/resource/virtualnetwork/VirtualRoutingResource.java:protected synchronized Answer execute (final DhcpEntryCommand cmd) { } before. I've not tried 4.0.0 so far, thus I cannot say whether it might be affected or not. As a workaround, I've patched dhcp_entry.sh to re-evaluate the positional parameters using 'set -- ${@}' (will attach a patch, also one for logging). -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Updated] (CLOUDSTACK-585) DHCP entry provisioning is broken in the KVM agent
[ https://issues.apache.org/jira/browse/CLOUDSTACK-585?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Wolfram Schlich updated CLOUDSTACK-585: --- Attachment: 2-dhcp_entry.sh.fix.patch Patch for /usr/lib64/cloud/agent/scripts/network/domr/dhcp_entry.sh to work around wrongly constructed arguments (re-evaluates the positional parameters) DHCP entry provisioning is broken in the KVM agent -- Key: CLOUDSTACK-585 URL: https://issues.apache.org/jira/browse/CLOUDSTACK-585 Project: CloudStack Issue Type: Bug Security Level: Public(Anyone can view this level - this is the default.) Components: Hypervisor Controller, KVM Environment: CloudStack 3.0.2 KVM agent running on Fedora 14 Reporter: Wolfram Schlich Labels: agent, kvm Attachments: 1-dhcp_entry.sh.log.patch, 2-dhcp_entry.sh.fix.patch When adding an instance to a routerVM DHCP configuration, it seems that the KVM agent calls /usr/lib64/cloud/agent/scripts/network/domr/dhcp_entry.sh with wrongly constructed command line arguments, making the script fail to add correct entries (specifically default router, DNS servers and static routes) for that instance to the routerVM's /etc/dhcphosts.txt + /etc/dhcpopts.txt. Especially adding a specific default gateway fails, so the routerVM will always announce itself as the default router, because the correct entry in /etc/dhcpopts specifying the gateway of the instance's default network as the gateway is missing. This is especially nasty for non-default/additional networks of an instance, messing up the default routing. Examples: Management server log entry: 2012-11-29 14:36:37,764 DEBUG [resource.virtualnetwork.VirtualRoutingResource] (agentRequest-Handler-3:null) Executing: /usr/lib64/cloud/agent/./scripts/network/domr/dhcp_entry.sh -r 169.254.0.122 -v 172.31.2.233 -m 06:b5:88:00:02:30 -n vmname -d 172.31.2.1 -N 172.31.2.201 END-OF-LINE Notice the double spaces before -d and -N (and the extra space at the EOL). After patching /usr/lib64/cloud/agent/scripts/network/domr/dhcp_entry.sh to do meaningful logging, it's clear that the script does not get called with -d, but -d instead (same for -N), so with an extra space before the dash. Thus, getopts fails to parse/recognize these two arguments correctly and passed empty values for $dfltrt and $dns to the /root/edithosts.sh being called on the routerVM. It's also clear that the CloudStack KVM Java agent calls it with the wrongly constructed command line, because if a shell would interpret this command line, it would just ignore the extra spaces itself. I've not been able to dig it down, but I somehow suspect that one of ./utils/src/com/cloud/utils/script/Script.java:protected String buildCommandLine(String[] command) { } ./utils/src/com/cloud/utils/script/Script.java:public String execute(OutputInterpreter interpreter) { } might mess up building the command line of the command that had been built by ./core/src/com/cloud/agent/resource/virtualnetwork/VirtualRoutingResource.java:protected synchronized Answer execute (final DhcpEntryCommand cmd) { } before. I've not tried 4.0.0 so far, thus I cannot say whether it might be affected or not. As a workaround, I've patched dhcp_entry.sh to re-evaluate the positional parameters using 'set -- ${@}' (will attach a patch, also one for logging). -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Updated] (CLOUDSTACK-585) DHCP entry provisioning is broken in the KVM agent
[ https://issues.apache.org/jira/browse/CLOUDSTACK-585?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Wolfram Schlich updated CLOUDSTACK-585: --- Attachment: 1-dhcp_entry.sh.log.patch Patch for /usr/lib64/cloud/agent/scripts/network/domr/dhcp_entry.sh to enable logging of wrongly constructed arguments and empty parameters passed to routerVM:/root/edithosts.sh DHCP entry provisioning is broken in the KVM agent -- Key: CLOUDSTACK-585 URL: https://issues.apache.org/jira/browse/CLOUDSTACK-585 Project: CloudStack Issue Type: Bug Security Level: Public(Anyone can view this level - this is the default.) Components: Hypervisor Controller, KVM Environment: CloudStack 3.0.2 KVM agent running on Fedora 14 Reporter: Wolfram Schlich Labels: agent, kvm Attachments: 1-dhcp_entry.sh.log.patch, 2-dhcp_entry.sh.fix.patch When adding an instance to a routerVM DHCP configuration, it seems that the KVM agent calls /usr/lib64/cloud/agent/scripts/network/domr/dhcp_entry.sh with wrongly constructed command line arguments, making the script fail to add correct entries (specifically default router, DNS servers and static routes) for that instance to the routerVM's /etc/dhcphosts.txt + /etc/dhcpopts.txt. Especially adding a specific default gateway fails, so the routerVM will always announce itself as the default router, because the correct entry in /etc/dhcpopts specifying the gateway of the instance's default network as the gateway is missing. This is especially nasty for non-default/additional networks of an instance, messing up the default routing. Examples: Management server log entry: 2012-11-29 14:36:37,764 DEBUG [resource.virtualnetwork.VirtualRoutingResource] (agentRequest-Handler-3:null) Executing: /usr/lib64/cloud/agent/./scripts/network/domr/dhcp_entry.sh -r 169.254.0.122 -v 172.31.2.233 -m 06:b5:88:00:02:30 -n vmname -d 172.31.2.1 -N 172.31.2.201 END-OF-LINE Notice the double spaces before -d and -N (and the extra space at the EOL). After patching /usr/lib64/cloud/agent/scripts/network/domr/dhcp_entry.sh to do meaningful logging, it's clear that the script does not get called with -d, but -d instead (same for -N), so with an extra space before the dash. Thus, getopts fails to parse/recognize these two arguments correctly and passed empty values for $dfltrt and $dns to the /root/edithosts.sh being called on the routerVM. It's also clear that the CloudStack KVM Java agent calls it with the wrongly constructed command line, because if a shell would interpret this command line, it would just ignore the extra spaces itself. I've not been able to dig it down, but I somehow suspect that one of ./utils/src/com/cloud/utils/script/Script.java:protected String buildCommandLine(String[] command) { } ./utils/src/com/cloud/utils/script/Script.java:public String execute(OutputInterpreter interpreter) { } might mess up building the command line of the command that had been built by ./core/src/com/cloud/agent/resource/virtualnetwork/VirtualRoutingResource.java:protected synchronized Answer execute (final DhcpEntryCommand cmd) { } before. I've not tried 4.0.0 so far, thus I cannot say whether it might be affected or not. As a workaround, I've patched dhcp_entry.sh to re-evaluate the positional parameters using 'set -- ${@}' (will attach a patch, also one for logging). -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Updated] (CLOUDSTACK-585) DHCP entry provisioning is broken in the KVM agent
[ https://issues.apache.org/jira/browse/CLOUDSTACK-585?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Wolfram Schlich updated CLOUDSTACK-585: --- Description: When adding an instance to a routerVM DHCP configuration, it seems that the KVM agent calls /usr/lib64/cloud/agent/scripts/network/domr/dhcp_entry.sh with wrongly constructed command line arguments, making the script fail to add correct entries (specifically default router, DNS servers and static routes) for that instance to the routerVM's /etc/dhcphosts.txt + /etc/dhcpopts.txt. Especially adding a specific default gateway fails, so the routerVM will always announce itself as the default router, because the correct entry in /etc/dhcpopts specifying the gateway of the instance's default network as the gateway is missing. This is especially nasty for non-default/additional networks of an instance, messing up the default routing. Examples: Management server log entry: codepre2012-11-29 14:36:37,764 DEBUG [resource.virtualnetwork.VirtualRoutingResource] (agentRequest-Handler-3:null) Executing: /usr/lib64/cloud/agent/./scripts/network/domr/dhcp_entry.sh -r 169.254.0.122 -v 172.31.2.233 -m 06:b5:88:00:02:30 -n vmname -d 172.31.2.1 -N 172.31.2.201 /pre/code Notice the double spaces before -d and -N (and the extra space at the EOL). After patching /usr/lib64/cloud/agent/scripts/network/domr/dhcp_entry.sh to do meaningful logging, it's clear that the script does not get called with -d, but -d instead (same for -N), so with an extra space before the dash. Thus, getopts fails to parse/recognize these two arguments correctly and passed empty values for $dfltrt and $dns to the /root/edithosts.sh being called on the routerVM. It's also clear that the CloudStack KVM Java agent calls it with the wrongly constructed command line, because if a shell would interpret this command line, it would just ignore the extra spaces itself. I've not been able to dig it down, but I somehow suspect that one of ./utils/src/com/cloud/utils/script/Script.java:protected String buildCommandLine(String[] command) { } ./utils/src/com/cloud/utils/script/Script.java:public String execute(OutputInterpreter interpreter) { } might mess up building the command line of the command that had been built by ./core/src/com/cloud/agent/resource/virtualnetwork/VirtualRoutingResource.java:protected synchronized Answer execute (final DhcpEntryCommand cmd) { } before. I've not tried 4.0.0 so far, thus I cannot say whether it might be affected or not. As a workaround, I've patched dhcp_entry.sh to re-evaluate the positional parameters using 'set -- ${@}' (will attach a patch, also one for logging). was: When adding an instance to a routerVM DHCP configuration, it seems that the KVM agent calls /usr/lib64/cloud/agent/scripts/network/domr/dhcp_entry.sh with wrongly constructed command line arguments, making the script fail to add correct entries (specifically default router, DNS servers and static routes) for that instance to the routerVM's /etc/dhcphosts.txt + /etc/dhcpopts.txt. Especially adding a specific default gateway fails, so the routerVM will always announce itself as the default router, because the correct entry in /etc/dhcpopts specifying the gateway of the instance's default network as the gateway is missing. This is especially nasty for non-default/additional networks of an instance, messing up the default routing. Examples: Management server log entry: 2012-11-29 14:36:37,764 DEBUG [resource.virtualnetwork.VirtualRoutingResource] (agentRequest-Handler-3:null) Executing: /usr/lib64/cloud/agent/./scripts/network/domr/dhcp_entry.sh -r 169.254.0.122 -v 172.31.2.233 -m 06:b5:88:00:02:30 -n vmname -d 172.31.2.1 -N 172.31.2.201 END-OF-LINE Notice the double spaces before -d and -N (and the extra space at the EOL). After patching /usr/lib64/cloud/agent/scripts/network/domr/dhcp_entry.sh to do meaningful logging, it's clear that the script does not get called with -d, but -d instead (same for -N), so with an extra space before the dash. Thus, getopts fails to parse/recognize these two arguments correctly and passed empty values for $dfltrt and $dns to the /root/edithosts.sh being called on the routerVM. It's also clear that the CloudStack KVM Java agent calls it with the wrongly constructed command line, because if a shell would interpret this command line, it would just ignore the extra spaces itself. I've not been able to dig it down, but I somehow suspect that one of ./utils/src/com/cloud/utils/script/Script.java:protected String buildCommandLine(String[] command) { } ./utils/src/com/cloud/utils/script/Script.java:public String execute(OutputInterpreter interpreter) { } might mess up building the command line of the command that had been built by ./core/src/com/cloud/agent/resource/virtualnetwork/VirtualRoutingResource.java:protected synchronized Answer
[jira] [Updated] (CLOUDSTACK-585) DHCP entry provisioning is broken in the KVM agent
[ https://issues.apache.org/jira/browse/CLOUDSTACK-585?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Wolfram Schlich updated CLOUDSTACK-585: --- Description: When adding an instance to a routerVM DHCP configuration, it seems that the KVM agent calls /usr/lib64/cloud/agent/scripts/network/domr/dhcp_entry.sh with wrongly constructed command line arguments, making the script fail to add correct entries (specifically default router, DNS servers and static routes) for that instance to the routerVM's /etc/dhcphosts.txt + /etc/dhcpopts.txt. Especially adding a specific default gateway fails, so the routerVM will always announce itself as the default router, because the correct entry in /etc/dhcpopts specifying the gateway of the instance's default network as the gateway is missing. This is especially nasty for non-default/additional networks of an instance, messing up the default routing. Examples: Management server log entry: 2012-11-29 14:36:37,764 DEBUG [resource.virtualnetwork.VirtualRoutingResource] (agentRequest-Handler-3:null) Executing: /usr/lib64/cloud/agent/./scripts/network/domr/dhcp_entry.sh -r 169.254.0.122 -v 172.31.2.233 -m 06:b5:88:00:02:30 -n vmname -d 172.31.2.1 -N 172.31.2.201 Notice the double spaces before -d and -N (and the extra space at the EOL) -- WARNING: you have to view the source of this description to actually _see_ the double spaces!) After patching /usr/lib64/cloud/agent/scripts/network/domr/dhcp_entry.sh to do meaningful logging, it's clear that the script does not get called with -d, but -d instead (same for -N), so with an extra space before the dash. Thus, getopts fails to parse/recognize these two arguments correctly and passed empty values for $dfltrt and $dns to the /root/edithosts.sh being called on the routerVM. It's also clear that the CloudStack KVM Java agent calls it with the wrongly constructed command line, because if a shell would interpret this command line, it would just ignore the extra spaces itself. I've not been able to dig it down, but I somehow suspect that one of ./utils/src/com/cloud/utils/script/Script.java:protected String buildCommandLine(String[] command) { } ./utils/src/com/cloud/utils/script/Script.java:public String execute(OutputInterpreter interpreter) { } might mess up building the command line of the command that had been built by ./core/src/com/cloud/agent/resource/virtualnetwork/VirtualRoutingResource.java:protected synchronized Answer execute (final DhcpEntryCommand cmd) { } before. I've not tried 4.0.0 so far, thus I cannot say whether it might be affected or not. As a workaround, I've patched dhcp_entry.sh to re-evaluate the positional parameters using 'set -- ${@}' (will attach a patch, also one for logging). was: When adding an instance to a routerVM DHCP configuration, it seems that the KVM agent calls /usr/lib64/cloud/agent/scripts/network/domr/dhcp_entry.sh with wrongly constructed command line arguments, making the script fail to add correct entries (specifically default router, DNS servers and static routes) for that instance to the routerVM's /etc/dhcphosts.txt + /etc/dhcpopts.txt. Especially adding a specific default gateway fails, so the routerVM will always announce itself as the default router, because the correct entry in /etc/dhcpopts specifying the gateway of the instance's default network as the gateway is missing. This is especially nasty for non-default/additional networks of an instance, messing up the default routing. Examples: Management server log entry: codepre2012-11-29 14:36:37,764 DEBUG [resource.virtualnetwork.VirtualRoutingResource] (agentRequest-Handler-3:null) Executing: /usr/lib64/cloud/agent/./scripts/network/domr/dhcp_entry.sh -r 169.254.0.122 -v 172.31.2.233 -m 06:b5:88:00:02:30 -n vmname -d 172.31.2.1 -N 172.31.2.201 /pre/code Notice the double spaces before -d and -N (and the extra space at the EOL). After patching /usr/lib64/cloud/agent/scripts/network/domr/dhcp_entry.sh to do meaningful logging, it's clear that the script does not get called with -d, but -d instead (same for -N), so with an extra space before the dash. Thus, getopts fails to parse/recognize these two arguments correctly and passed empty values for $dfltrt and $dns to the /root/edithosts.sh being called on the routerVM. It's also clear that the CloudStack KVM Java agent calls it with the wrongly constructed command line, because if a shell would interpret this command line, it would just ignore the extra spaces itself. I've not been able to dig it down, but I somehow suspect that one of ./utils/src/com/cloud/utils/script/Script.java:protected String buildCommandLine(String[] command) { } ./utils/src/com/cloud/utils/script/Script.java:public String execute(OutputInterpreter interpreter) { } might mess up building the command line of the command that had been built by
[jira] [Updated] (CLOUDSTACK-585) DHCP entry provisioning is broken in the KVM agent
[ https://issues.apache.org/jira/browse/CLOUDSTACK-585?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Wolfram Schlich updated CLOUDSTACK-585: --- Description: When adding an instance to a routerVM DHCP configuration, it seems that the KVM agent calls /usr/lib64/cloud/agent/scripts/network/domr/dhcp_entry.sh with wrongly constructed command line arguments, making the script fail to add correct entries (specifically default router, DNS servers and static routes) for that instance to the routerVM's /etc/dhcphosts.txt + /etc/dhcpopts.txt. Especially adding a specific default gateway fails, so the routerVM will always announce itself as the default router, because the correct entry in /etc/dhcpopts specifying the gateway of the instance's default network as the gateway is missing. This is especially nasty for non-default/additional networks of an instance, messing up the default routing. Examples: Management server log entry: {noformat} 2012-11-29 14:36:37,764 DEBUG [resource.virtualnetwork.VirtualRoutingResource] (agentRequest-Handler-3:null) Executing: /usr/lib64/cloud/agent/./scripts/network/domr/dhcp_entry.sh -r 169.254.0.122 -v 172.31.2.233 -m 06:b5:88:00:02:30 -n vmname -d 172.31.2.1 -N 172.31.2.201 {noformat} Notice the double spaces before -d and -N (and the extra space at the EOL) -- WARNING: you have to view the source of this description to actually _see_ the double spaces!) After patching /usr/lib64/cloud/agent/scripts/network/domr/dhcp_entry.sh to do meaningful logging, it's clear that the script does not get called with -d, but -d instead (same for -N), so with an extra space before the dash. Thus, getopts fails to parse/recognize these two arguments correctly and passed empty values for $dfltrt and $dns to the /root/edithosts.sh being called on the routerVM. It's also clear that the CloudStack KVM Java agent calls it with the wrongly constructed command line, because if a shell would interpret this command line, it would just ignore the extra spaces itself. I've not been able to dig it down, but I somehow suspect that one of ./utils/src/com/cloud/utils/script/Script.java:protected String buildCommandLine(String[] command) { } ./utils/src/com/cloud/utils/script/Script.java:public String execute(OutputInterpreter interpreter) { } might mess up building the command line of the command that had been built by ./core/src/com/cloud/agent/resource/virtualnetwork/VirtualRoutingResource.java:protected synchronized Answer execute (final DhcpEntryCommand cmd) { } before. I've not tried 4.0.0 so far, thus I cannot say whether it might be affected or not. As a workaround, I've patched dhcp_entry.sh to re-evaluate the positional parameters using 'set -- ${@}' (will attach a patch, also one for logging). was: When adding an instance to a routerVM DHCP configuration, it seems that the KVM agent calls /usr/lib64/cloud/agent/scripts/network/domr/dhcp_entry.sh with wrongly constructed command line arguments, making the script fail to add correct entries (specifically default router, DNS servers and static routes) for that instance to the routerVM's /etc/dhcphosts.txt + /etc/dhcpopts.txt. Especially adding a specific default gateway fails, so the routerVM will always announce itself as the default router, because the correct entry in /etc/dhcpopts specifying the gateway of the instance's default network as the gateway is missing. This is especially nasty for non-default/additional networks of an instance, messing up the default routing. Examples: Management server log entry: 2012-11-29 14:36:37,764 DEBUG [resource.virtualnetwork.VirtualRoutingResource] (agentRequest-Handler-3:null) Executing: /usr/lib64/cloud/agent/./scripts/network/domr/dhcp_entry.sh -r 169.254.0.122 -v 172.31.2.233 -m 06:b5:88:00:02:30 -n vmname -d 172.31.2.1 -N 172.31.2.201 Notice the double spaces before -d and -N (and the extra space at the EOL) -- WARNING: you have to view the source of this description to actually _see_ the double spaces!) After patching /usr/lib64/cloud/agent/scripts/network/domr/dhcp_entry.sh to do meaningful logging, it's clear that the script does not get called with -d, but -d instead (same for -N), so with an extra space before the dash. Thus, getopts fails to parse/recognize these two arguments correctly and passed empty values for $dfltrt and $dns to the /root/edithosts.sh being called on the routerVM. It's also clear that the CloudStack KVM Java agent calls it with the wrongly constructed command line, because if a shell would interpret this command line, it would just ignore the extra spaces itself. I've not been able to dig it down, but I somehow suspect that one of ./utils/src/com/cloud/utils/script/Script.java:protected String buildCommandLine(String[] command) { } ./utils/src/com/cloud/utils/script/Script.java:public String execute(OutputInterpreter interpreter) { }
[jira] [Updated] (CLOUDSTACK-585) DHCP entry provisioning is broken in the KVM agent
[ https://issues.apache.org/jira/browse/CLOUDSTACK-585?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Wolfram Schlich updated CLOUDSTACK-585: --- Description: When adding an instance to a routerVM DHCP configuration, it seems that the KVM agent calls /usr/lib64/cloud/agent/scripts/network/domr/dhcp_entry.sh with wrongly constructed command line arguments, making the script fail to add correct entries (specifically default router, DNS servers and static routes) for that instance to the routerVM's /etc/dhcphosts.txt + /etc/dhcpopts.txt. Especially adding a specific default gateway fails, so the routerVM will always announce itself as the default router, because the correct entry in /etc/dhcpopts specifying the gateway of the instance's default network as the gateway is missing. This is especially nasty for non-default/additional networks of an instance, messing up the default routing. Examples: Management server log entry: 2012-11-29 14:36:37,764 DEBUG [resource.virtualnetwork.VirtualRoutingResource] (agentRequest-Handler-3:null) Executing: /usr/lib64/cloud/agent/./scripts/network/domr/dhcp_entry.sh -r 169.254.0.122 -v 172.31.2.233 -m 06:b5:88:00:02:30 -n vmname -d 172.31.2.1 -N 172.31.2.201 Notice the double spaces before -d and -N (and the extra space at the EOL) -- WARNING: you have to view the source of this description to actually _see_ the double spaces!) After patching /usr/lib64/cloud/agent/scripts/network/domr/dhcp_entry.sh to do meaningful logging, it's clear that the script does not get called with -d, but -d instead (same for -N), so with an extra space before the dash. Thus, getopts fails to parse/recognize these two arguments correctly and passed empty values for $dfltrt and $dns to the /root/edithosts.sh being called on the routerVM. It's also clear that the CloudStack KVM Java agent calls it with the wrongly constructed command line, because if a shell would interpret this command line, it would just ignore the extra spaces itself. I've not been able to dig it down, but I somehow suspect that one of ./utils/src/com/cloud/utils/script/Script.java:protected String buildCommandLine(String[] command) { } ./utils/src/com/cloud/utils/script/Script.java:public String execute(OutputInterpreter interpreter) { } might mess up building the command line of the command that had been built by ./core/src/com/cloud/agent/resource/virtualnetwork/VirtualRoutingResource.java:protected synchronized Answer execute (final DhcpEntryCommand cmd) { } before. I've not tried 4.0.0 so far, thus I cannot say whether it might be affected or not. As a workaround, I've patched dhcp_entry.sh to re-evaluate the positional parameters using 'set -- ${@}' (will attach a patch, also one for logging). was: When adding an instance to a routerVM DHCP configuration, it seems that the KVM agent calls /usr/lib64/cloud/agent/scripts/network/domr/dhcp_entry.sh with wrongly constructed command line arguments, making the script fail to add correct entries (specifically default router, DNS servers and static routes) for that instance to the routerVM's /etc/dhcphosts.txt + /etc/dhcpopts.txt. Especially adding a specific default gateway fails, so the routerVM will always announce itself as the default router, because the correct entry in /etc/dhcpopts specifying the gateway of the instance's default network as the gateway is missing. This is especially nasty for non-default/additional networks of an instance, messing up the default routing. Examples: Management server log entry: {noformat} 2012-11-29 14:36:37,764 DEBUG [resource.virtualnetwork.VirtualRoutingResource] (agentRequest-Handler-3:null) Executing: /usr/lib64/cloud/agent/./scripts/network/domr/dhcp_entry.sh -r 169.254.0.122 -v 172.31.2.233 -m 06:b5:88:00:02:30 -n vmname -d 172.31.2.1 -N 172.31.2.201 {noformat} Notice the double spaces before -d and -N (and the extra space at the EOL) -- WARNING: you have to view the source of this description to actually _see_ the double spaces!) After patching /usr/lib64/cloud/agent/scripts/network/domr/dhcp_entry.sh to do meaningful logging, it's clear that the script does not get called with -d, but -d instead (same for -N), so with an extra space before the dash. Thus, getopts fails to parse/recognize these two arguments correctly and passed empty values for $dfltrt and $dns to the /root/edithosts.sh being called on the routerVM. It's also clear that the CloudStack KVM Java agent calls it with the wrongly constructed command line, because if a shell would interpret this command line, it would just ignore the extra spaces itself. I've not been able to dig it down, but I somehow suspect that one of ./utils/src/com/cloud/utils/script/Script.java:protected String buildCommandLine(String[] command) { } ./utils/src/com/cloud/utils/script/Script.java:public String execute(OutputInterpreter interpreter) { }
[jira] [Comment Edited] (CLOUDSTACK-585) DHCP entry provisioning is broken in the KVM agent
[ https://issues.apache.org/jira/browse/CLOUDSTACK-585?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13510533#comment-13510533 ] Wolfram Schlich edited comment on CLOUDSTACK-585 at 12/5/12 3:37 PM: - 2-dhcp_entry.sh.fix.patch: Patch for /usr/lib64/cloud/agent/scripts/network/domr/dhcp_entry.sh to work around wrongly constructed arguments (re-evaluates the positional parameters) was (Author: wschlich): Patch for /usr/lib64/cloud/agent/scripts/network/domr/dhcp_entry.sh to work around wrongly constructed arguments (re-evaluates the positional parameters) DHCP entry provisioning is broken in the KVM agent -- Key: CLOUDSTACK-585 URL: https://issues.apache.org/jira/browse/CLOUDSTACK-585 Project: CloudStack Issue Type: Bug Security Level: Public(Anyone can view this level - this is the default.) Components: Hypervisor Controller, KVM Environment: CloudStack 3.0.2 KVM agent running on Fedora 14 Reporter: Wolfram Schlich Labels: agent, kvm Attachments: 1-dhcp_entry.sh.log.patch, 2-dhcp_entry.sh.fix.patch When adding an instance to a routerVM DHCP configuration, it seems that the KVM agent calls /usr/lib64/cloud/agent/scripts/network/domr/dhcp_entry.sh with wrongly constructed command line arguments, making the script fail to add correct entries (specifically default router, DNS servers and static routes) for that instance to the routerVM's /etc/dhcphosts.txt + /etc/dhcpopts.txt. Especially adding a specific default gateway fails, so the routerVM will always announce itself as the default router, because the correct entry in /etc/dhcpopts specifying the gateway of the instance's default network as the gateway is missing. This is especially nasty for non-default/additional networks of an instance, messing up the default routing. Examples: Management server log entry: 2012-11-29 14:36:37,764 DEBUG [resource.virtualnetwork.VirtualRoutingResource] (agentRequest-Handler-3:null) Executing: /usr/lib64/cloud/agent/./scripts/network/domr/dhcp_entry.sh -r 169.254.0.122 -v 172.31.2.233 -m 06:b5:88:00:02:30 -n vmname -d 172.31.2.1 -N 172.31.2.201 Notice the double spaces before -d and -N (and the extra space at the EOL) -- WARNING: you have to view the source of this description to actually _see_ the double spaces!) After patching /usr/lib64/cloud/agent/scripts/network/domr/dhcp_entry.sh to do meaningful logging, it's clear that the script does not get called with -d, but -d instead (same for -N), so with an extra space before the dash. Thus, getopts fails to parse/recognize these two arguments correctly and passed empty values for $dfltrt and $dns to the /root/edithosts.sh being called on the routerVM. It's also clear that the CloudStack KVM Java agent calls it with the wrongly constructed command line, because if a shell would interpret this command line, it would just ignore the extra spaces itself. I've not been able to dig it down, but I somehow suspect that one of ./utils/src/com/cloud/utils/script/Script.java:protected String buildCommandLine(String[] command) { } ./utils/src/com/cloud/utils/script/Script.java:public String execute(OutputInterpreter interpreter) { } might mess up building the command line of the command that had been built by ./core/src/com/cloud/agent/resource/virtualnetwork/VirtualRoutingResource.java:protected synchronized Answer execute (final DhcpEntryCommand cmd) { } before. I've not tried 4.0.0 so far, thus I cannot say whether it might be affected or not. As a workaround, I've patched dhcp_entry.sh to re-evaluate the positional parameters using 'set -- ${@}' (will attach a patch, also one for logging). -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Comment Edited] (CLOUDSTACK-585) DHCP entry provisioning is broken in the KVM agent
[ https://issues.apache.org/jira/browse/CLOUDSTACK-585?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13510532#comment-13510532 ] Wolfram Schlich edited comment on CLOUDSTACK-585 at 12/5/12 3:37 PM: - 1-dhcp_entry.sh.log.patch: Patch for /usr/lib64/cloud/agent/scripts/network/domr/dhcp_entry.sh to enable logging of wrongly constructed arguments and empty parameters passed to routerVM:/root/edithosts.sh was (Author: wschlich): Patch for /usr/lib64/cloud/agent/scripts/network/domr/dhcp_entry.sh to enable logging of wrongly constructed arguments and empty parameters passed to routerVM:/root/edithosts.sh DHCP entry provisioning is broken in the KVM agent -- Key: CLOUDSTACK-585 URL: https://issues.apache.org/jira/browse/CLOUDSTACK-585 Project: CloudStack Issue Type: Bug Security Level: Public(Anyone can view this level - this is the default.) Components: Hypervisor Controller, KVM Environment: CloudStack 3.0.2 KVM agent running on Fedora 14 Reporter: Wolfram Schlich Labels: agent, kvm Attachments: 1-dhcp_entry.sh.log.patch, 2-dhcp_entry.sh.fix.patch When adding an instance to a routerVM DHCP configuration, it seems that the KVM agent calls /usr/lib64/cloud/agent/scripts/network/domr/dhcp_entry.sh with wrongly constructed command line arguments, making the script fail to add correct entries (specifically default router, DNS servers and static routes) for that instance to the routerVM's /etc/dhcphosts.txt + /etc/dhcpopts.txt. Especially adding a specific default gateway fails, so the routerVM will always announce itself as the default router, because the correct entry in /etc/dhcpopts specifying the gateway of the instance's default network as the gateway is missing. This is especially nasty for non-default/additional networks of an instance, messing up the default routing. Examples: Management server log entry: 2012-11-29 14:36:37,764 DEBUG [resource.virtualnetwork.VirtualRoutingResource] (agentRequest-Handler-3:null) Executing: /usr/lib64/cloud/agent/./scripts/network/domr/dhcp_entry.sh -r 169.254.0.122 -v 172.31.2.233 -m 06:b5:88:00:02:30 -n vmname -d 172.31.2.1 -N 172.31.2.201 Notice the double spaces before -d and -N (and the extra space at the EOL) -- WARNING: you have to view the source of this description to actually _see_ the double spaces!) After patching /usr/lib64/cloud/agent/scripts/network/domr/dhcp_entry.sh to do meaningful logging, it's clear that the script does not get called with -d, but -d instead (same for -N), so with an extra space before the dash. Thus, getopts fails to parse/recognize these two arguments correctly and passed empty values for $dfltrt and $dns to the /root/edithosts.sh being called on the routerVM. It's also clear that the CloudStack KVM Java agent calls it with the wrongly constructed command line, because if a shell would interpret this command line, it would just ignore the extra spaces itself. I've not been able to dig it down, but I somehow suspect that one of ./utils/src/com/cloud/utils/script/Script.java:protected String buildCommandLine(String[] command) { } ./utils/src/com/cloud/utils/script/Script.java:public String execute(OutputInterpreter interpreter) { } might mess up building the command line of the command that had been built by ./core/src/com/cloud/agent/resource/virtualnetwork/VirtualRoutingResource.java:protected synchronized Answer execute (final DhcpEntryCommand cmd) { } before. I've not tried 4.0.0 so far, thus I cannot say whether it might be affected or not. As a workaround, I've patched dhcp_entry.sh to re-evaluate the positional parameters using 'set -- ${@}' (will attach a patch, also one for logging). -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira
Reminder: IRC Meeting Today
Howdy all, Reminder - IRC meeting today at 17:00 UTC, which is 12:00 Eastern, 09:00 Pacific. We meet in #cloudstack-meeting on Freenode. If you're new to the meetings, remember that the purpose for real-time meetings is for status reports and discussion only - topics that require a decision are always discussed and decided on the mailing list. We use Meetbot to keep meeting minutes. If you aren't familiar with Meetbot, the guide can be found here: http://meetbot.debian.net/Manual.html Best, jzb -- Joe Brockmeier http://dissociatedpress.net/ Twitter: @jzb
Re: [DISCUSS] changing the format of commit message subject lines
On Tue, Dec 04, 2012 at 03:35:48PM -0500, David Nalley wrote: Hi folks, I'd like to propose that we change the commit message subject line from: git commit: $short_commit_msg to: git commit: $branch - $short_commit_msg Any other opinions? Works for me. Would be much easier to see what's going where just by skimming the messages... +1 Best, jzb -- Joe Brockmeier http://dissociatedpress.net/ Twitter: @jzb
[jira] [Commented] (CLOUDSTACK-460) Cloudstack 4.0 Cannot ssh into system vms
[ https://issues.apache.org/jira/browse/CLOUDSTACK-460?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13510569#comment-13510569 ] Joe Brockmeier commented on CLOUDSTACK-460: --- There's time! I've added the commit from master. It's: 2f99c25a046b39f8377fb2cdacaedb7404a59481 Cloudstack 4.0 Cannot ssh into system vms - Key: CLOUDSTACK-460 URL: https://issues.apache.org/jira/browse/CLOUDSTACK-460 Project: CloudStack Issue Type: Bug Security Level: Public(Anyone can view this level - this is the default.) Components: Install and Setup, KVM Affects Versions: 4.0.0 Environment: Ubuntu 12.04 Reporter: jason Labels: newbie Fix For: 4.0.0, 4.0.1 Original Estimate: 168h Remaining Estimate: 168h Cannot ssh into system vms, as far as i can see the the keys match up and im also using the Link Local IP Adddress. I Have reinstalled 3 times to make sure its not user error please help thanks jason -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Created] (CLOUDSTACK-586) exportfs -a gives 'no_subtree_check warning' when following the Installation guide
Max Thoursie created CLOUDSTACK-586: --- Summary: exportfs -a gives 'no_subtree_check warning' when following the Installation guide Key: CLOUDSTACK-586 URL: https://issues.apache.org/jira/browse/CLOUDSTACK-586 Project: CloudStack Issue Type: Bug Security Level: Public (Anyone can view this level - this is the default.) Components: Doc Affects Versions: 4.0.0 Environment: Ubuntu 12.04.1 LTS (GNU/Linux 3.2.0-29-generic x86_64) Reporter: Max Thoursie Priority: Minor The installation guide says, under 4.5.5.1 and 4.5.5.2 2. To configure the new directories as NFS exports, edit /etc/exports. Export the NFS share(s) with rw,async,no_root_squash. For example: # vi /etc/exports Insert the following line. /export *(rw,async,no_root_squash) Following these instructions, Ubuntu prints the following warning max@ubuntu:~$ sudo exportfs -a exportfs: /etc/exports [1]: Neither 'subtree_check' or 'no_subtree_check' specified for export *:/export. Assuming default behaviour ('no_subtree_check'). NOTE: this default has changed since nfs-utils version 1.0.x -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira
Re: SSVM Network Configuration Issue
All, I was wondering if anyone else is experiencing this problem when using secondary storage on a devcloud-style VM with a host-only and NAT adapter. One aspect of this issue that seems interesting is that following route table from the SSVM: root@s-5-TEST:~# route Kernel IP routing table Destination Gateway Genmask Flags Metric RefUse Iface 10.0.3.2192.168.56.1255.255.255.255 UGH 0 00 eth1 10.0.3.0* 255.255.255.0 U 0 00 eth2 192.168.56.0* 255.255.255.0 U 0 00 eth1 192.168.56.0* 255.255.255.0 U 0 00 eth3 link-local * 255.255.0.0 U 0 00 eth0 default 10.0.3.20.0.0.0 UG0 00 eth2 In particular, the gateways for the management and guest networks do not match to the configuration provided to the management server (i.e. 10.0.3.2 is the gateway for the 10.0.3.0/24 network and 192.168.56.1 is the gateway for the 192.168.56.0/24 network). With this configuration, the SSVM has a socket connection to the management server, but is in alert state. Finally, when I remove the host-only NIC and use only a NAT adapter the SSVM's networking works as expecting leading me to believe that the segregated network configuration is at the root of the problem. Until I can get the networking on the SSVM configured, I am unable to complete the testing of the S3-backed Secondary Storage enhancement. Thank you for your help, -John On Dec 3, 2012, at 4:46 PM, John Burwell jburw...@basho.com wrote: All, I am setting up a multi-zone devcloud configuration on VirtualBox 4.2.4 using the Ubuntu 12.04.1 and Xen 4.1. I have configured the base management server VM (zone1) to serve as both the zone1, as well as, the management server (running MySql) with eth0 as a host-only adapter and a static IP of 192.168.56.15 and eth1 as a NAT adapter (see the attached zone1-interfaces file for the exact network configuration on the VM). The management and guest networks are configured as follows: Zone 1 Management: 192.168.56.100-149 gw 192.168.56.1 dns 10.0.3.2 (?) Guest: 10.0.3.200-10.0.3.220 gw 10.0.3.2 dns 8.8.8.8 Zone 2 Management: 192.168.56.150-200 gw 192.68.56.1 dns 10.0.3.2 (?) Guest: 10.0.3.221-240 gw 10.0.3.2 dns 8.8.8.8 The management server deploys and starts without error. I then populate the configuration it using the attached Marvin configuration file (zone1.devcloud.cfg) and restart the management server in order to allow the global configuration option changes to take effect. Following the restart, the CPVM and SSVM start without error. Unfortunately, they drop into alert status, and the SSVM is unable to connect outbound through the guest network (very important for my tests because I am testing S3-backed secondary storage). From the diagnostic checks I have performed on the management server and the SSVM, it appears that the daemon on the SSVM is connecting back to the management server. I have attached a set of diagnostic information from the management server (mgmtsvr-zone1-diagnostics.log) and SSVM server (ssvm-zone1-diagnostics.log) that includes the results of ifconfig, route, netstat and ping checks, as well as, other information (e.g. the contents of /var/cache/cloud/cmdline on the SSVM). Finally, I have attached the vmops log from the management server (vmops-zone1.log). What changes need to be made to management server configuration in order to start up an SSVM that can communicate with the secondary storage NFS volumes, management server, and connect to hosts on the Internet? Thanks for your help, -John ssvm-zone1-diagnostics.log vmops-zone1.tar.gz mgmtsvr-zone1-diagnostics.log zone1-interfaces zone1.devcloud.cfg
Re: SSVM Network Configuration Issue
I think it may have something to do with your DNS. I've noticed that it drops routes in for my DNS servers, forcing them to use the management network. So in this case perhaps you have your DNS configured as 10.0.3.2, so it forces that to use eth1... just a guess. If thats the issue (or something like that) then you'll want to set up a dnsmasq service on on the 192.168 network to use, or whatever to keep your DNS from routing. On Wed, Dec 5, 2012 at 10:10 AM, John Burwell jburw...@basho.com wrote: All, I was wondering if anyone else is experiencing this problem when using secondary storage on a devcloud-style VM with a host-only and NAT adapter. One aspect of this issue that seems interesting is that following route table from the SSVM: root@s-5-TEST:~# route Kernel IP routing table Destination Gateway Genmask Flags Metric RefUse Iface 10.0.3.2192.168.56.1255.255.255.255 UGH 0 00 eth1 10.0.3.0* 255.255.255.0 U 0 00 eth2 192.168.56.0* 255.255.255.0 U 0 00 eth1 192.168.56.0* 255.255.255.0 U 0 00 eth3 link-local * 255.255.0.0 U 0 00 eth0 default 10.0.3.20.0.0.0 UG0 00 eth2 In particular, the gateways for the management and guest networks do not match to the configuration provided to the management server (i.e. 10.0.3.2 is the gateway for the 10.0.3.0/24 network and 192.168.56.1 is the gateway for the 192.168.56.0/24 network). With this configuration, the SSVM has a socket connection to the management server, but is in alert state. Finally, when I remove the host-only NIC and use only a NAT adapter the SSVM's networking works as expecting leading me to believe that the segregated network configuration is at the root of the problem. Until I can get the networking on the SSVM configured, I am unable to complete the testing of the S3-backed Secondary Storage enhancement. Thank you for your help, -John On Dec 3, 2012, at 4:46 PM, John Burwell jburw...@basho.com wrote: All, I am setting up a multi-zone devcloud configuration on VirtualBox 4.2.4 using the Ubuntu 12.04.1 and Xen 4.1. I have configured the base management server VM (zone1) to serve as both the zone1, as well as, the management server (running MySql) with eth0 as a host-only adapter and a static IP of 192.168.56.15 and eth1 as a NAT adapter (see the attached zone1-interfaces file for the exact network configuration on the VM). The management and guest networks are configured as follows: Zone 1 Management: 192.168.56.100-149 gw 192.168.56.1 dns 10.0.3.2 (?) Guest: 10.0.3.200-10.0.3.220 gw 10.0.3.2 dns 8.8.8.8 Zone 2 Management: 192.168.56.150-200 gw 192.68.56.1 dns 10.0.3.2 (?) Guest: 10.0.3.221-240 gw 10.0.3.2 dns 8.8.8.8 The management server deploys and starts without error. I then populate the configuration it using the attached Marvin configuration file (zone1.devcloud.cfg) and restart the management server in order to allow the global configuration option changes to take effect. Following the restart, the CPVM and SSVM start without error. Unfortunately, they drop into alert status, and the SSVM is unable to connect outbound through the guest network (very important for my tests because I am testing S3-backed secondary storage). From the diagnostic checks I have performed on the management server and the SSVM, it appears that the daemon on the SSVM is connecting back to the management server. I have attached a set of diagnostic information from the management server (mgmtsvr-zone1-diagnostics.log) and SSVM server (ssvm-zone1-diagnostics.log) that includes the results of ifconfig, route, netstat and ping checks, as well as, other information (e.g. the contents of /var/cache/cloud/cmdline on the SSVM). Finally, I have attached the vmops log from the management server (vmops-zone1.log). What changes need to be made to management server configuration in order to start up an SSVM that can communicate with the secondary storage NFS volumes, management server, and connect to hosts on the Internet? Thanks for your help, -John ssvm-zone1-diagnostics.log vmops-zone1.tar.gz mgmtsvr-zone1-diagnostics.log zone1-interfaces zone1.devcloud.cfg
[jira] [Commented] (CLOUDSTACK-585) DHCP entry provisioning is broken in the KVM agent
[ https://issues.apache.org/jira/browse/CLOUDSTACK-585?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13510614#comment-13510614 ] Anthony Xu commented on CLOUDSTACK-585: --- Hi wolfram, I can't reproduce the issue, see following example, I put double spaces before -d and -N, at the end of the output, + ssh -p 3922 -o StrictHostKeyChecking=no -i /root/.ssh/id_rsa.cloud root@169.254.2.144 '/root/edithosts.sh 06:55:22:00:00:11 10.223.170.23 anthony 10.223.170.1 10.223.110.254 ' looks good to me and I checked /etc/dhcpopts.txt in route VM, gateway and NDS is pragrammed correctly 10_223_170_23,3,10.223.170.1 10_223_170_23,6,10.223.110.254 I executed the bellow command in XenServer host, KVM might be different, cloud you do the similar test in KVM and post the output? Anthony bash -x dhcp_entry.sh -r 169.254.2.144 -v 10.223.170.23 -m 06:55:22:00:00:11 -n anthony -d 10.223.170.1 -N 10.223.110.254 + cert=/root/.ssh/id_rsa.cloud + domrIp= + vmMac= + vmIp= + vmName= + staticrt= + dfltrt= + dns= + getopts r:m:v:n:d:s:N: OPTION + case $OPTION in + domrIp=169.254.2.144 + getopts r:m:v:n:d:s:N: OPTION + case $OPTION in + vmIp=10.223.170.23 + getopts r:m:v:n:d:s:N: OPTION + case $OPTION in + vmMac=06:55:22:00:00:11 + getopts r:m:v:n:d:s:N: OPTION + case $OPTION in + vmName=anthony + getopts r:m:v:n:d:s:N: OPTION + case $OPTION in + dfltrt=10.223.170.1 + getopts r:m:v:n:d:s:N: OPTION + case $OPTION in + dns=10.223.110.254 + getopts r:m:v:n:d:s:N: OPTION + add_dhcp_entry 169.254.2.144 06:55:22:00:00:11 10.223.170.23 anthony 10.223.170.1 10.223.110.254 + local domr=169.254.2.144 + local mac=06:55:22:00:00:11 + local ip=10.223.170.23 + local vm=anthony + local dfltrt=10.223.170.1 + local ns=10.223.110.254 + local staticrt= + ssh -p 3922 -o StrictHostKeyChecking=no -i /root/.ssh/id_rsa.cloud root@169.254.2.144 '/root/edithosts.sh 06:55:22:00:00:11 10.223.170.23 anthony 10.223.170.1 10.223.110.254 ' + return 0 + exit 0 DHCP entry provisioning is broken in the KVM agent -- Key: CLOUDSTACK-585 URL: https://issues.apache.org/jira/browse/CLOUDSTACK-585 Project: CloudStack Issue Type: Bug Security Level: Public(Anyone can view this level - this is the default.) Components: Hypervisor Controller, KVM Environment: CloudStack 3.0.2 KVM agent running on Fedora 14 Reporter: Wolfram Schlich Labels: agent, kvm Attachments: 1-dhcp_entry.sh.log.patch, 2-dhcp_entry.sh.fix.patch When adding an instance to a routerVM DHCP configuration, it seems that the KVM agent calls /usr/lib64/cloud/agent/scripts/network/domr/dhcp_entry.sh with wrongly constructed command line arguments, making the script fail to add correct entries (specifically default router, DNS servers and static routes) for that instance to the routerVM's /etc/dhcphosts.txt + /etc/dhcpopts.txt. Especially adding a specific default gateway fails, so the routerVM will always announce itself as the default router, because the correct entry in /etc/dhcpopts specifying the gateway of the instance's default network as the gateway is missing. This is especially nasty for non-default/additional networks of an instance, messing up the default routing. Examples: Management server log entry: 2012-11-29 14:36:37,764 DEBUG [resource.virtualnetwork.VirtualRoutingResource] (agentRequest-Handler-3:null) Executing: /usr/lib64/cloud/agent/./scripts/network/domr/dhcp_entry.sh -r 169.254.0.122 -v 172.31.2.233 -m 06:b5:88:00:02:30 -n vmname -d 172.31.2.1 -N 172.31.2.201 Notice the double spaces before -d and -N (and the extra space at the EOL) -- WARNING: you have to view the source of this description to actually _see_ the double spaces!) After patching /usr/lib64/cloud/agent/scripts/network/domr/dhcp_entry.sh to do meaningful logging, it's clear that the script does not get called with -d, but -d instead (same for -N), so with an extra space before the dash. Thus, getopts fails to parse/recognize these two arguments correctly and passed empty values for $dfltrt and $dns to the /root/edithosts.sh being called on the routerVM. It's also clear that the CloudStack KVM Java agent calls it with the wrongly constructed command line, because if a shell would interpret this command line, it would just ignore the extra spaces itself. I've not been able to dig it down, but I somehow suspect that one of ./utils/src/com/cloud/utils/script/Script.java:protected String buildCommandLine(String[] command) { } ./utils/src/com/cloud/utils/script/Script.java:public String execute(OutputInterpreter interpreter) { } might mess up building the command line of the command that had been built by
[jira] [Updated] (CLOUDSTACK-585) DHCP entry provisioning is broken in the KVM agent
[ https://issues.apache.org/jira/browse/CLOUDSTACK-585?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Wolfram Schlich updated CLOUDSTACK-585: --- Description: When adding an instance to a routerVM DHCP configuration, it seems that the KVM agent calls /usr/lib64/cloud/agent/scripts/network/domr/dhcp_entry.sh with wrongly constructed command line arguments, making the script fail to add correct entries (specifically default router, DNS servers and static routes) for that instance to the routerVM's /etc/dhcphosts.txt + /etc/dhcpopts.txt. Especially adding a specific default gateway fails, so the routerVM will always announce itself as the default router, because the correct entry in /etc/dhcpopts specifying the gateway of the instance's default network as the gateway is missing. This is especially nasty for non-default/additional networks of an instance, messing up the default routing. Examples: Management server log entry: 2012-11-29 14:36:37,764 DEBUG [resource.virtualnetwork.VirtualRoutingResource] (agentRequest-Handler-3:null) Executing: /usr/lib64/cloud/agent/./scripts/network/domr/dhcp_entry.sh -r 169.254.0.122 -v 172.31.2.233 -m 06:b5:88:00:02:30 -n vmname -d 172.31.2.1 -N 172.31.2.201 Notice the double spaces before -d and -N (and the extra space at the EOL) -- WARNING: you have to view the source of this description to actually _see_ the double spaces!) After patching /usr/lib64/cloud/agent/scripts/network/domr/dhcp_entry.sh to do meaningful logging, it's clear that the script does not get called with -d, but -d instead (same for -N), so with an extra space before the dash. Thus, getopts fails to parse/recognize these two arguments correctly and passed empty values for $dfltrt and $dns to the /root/edithosts.sh being called on the routerVM. It's also clear that the CloudStack KVM Java agent calls it with the wrongly constructed command line, because if a shell would interpret this command line, it would just ignore the extra spaces itself. I've not been able to dig it down, but I somehow suspect that one of ./utils/src/com/cloud/utils/script/Script.java:protected String buildCommandLine(String[] command) { } ./utils/src/com/cloud/utils/script/Script.java:public String execute(OutputInterpreter interpreter) { } might mess up building the command line of the command that had been built by ./core/src/com/cloud/agent/resource/virtualnetwork/VirtualRoutingResource.java:protected synchronized Answer execute (final DhcpEntryCommand cmd) { } before. I've not tried 4.0.0 so far, thus I cannot say whether it might be affected or not. As a workaround, I've patched dhcp_entry.sh to re-evaluate the positional parameters using 'set -- ${@}' (will attach a patch, also one for logging). was: DHCP entry provisioning is broken in the KVM agent -- Key: CLOUDSTACK-585 URL: https://issues.apache.org/jira/browse/CLOUDSTACK-585 Project: CloudStack Issue Type: Bug Security Level: Public(Anyone can view this level - this is the default.) Components: Hypervisor Controller, KVM Environment: CloudStack 3.0.2 KVM agent running on Fedora 14 Reporter: Wolfram Schlich Labels: agent, kvm Attachments: 1-dhcp_entry.sh.log.patch, 2-dhcp_entry.sh.fix.patch When adding an instance to a routerVM DHCP configuration, it seems that the KVM agent calls /usr/lib64/cloud/agent/scripts/network/domr/dhcp_entry.sh with wrongly constructed command line arguments, making the script fail to add correct entries (specifically default router, DNS servers and static routes) for that instance to the routerVM's /etc/dhcphosts.txt + /etc/dhcpopts.txt. Especially adding a specific default gateway fails, so the routerVM will always announce itself as the default router, because the correct entry in /etc/dhcpopts specifying the gateway of the instance's default network as the gateway is missing. This is especially nasty for non-default/additional networks of an instance, messing up the default routing. Examples: Management server log entry: 2012-11-29 14:36:37,764 DEBUG [resource.virtualnetwork.VirtualRoutingResource] (agentRequest-Handler-3:null) Executing: /usr/lib64/cloud/agent/./scripts/network/domr/dhcp_entry.sh -r 169.254.0.122 -v 172.31.2.233 -m 06:b5:88:00:02:30 -n vmname -d 172.31.2.1 -N 172.31.2.201 Notice the double spaces before -d and -N (and the extra space at the EOL) -- WARNING: you have to view the source of this description to actually _see_ the double spaces!) After patching /usr/lib64/cloud/agent/scripts/network/domr/dhcp_entry.sh to do meaningful logging, it's clear that the script does not get called with -d, but -d instead (same for -N), so with an extra space before the dash. Thus, getopts fails to
[jira] [Updated] (CLOUDSTACK-585) DHCP entry provisioning is broken in the KVM agent
[ https://issues.apache.org/jira/browse/CLOUDSTACK-585?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Wolfram Schlich updated CLOUDSTACK-585: --- Description:(was: When adding an instance to a routerVM DHCP configuration, it seems that the KVM agent calls /usr/lib64/cloud/agent/scripts/network/domr/dhcp_entry.sh with wrongly constructed command line arguments, making the script fail to add correct entries (specifically default router, DNS servers and static routes) for that instance to the routerVM's /etc/dhcphosts.txt + /etc/dhcpopts.txt. Especially adding a specific default gateway fails, so the routerVM will always announce itself as the default router, because the correct entry in /etc/dhcpopts specifying the gateway of the instance's default network as the gateway is missing. This is especially nasty for non-default/additional networks of an instance, messing up the default routing. Examples: Management server log entry: 2012-11-29 14:36:37,764 DEBUG [resource.virtualnetwork.VirtualRoutingResource] (agentRequest-Handler-3:null) Executing: /usr/lib64/cloud/agent/./scripts/network/domr/dhcp_entry.sh -r 169.254.0.122 -v 172.31.2.233 -m 06:b5:88:00:02:30 -n vmname -d 172.31.2.1 -N 172.31.2.201 Notice the double spaces before -d and -N (and the extra space at the EOL) -- WARNING: you have to view the source of this description to actually _see_ the double spaces!) After patching /usr/lib64/cloud/agent/scripts/network/domr/dhcp_entry.sh to do meaningful logging, it's clear that the script does not get called with -d, but -d instead (same for -N), so with an extra space before the dash. Thus, getopts fails to parse/recognize these two arguments correctly and passed empty values for $dfltrt and $dns to the /root/edithosts.sh being called on the routerVM. It's also clear that the CloudStack KVM Java agent calls it with the wrongly constructed command line, because if a shell would interpret this command line, it would just ignore the extra spaces itself. I've not been able to dig it down, but I somehow suspect that one of ./utils/src/com/cloud/utils/script/Script.java:protected String buildCommandLine(String[] command) { } ./utils/src/com/cloud/utils/script/Script.java:public String execute(OutputInterpreter interpreter) { } might mess up building the command line of the command that had been built by ./core/src/com/cloud/agent/resource/virtualnetwork/VirtualRoutingResource.java:protected synchronized Answer execute (final DhcpEntryCommand cmd) { } before. I've not tried 4.0.0 so far, thus I cannot say whether it might be affected or not. As a workaround, I've patched dhcp_entry.sh to re-evaluate the positional parameters using 'set -- ${@}' (will attach a patch, also one for logging).) DHCP entry provisioning is broken in the KVM agent -- Key: CLOUDSTACK-585 URL: https://issues.apache.org/jira/browse/CLOUDSTACK-585 Project: CloudStack Issue Type: Bug Security Level: Public(Anyone can view this level - this is the default.) Components: Hypervisor Controller, KVM Environment: CloudStack 3.0.2 KVM agent running on Fedora 14 Reporter: Wolfram Schlich Labels: agent, kvm Attachments: 1-dhcp_entry.sh.log.patch, 2-dhcp_entry.sh.fix.patch -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira
RE: SSVM Network Configuration Issue
Hi join, Try following, Set global configuration management.network.cidr to your management server CIDR, if this configuration is not available in UI, you can change it in DB directly. Restart management, Stop/Start SSVM and CPVM. And could you post cat /proc/cmdline in SSVM? Anthony -Original Message- From: John Burwell [mailto:jburw...@basho.com] Sent: Wednesday, December 05, 2012 9:11 AM To: cloudstack-dev@incubator.apache.org Subject: Re: SSVM Network Configuration Issue All, I was wondering if anyone else is experiencing this problem when using secondary storage on a devcloud-style VM with a host-only and NAT adapter. One aspect of this issue that seems interesting is that following route table from the SSVM: root@s-5-TEST:~# route Kernel IP routing table Destination Gateway Genmask Flags Metric RefUse Iface 10.0.3.2192.168.56.1255.255.255.255 UGH 0 00 eth1 10.0.3.0* 255.255.255.0 U 0 00 eth2 192.168.56.0* 255.255.255.0 U 0 00 eth1 192.168.56.0* 255.255.255.0 U 0 00 eth3 link-local * 255.255.0.0 U 0 00 eth0 default 10.0.3.20.0.0.0 UG0 00 eth2 In particular, the gateways for the management and guest networks do not match to the configuration provided to the management server (i.e. 10.0.3.2 is the gateway for the 10.0.3.0/24 network and 192.168.56.1 is the gateway for the 192.168.56.0/24 network). With this configuration, the SSVM has a socket connection to the management server, but is in alert state. Finally, when I remove the host-only NIC and use only a NAT adapter the SSVM's networking works as expecting leading me to believe that the segregated network configuration is at the root of the problem. Until I can get the networking on the SSVM configured, I am unable to complete the testing of the S3-backed Secondary Storage enhancement. Thank you for your help, -John On Dec 3, 2012, at 4:46 PM, John Burwell jburw...@basho.com wrote: All, I am setting up a multi-zone devcloud configuration on VirtualBox 4.2.4 using the Ubuntu 12.04.1 and Xen 4.1. I have configured the base management server VM (zone1) to serve as both the zone1, as well as, the management server (running MySql) with eth0 as a host-only adapter and a static IP of 192.168.56.15 and eth1 as a NAT adapter (see the attached zone1-interfaces file for the exact network configuration on the VM). The management and guest networks are configured as follows: Zone 1 Management: 192.168.56.100-149 gw 192.168.56.1 dns 10.0.3.2 (?) Guest: 10.0.3.200-10.0.3.220 gw 10.0.3.2 dns 8.8.8.8 Zone 2 Management: 192.168.56.150-200 gw 192.68.56.1 dns 10.0.3.2 (?) Guest: 10.0.3.221-240 gw 10.0.3.2 dns 8.8.8.8 The management server deploys and starts without error. I then populate the configuration it using the attached Marvin configuration file (zone1.devcloud.cfg) and restart the management server in order to allow the global configuration option changes to take effect. Following the restart, the CPVM and SSVM start without error. Unfortunately, they drop into alert status, and the SSVM is unable to connect outbound through the guest network (very important for my tests because I am testing S3-backed secondary storage). From the diagnostic checks I have performed on the management server and the SSVM, it appears that the daemon on the SSVM is connecting back to the management server. I have attached a set of diagnostic information from the management server (mgmtsvr-zone1-diagnostics.log) and SSVM server (ssvm-zone1-diagnostics.log) that includes the results of ifconfig, route, netstat and ping checks, as well as, other information (e.g. the contents of /var/cache/cloud/cmdline on the SSVM). Finally, I have attached the vmops log from the management server (vmops-zone1.log). What changes need to be made to management server configuration in order to start up an SSVM that can communicate with the secondary storage NFS volumes, management server, and connect to hosts on the Internet? Thanks for your help, -John ssvm-zone1-diagnostics.log vmops-zone1.tar.gz mgmtsvr-zone1-diagnostics.log zone1-interfaces zone1.devcloud.cfg
[jira] [Commented] (CLOUDSTACK-585) DHCP entry provisioning is broken in the KVM agent
[ https://issues.apache.org/jira/browse/CLOUDSTACK-585?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13510624#comment-13510624 ] Wolfram Schlich commented on CLOUDSTACK-585: Crappy auto-update bs when you click into a field! Argh. Anyway. Anthony, thanks for your response. Unfortunately, you didn't get the point at all :-) Please re-read the description. I don't know how to make it even more clear than I already did. Thanks. DHCP entry provisioning is broken in the KVM agent -- Key: CLOUDSTACK-585 URL: https://issues.apache.org/jira/browse/CLOUDSTACK-585 Project: CloudStack Issue Type: Bug Security Level: Public(Anyone can view this level - this is the default.) Components: Hypervisor Controller, KVM Environment: CloudStack 3.0.2 KVM agent running on Fedora 14 Reporter: Wolfram Schlich Labels: agent, kvm Attachments: 1-dhcp_entry.sh.log.patch, 2-dhcp_entry.sh.fix.patch When adding an instance to a routerVM DHCP configuration, it seems that the KVM agent calls /usr/lib64/cloud/agent/scripts/network/domr/dhcp_entry.sh with wrongly constructed command line arguments, making the script fail to add correct entries (specifically default router, DNS servers and static routes) for that instance to the routerVM's /etc/dhcphosts.txt + /etc/dhcpopts.txt. Especially adding a specific default gateway fails, so the routerVM will always announce itself as the default router, because the correct entry in /etc/dhcpopts specifying the gateway of the instance's default network as the gateway is missing. This is especially nasty for non-default/additional networks of an instance, messing up the default routing. Examples: Management server log entry: 2012-11-29 14:36:37,764 DEBUG [resource.virtualnetwork.VirtualRoutingResource] (agentRequest-Handler-3:null) Executing: /usr/lib64/cloud/agent/./scripts/network/domr/dhcp_entry.sh -r 169.254.0.122 -v 172.31.2.233 -m 06:b5:88:00:02:30 -n vmname -d 172.31.2.1 -N 172.31.2.201 Notice the double spaces before -d and -N (and the extra space at the EOL) -- WARNING: you have to view the source of this description to actually _see_ the double spaces!) After patching /usr/lib64/cloud/agent/scripts/network/domr/dhcp_entry.sh to do meaningful logging, it's clear that the script does not get called with -d, but -d instead (same for -N), so with an extra space before the dash. Thus, getopts fails to parse/recognize these two arguments correctly and passed empty values for $dfltrt and $dns to the /root/edithosts.sh being called on the routerVM. It's also clear that the CloudStack KVM Java agent calls it with the wrongly constructed command line, because if a shell would interpret this command line, it would just ignore the extra spaces itself. I've not been able to dig it down, but I somehow suspect that one of ./utils/src/com/cloud/utils/script/Script.java:protected String buildCommandLine(String[] command) { } ./utils/src/com/cloud/utils/script/Script.java:public String execute(OutputInterpreter interpreter) { } might mess up building the command line of the command that had been built by ./core/src/com/cloud/agent/resource/virtualnetwork/VirtualRoutingResource.java:protected synchronized Answer execute (final DhcpEntryCommand cmd) { } before. I've not tried 4.0.0 so far, thus I cannot say whether it might be affected or not. As a workaround, I've patched dhcp_entry.sh to re-evaluate the positional parameters using 'set -- ${@}' (will attach a patch, also one for logging). -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Commented] (CLOUDSTACK-585) DHCP entry provisioning is broken in the KVM agent
[ https://issues.apache.org/jira/browse/CLOUDSTACK-585?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13510628#comment-13510628 ] Wolfram Schlich commented on CLOUDSTACK-585: Ok I'll try: You cannot reproduce it by running the command on the shell as you did! You have to let the KVM agent do that task (call dhcp_entry.sh). You can only simulate how the KVM agent does it by also quoting the arguments wrong on a shell: bash -x dhcp_entry.sh -r 169.254.2.144 -v 10.223.170.23 -m 06:55:22:00:00:11 -n anthony -d 10.223.170.1 -N 10.223.110.254 But that doesn't help you at all I believe. DHCP entry provisioning is broken in the KVM agent -- Key: CLOUDSTACK-585 URL: https://issues.apache.org/jira/browse/CLOUDSTACK-585 Project: CloudStack Issue Type: Bug Security Level: Public(Anyone can view this level - this is the default.) Components: Hypervisor Controller, KVM Environment: CloudStack 3.0.2 KVM agent running on Fedora 14 Reporter: Wolfram Schlich Labels: agent, kvm Attachments: 1-dhcp_entry.sh.log.patch, 2-dhcp_entry.sh.fix.patch When adding an instance to a routerVM DHCP configuration, it seems that the KVM agent calls /usr/lib64/cloud/agent/scripts/network/domr/dhcp_entry.sh with wrongly constructed command line arguments, making the script fail to add correct entries (specifically default router, DNS servers and static routes) for that instance to the routerVM's /etc/dhcphosts.txt + /etc/dhcpopts.txt. Especially adding a specific default gateway fails, so the routerVM will always announce itself as the default router, because the correct entry in /etc/dhcpopts specifying the gateway of the instance's default network as the gateway is missing. This is especially nasty for non-default/additional networks of an instance, messing up the default routing. Examples: Management server log entry: 2012-11-29 14:36:37,764 DEBUG [resource.virtualnetwork.VirtualRoutingResource] (agentRequest-Handler-3:null) Executing: /usr/lib64/cloud/agent/./scripts/network/domr/dhcp_entry.sh -r 169.254.0.122 -v 172.31.2.233 -m 06:b5:88:00:02:30 -n vmname -d 172.31.2.1 -N 172.31.2.201 Notice the double spaces before -d and -N (and the extra space at the EOL) -- WARNING: you have to view the source of this description to actually _see_ the double spaces!) After patching /usr/lib64/cloud/agent/scripts/network/domr/dhcp_entry.sh to do meaningful logging, it's clear that the script does not get called with -d, but -d instead (same for -N), so with an extra space before the dash. Thus, getopts fails to parse/recognize these two arguments correctly and passed empty values for $dfltrt and $dns to the /root/edithosts.sh being called on the routerVM. It's also clear that the CloudStack KVM Java agent calls it with the wrongly constructed command line, because if a shell would interpret this command line, it would just ignore the extra spaces itself. I've not been able to dig it down, but I somehow suspect that one of ./utils/src/com/cloud/utils/script/Script.java:protected String buildCommandLine(String[] command) { } ./utils/src/com/cloud/utils/script/Script.java:public String execute(OutputInterpreter interpreter) { } might mess up building the command line of the command that had been built by ./core/src/com/cloud/agent/resource/virtualnetwork/VirtualRoutingResource.java:protected synchronized Answer execute (final DhcpEntryCommand cmd) { } before. I've not tried 4.0.0 so far, thus I cannot say whether it might be affected or not. As a workaround, I've patched dhcp_entry.sh to re-evaluate the positional parameters using 'set -- ${@}' (will attach a patch, also one for logging). -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Commented] (CLOUDSTACK-585) DHCP entry provisioning is broken in the KVM agent
[ https://issues.apache.org/jira/browse/CLOUDSTACK-585?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13510634#comment-13510634 ] Wolfram Schlich commented on CLOUDSTACK-585: The issue is NOT with dhcp_entry.sh, but with the command line construction that the KVM agent uses to execute it! DHCP entry provisioning is broken in the KVM agent -- Key: CLOUDSTACK-585 URL: https://issues.apache.org/jira/browse/CLOUDSTACK-585 Project: CloudStack Issue Type: Bug Security Level: Public(Anyone can view this level - this is the default.) Components: Hypervisor Controller, KVM Environment: CloudStack 3.0.2 KVM agent running on Fedora 14 Reporter: Wolfram Schlich Labels: agent, kvm Attachments: 1-dhcp_entry.sh.log.patch, 2-dhcp_entry.sh.fix.patch When adding an instance to a routerVM DHCP configuration, it seems that the KVM agent calls /usr/lib64/cloud/agent/scripts/network/domr/dhcp_entry.sh with wrongly constructed command line arguments, making the script fail to add correct entries (specifically default router, DNS servers and static routes) for that instance to the routerVM's /etc/dhcphosts.txt + /etc/dhcpopts.txt. Especially adding a specific default gateway fails, so the routerVM will always announce itself as the default router, because the correct entry in /etc/dhcpopts specifying the gateway of the instance's default network as the gateway is missing. This is especially nasty for non-default/additional networks of an instance, messing up the default routing. Examples: Management server log entry: 2012-11-29 14:36:37,764 DEBUG [resource.virtualnetwork.VirtualRoutingResource] (agentRequest-Handler-3:null) Executing: /usr/lib64/cloud/agent/./scripts/network/domr/dhcp_entry.sh -r 169.254.0.122 -v 172.31.2.233 -m 06:b5:88:00:02:30 -n vmname -d 172.31.2.1 -N 172.31.2.201 Notice the double spaces before -d and -N (and the extra space at the EOL) -- WARNING: you have to view the source of this description to actually _see_ the double spaces!) After patching /usr/lib64/cloud/agent/scripts/network/domr/dhcp_entry.sh to do meaningful logging, it's clear that the script does not get called with -d, but -d instead (same for -N), so with an extra space before the dash. Thus, getopts fails to parse/recognize these two arguments correctly and passed empty values for $dfltrt and $dns to the /root/edithosts.sh being called on the routerVM. It's also clear that the CloudStack KVM Java agent calls it with the wrongly constructed command line, because if a shell would interpret this command line, it would just ignore the extra spaces itself. I've not been able to dig it down, but I somehow suspect that one of ./utils/src/com/cloud/utils/script/Script.java:protected String buildCommandLine(String[] command) { } ./utils/src/com/cloud/utils/script/Script.java:public String execute(OutputInterpreter interpreter) { } might mess up building the command line of the command that had been built by ./core/src/com/cloud/agent/resource/virtualnetwork/VirtualRoutingResource.java:protected synchronized Answer execute (final DhcpEntryCommand cmd) { } before. I've not tried 4.0.0 so far, thus I cannot say whether it might be affected or not. As a workaround, I've patched dhcp_entry.sh to re-evaluate the positional parameters using 'set -- ${@}' (will attach a patch, also one for logging). -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Commented] (CLOUDSTACK-585) DHCP entry provisioning is broken in the KVM agent
[ https://issues.apache.org/jira/browse/CLOUDSTACK-585?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13510639#comment-13510639 ] Anthony Xu commented on CLOUDSTACK-585: --- I can reproduce the issue with quote, and your fix works:-) please check in your fix. DHCP entry provisioning is broken in the KVM agent -- Key: CLOUDSTACK-585 URL: https://issues.apache.org/jira/browse/CLOUDSTACK-585 Project: CloudStack Issue Type: Bug Security Level: Public(Anyone can view this level - this is the default.) Components: Hypervisor Controller, KVM Environment: CloudStack 3.0.2 KVM agent running on Fedora 14 Reporter: Wolfram Schlich Labels: agent, kvm Attachments: 1-dhcp_entry.sh.log.patch, 2-dhcp_entry.sh.fix.patch When adding an instance to a routerVM DHCP configuration, it seems that the KVM agent calls /usr/lib64/cloud/agent/scripts/network/domr/dhcp_entry.sh with wrongly constructed command line arguments, making the script fail to add correct entries (specifically default router, DNS servers and static routes) for that instance to the routerVM's /etc/dhcphosts.txt + /etc/dhcpopts.txt. Especially adding a specific default gateway fails, so the routerVM will always announce itself as the default router, because the correct entry in /etc/dhcpopts specifying the gateway of the instance's default network as the gateway is missing. This is especially nasty for non-default/additional networks of an instance, messing up the default routing. Examples: Management server log entry: 2012-11-29 14:36:37,764 DEBUG [resource.virtualnetwork.VirtualRoutingResource] (agentRequest-Handler-3:null) Executing: /usr/lib64/cloud/agent/./scripts/network/domr/dhcp_entry.sh -r 169.254.0.122 -v 172.31.2.233 -m 06:b5:88:00:02:30 -n vmname -d 172.31.2.1 -N 172.31.2.201 Notice the double spaces before -d and -N (and the extra space at the EOL) -- WARNING: you have to view the source of this description to actually _see_ the double spaces!) After patching /usr/lib64/cloud/agent/scripts/network/domr/dhcp_entry.sh to do meaningful logging, it's clear that the script does not get called with -d, but -d instead (same for -N), so with an extra space before the dash. Thus, getopts fails to parse/recognize these two arguments correctly and passed empty values for $dfltrt and $dns to the /root/edithosts.sh being called on the routerVM. It's also clear that the CloudStack KVM Java agent calls it with the wrongly constructed command line, because if a shell would interpret this command line, it would just ignore the extra spaces itself. I've not been able to dig it down, but I somehow suspect that one of ./utils/src/com/cloud/utils/script/Script.java:protected String buildCommandLine(String[] command) { } ./utils/src/com/cloud/utils/script/Script.java:public String execute(OutputInterpreter interpreter) { } might mess up building the command line of the command that had been built by ./core/src/com/cloud/agent/resource/virtualnetwork/VirtualRoutingResource.java:protected synchronized Answer execute (final DhcpEntryCommand cmd) { } before. I've not tried 4.0.0 so far, thus I cannot say whether it might be affected or not. As a workaround, I've patched dhcp_entry.sh to re-evaluate the positional parameters using 'set -- ${@}' (will attach a patch, also one for logging). -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Commented] (CLOUDSTACK-585) DHCP entry provisioning is broken in the KVM agent
[ https://issues.apache.org/jira/browse/CLOUDSTACK-585?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13510643#comment-13510643 ] Wolfram Schlich commented on CLOUDSTACK-585: NO!!! The fix is only a *workaround*! It does not fix the cause of the problem! It's inside the KVM agent sourcecode! Please... DHCP entry provisioning is broken in the KVM agent -- Key: CLOUDSTACK-585 URL: https://issues.apache.org/jira/browse/CLOUDSTACK-585 Project: CloudStack Issue Type: Bug Security Level: Public(Anyone can view this level - this is the default.) Components: Hypervisor Controller, KVM Environment: CloudStack 3.0.2 KVM agent running on Fedora 14 Reporter: Wolfram Schlich Labels: agent, kvm Attachments: 1-dhcp_entry.sh.log.patch, 2-dhcp_entry.sh.fix.patch When adding an instance to a routerVM DHCP configuration, it seems that the KVM agent calls /usr/lib64/cloud/agent/scripts/network/domr/dhcp_entry.sh with wrongly constructed command line arguments, making the script fail to add correct entries (specifically default router, DNS servers and static routes) for that instance to the routerVM's /etc/dhcphosts.txt + /etc/dhcpopts.txt. Especially adding a specific default gateway fails, so the routerVM will always announce itself as the default router, because the correct entry in /etc/dhcpopts specifying the gateway of the instance's default network as the gateway is missing. This is especially nasty for non-default/additional networks of an instance, messing up the default routing. Examples: Management server log entry: 2012-11-29 14:36:37,764 DEBUG [resource.virtualnetwork.VirtualRoutingResource] (agentRequest-Handler-3:null) Executing: /usr/lib64/cloud/agent/./scripts/network/domr/dhcp_entry.sh -r 169.254.0.122 -v 172.31.2.233 -m 06:b5:88:00:02:30 -n vmname -d 172.31.2.1 -N 172.31.2.201 Notice the double spaces before -d and -N (and the extra space at the EOL) -- WARNING: you have to view the source of this description to actually _see_ the double spaces!) After patching /usr/lib64/cloud/agent/scripts/network/domr/dhcp_entry.sh to do meaningful logging, it's clear that the script does not get called with -d, but -d instead (same for -N), so with an extra space before the dash. Thus, getopts fails to parse/recognize these two arguments correctly and passed empty values for $dfltrt and $dns to the /root/edithosts.sh being called on the routerVM. It's also clear that the CloudStack KVM Java agent calls it with the wrongly constructed command line, because if a shell would interpret this command line, it would just ignore the extra spaces itself. I've not been able to dig it down, but I somehow suspect that one of ./utils/src/com/cloud/utils/script/Script.java:protected String buildCommandLine(String[] command) { } ./utils/src/com/cloud/utils/script/Script.java:public String execute(OutputInterpreter interpreter) { } might mess up building the command line of the command that had been built by ./core/src/com/cloud/agent/resource/virtualnetwork/VirtualRoutingResource.java:protected synchronized Answer execute (final DhcpEntryCommand cmd) { } before. I've not tried 4.0.0 so far, thus I cannot say whether it might be affected or not. As a workaround, I've patched dhcp_entry.sh to re-evaluate the positional parameters using 'set -- ${@}' (will attach a patch, also one for logging). -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Comment Edited] (CLOUDSTACK-585) DHCP entry provisioning is broken in the KVM agent
[ https://issues.apache.org/jira/browse/CLOUDSTACK-585?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13510646#comment-13510646 ] Wolfram Schlich edited comment on CLOUDSTACK-585 at 12/5/12 6:05 PM: - The command that is executed by the KVM agent seems to be constructed wrongly. Let me explain that using pseudocode: Here's what the Java KVM agent seems to be doing: exec(/usr/lib64/cloud/agent/scripts/network/domr/dhcp_entry.sh, -r, 169.254.2.144, -v, 10.223.170.23, -m, 06:55:22:00:00:11, -n, anthony, -d, 10.223.170.1, -N, 10.223.110.254 ); Please look at the Java source code files and methods that I referenced, then it will get more clear. was (Author: wschlich): The command that is execute by the KVM agent seems to be constructed wrongly. Let me explain that using pseudocode: Here's what the Java KVM agent seems to be doing: exec(/usr/lib64/cloud/agent/scripts/network/domr/dhcp_entry.sh, -r, 169.254.2.144, -v, 10.223.170.23, -m, 06:55:22:00:00:11, -n, anthony, -d, 10.223.170.1, -N, 10.223.110.254 ); Please look at the Java source code files and methods that I referenced, then it will get more clear. DHCP entry provisioning is broken in the KVM agent -- Key: CLOUDSTACK-585 URL: https://issues.apache.org/jira/browse/CLOUDSTACK-585 Project: CloudStack Issue Type: Bug Security Level: Public(Anyone can view this level - this is the default.) Components: Hypervisor Controller, KVM Environment: CloudStack 3.0.2 KVM agent running on Fedora 14 Reporter: Wolfram Schlich Labels: agent, kvm Attachments: 1-dhcp_entry.sh.log.patch, 2-dhcp_entry.sh.fix.patch When adding an instance to a routerVM DHCP configuration, it seems that the KVM agent calls /usr/lib64/cloud/agent/scripts/network/domr/dhcp_entry.sh with wrongly constructed command line arguments, making the script fail to add correct entries (specifically default router, DNS servers and static routes) for that instance to the routerVM's /etc/dhcphosts.txt + /etc/dhcpopts.txt. Especially adding a specific default gateway fails, so the routerVM will always announce itself as the default router, because the correct entry in /etc/dhcpopts specifying the gateway of the instance's default network as the gateway is missing. This is especially nasty for non-default/additional networks of an instance, messing up the default routing. Examples: Management server log entry: 2012-11-29 14:36:37,764 DEBUG [resource.virtualnetwork.VirtualRoutingResource] (agentRequest-Handler-3:null) Executing: /usr/lib64/cloud/agent/./scripts/network/domr/dhcp_entry.sh -r 169.254.0.122 -v 172.31.2.233 -m 06:b5:88:00:02:30 -n vmname -d 172.31.2.1 -N 172.31.2.201 Notice the double spaces before -d and -N (and the extra space at the EOL) -- WARNING: you have to view the source of this description to actually _see_ the double spaces!) After patching /usr/lib64/cloud/agent/scripts/network/domr/dhcp_entry.sh to do meaningful logging, it's clear that the script does not get called with -d, but -d instead (same for -N), so with an extra space before the dash. Thus, getopts fails to parse/recognize these two arguments correctly and passed empty values for $dfltrt and $dns to the /root/edithosts.sh being called on the routerVM. It's also clear that the CloudStack KVM Java agent calls it with the wrongly constructed command line, because if a shell would interpret this command line, it would just ignore the extra spaces itself. I've not been able to dig it down, but I somehow suspect that one of ./utils/src/com/cloud/utils/script/Script.java:protected String buildCommandLine(String[] command) { } ./utils/src/com/cloud/utils/script/Script.java:public String execute(OutputInterpreter interpreter) { } might mess up building the command line of the command that had been built by ./core/src/com/cloud/agent/resource/virtualnetwork/VirtualRoutingResource.java:protected synchronized Answer execute (final DhcpEntryCommand cmd) { } before. I've not tried 4.0.0 so far, thus I cannot say whether it might be affected or not. As a workaround, I've patched dhcp_entry.sh to re-evaluate the positional parameters using 'set -- ${@}' (will attach a patch, also one for logging). -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Commented] (CLOUDSTACK-585) DHCP entry provisioning is broken in the KVM agent
[ https://issues.apache.org/jira/browse/CLOUDSTACK-585?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13510646#comment-13510646 ] Wolfram Schlich commented on CLOUDSTACK-585: The command that is execute by the KVM agent seems to be constructed wrongly. Let me explain that using pseudocode: Here's what the Java KVM agent seems to be doing: exec(/usr/lib64/cloud/agent/scripts/network/domr/dhcp_entry.sh, -r, 169.254.2.144, -v, 10.223.170.23, -m, 06:55:22:00:00:11, -n, anthony, -d, 10.223.170.1, -N, 10.223.110.254 ); Please look at the Java source code files and methods that I referenced, then it will get more clear. DHCP entry provisioning is broken in the KVM agent -- Key: CLOUDSTACK-585 URL: https://issues.apache.org/jira/browse/CLOUDSTACK-585 Project: CloudStack Issue Type: Bug Security Level: Public(Anyone can view this level - this is the default.) Components: Hypervisor Controller, KVM Environment: CloudStack 3.0.2 KVM agent running on Fedora 14 Reporter: Wolfram Schlich Labels: agent, kvm Attachments: 1-dhcp_entry.sh.log.patch, 2-dhcp_entry.sh.fix.patch When adding an instance to a routerVM DHCP configuration, it seems that the KVM agent calls /usr/lib64/cloud/agent/scripts/network/domr/dhcp_entry.sh with wrongly constructed command line arguments, making the script fail to add correct entries (specifically default router, DNS servers and static routes) for that instance to the routerVM's /etc/dhcphosts.txt + /etc/dhcpopts.txt. Especially adding a specific default gateway fails, so the routerVM will always announce itself as the default router, because the correct entry in /etc/dhcpopts specifying the gateway of the instance's default network as the gateway is missing. This is especially nasty for non-default/additional networks of an instance, messing up the default routing. Examples: Management server log entry: 2012-11-29 14:36:37,764 DEBUG [resource.virtualnetwork.VirtualRoutingResource] (agentRequest-Handler-3:null) Executing: /usr/lib64/cloud/agent/./scripts/network/domr/dhcp_entry.sh -r 169.254.0.122 -v 172.31.2.233 -m 06:b5:88:00:02:30 -n vmname -d 172.31.2.1 -N 172.31.2.201 Notice the double spaces before -d and -N (and the extra space at the EOL) -- WARNING: you have to view the source of this description to actually _see_ the double spaces!) After patching /usr/lib64/cloud/agent/scripts/network/domr/dhcp_entry.sh to do meaningful logging, it's clear that the script does not get called with -d, but -d instead (same for -N), so with an extra space before the dash. Thus, getopts fails to parse/recognize these two arguments correctly and passed empty values for $dfltrt and $dns to the /root/edithosts.sh being called on the routerVM. It's also clear that the CloudStack KVM Java agent calls it with the wrongly constructed command line, because if a shell would interpret this command line, it would just ignore the extra spaces itself. I've not been able to dig it down, but I somehow suspect that one of ./utils/src/com/cloud/utils/script/Script.java:protected String buildCommandLine(String[] command) { } ./utils/src/com/cloud/utils/script/Script.java:public String execute(OutputInterpreter interpreter) { } might mess up building the command line of the command that had been built by ./core/src/com/cloud/agent/resource/virtualnetwork/VirtualRoutingResource.java:protected synchronized Answer execute (final DhcpEntryCommand cmd) { } before. I've not tried 4.0.0 so far, thus I cannot say whether it might be affected or not. As a workaround, I've patched dhcp_entry.sh to re-evaluate the positional parameters using 'set -- ${@}' (will attach a patch, also one for logging). -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira
CentOS System VM?
Has anyone looked into building a system VM that runs on a CentOS distro?
Re: MEMORY_CONSTRAINT_VIOLATIONMemory limits must satisfy:
On 04-Dec-2012, at 2:08 PM, Charles Moulliard ch0...@gmail.com wrote: Any idea on how to upgrade xen on Devcloud (= VB running Devian Wheezy + Xen 4.1) ? apt-get update apt-get upgrade On Tue, Dec 4, 2012 at 7:18 PM, Alex Huang alex.hu...@citrix.com wrote: Ah...Didn't realize you're using devCloud. Anthony is right in another email. This is most likely due to a mismatch between the xapi java stub library your deployment is using and the xapi your devcloud is on. --Alex -Original Message- From: Charles Moulliard [mailto:ch0...@gmail.com] Sent: Tuesday, December 04, 2012 9:11 AM To: cloudstack-dev Subject: Re: MEMORY_CONSTRAINT_VIOLATIONMemory limits must satisfy: Alex, There is no /var/log/xensource.log file on the XCP machine running in VirtualBox (= Devcloud2) but those one : ls -la /var/log -rw--- 1 root root20206 Dec 4 16:58 xcp-networkd.log -rw--- 1 root root 113299 Dec 4 17:04 xcp-squeezed.log -rw--- 1 root root27422 Dec 4 16:58 xcp-v6d.log -rw--- 1 root root 10351527 Dec 4 17:07 xcp-xapi.log drwxr-s--- 2 root adm 4096 Dec 3 08:32 xen On Tue, Dec 4, 2012 at 6:01 PM, Alex Huang alex.hu...@citrix.com wrote: /var/log/xensource.log -- Charles Moulliard Apache Committer / Sr. Enterprise Architect (RedHat) Twitter : @cmoulliard | Blog : http://cmoulliard.blogspot.com -- Charles Moulliard Apache Committer / Sr. Enterprise Architect (RedHat) Twitter : @cmoulliard | Blog : http://cmoulliard.blogspot.com
RE: CentOS System VM?
I guess you could, but why? -Original Message- From: Donal Lafferty [mailto:donal.laffe...@citrix.com] Sent: Wednesday, December 05, 2012 1:10 PM To: cloudstack-dev@incubator.apache.org Subject: CentOS System VM? Has anyone looked into building a system VM that runs on a CentOS distro?
RE: CentOS System VM?
CentOS has better support on Hyper-V -Original Message- From: Musayev, Ilya [mailto:imusa...@webmd.net] Sent: 05 December 2012 6:13 PM To: cloudstack-dev@incubator.apache.org Subject: RE: CentOS System VM? I guess you could, but why? -Original Message- From: Donal Lafferty [mailto:donal.laffe...@citrix.com] Sent: Wednesday, December 05, 2012 1:10 PM To: cloudstack-dev@incubator.apache.org Subject: CentOS System VM? Has anyone looked into building a system VM that runs on a CentOS distro?
[jira] [Commented] (CLOUDSTACK-585) DHCP entry provisioning is broken in the KVM agent
[ https://issues.apache.org/jira/browse/CLOUDSTACK-585?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13510653#comment-13510653 ] Anthony Xu commented on CLOUDSTACK-585: --- Why not, I see it as a fix. It's very easy to add one space somehow in java file, instead of checking every possible places in Java. It is a good practice to remove extra space in shell script. And I didn't find out how extra space is added.:-) in VirtualRoutingResource.java, protected synchronized Answer execute (final DhcpEntryCommand cmd) { final Script command = new Script(_dhcpEntryPath, _timeout, s_logger); command.add(-r, cmd.getAccessDetail(NetworkElementCommand.ROUTER_IP)); command.add(-v, cmd.getVmIpAddress()); command.add(-m, cmd.getVmMac()); command.add(-n, cmd.getVmName()); if (cmd.getDefaultRouter() != null) { command.add(-d, cmd.getDefaultRouter()); } if (cmd.getStaticRoutes() != null) { command.add(-s, cmd.getStaticRoutes()); } if (cmd.getDefaultDns() != null) { command.add(-N, cmd.getDefaultDns()); } final String result = command.execute(); return new Answer(cmd, result==null, result); } DHCP entry provisioning is broken in the KVM agent -- Key: CLOUDSTACK-585 URL: https://issues.apache.org/jira/browse/CLOUDSTACK-585 Project: CloudStack Issue Type: Bug Security Level: Public(Anyone can view this level - this is the default.) Components: Hypervisor Controller, KVM Environment: CloudStack 3.0.2 KVM agent running on Fedora 14 Reporter: Wolfram Schlich Labels: agent, kvm Attachments: 1-dhcp_entry.sh.log.patch, 2-dhcp_entry.sh.fix.patch When adding an instance to a routerVM DHCP configuration, it seems that the KVM agent calls /usr/lib64/cloud/agent/scripts/network/domr/dhcp_entry.sh with wrongly constructed command line arguments, making the script fail to add correct entries (specifically default router, DNS servers and static routes) for that instance to the routerVM's /etc/dhcphosts.txt + /etc/dhcpopts.txt. Especially adding a specific default gateway fails, so the routerVM will always announce itself as the default router, because the correct entry in /etc/dhcpopts specifying the gateway of the instance's default network as the gateway is missing. This is especially nasty for non-default/additional networks of an instance, messing up the default routing. Examples: Management server log entry: 2012-11-29 14:36:37,764 DEBUG [resource.virtualnetwork.VirtualRoutingResource] (agentRequest-Handler-3:null) Executing: /usr/lib64/cloud/agent/./scripts/network/domr/dhcp_entry.sh -r 169.254.0.122 -v 172.31.2.233 -m 06:b5:88:00:02:30 -n vmname -d 172.31.2.1 -N 172.31.2.201 Notice the double spaces before -d and -N (and the extra space at the EOL) -- WARNING: you have to view the source of this description to actually _see_ the double spaces!) After patching /usr/lib64/cloud/agent/scripts/network/domr/dhcp_entry.sh to do meaningful logging, it's clear that the script does not get called with -d, but -d instead (same for -N), so with an extra space before the dash. Thus, getopts fails to parse/recognize these two arguments correctly and passed empty values for $dfltrt and $dns to the /root/edithosts.sh being called on the routerVM. It's also clear that the CloudStack KVM Java agent calls it with the wrongly constructed command line, because if a shell would interpret this command line, it would just ignore the extra spaces itself. I've not been able to dig it down, but I somehow suspect that one of ./utils/src/com/cloud/utils/script/Script.java:protected String buildCommandLine(String[] command) { } ./utils/src/com/cloud/utils/script/Script.java:public String execute(OutputInterpreter interpreter) { } might mess up building the command line of the command that had been built by ./core/src/com/cloud/agent/resource/virtualnetwork/VirtualRoutingResource.java:protected synchronized Answer execute (final DhcpEntryCommand cmd) { } before. I've not tried 4.0.0 so far, thus I cannot say whether it might be affected or not. As a workaround, I've patched dhcp_entry.sh to re-evaluate the positional parameters using 'set -- ${@}' (will attach a patch, also one for logging). -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira
RE: CentOS System VM?
CS used to have Fedora 8 as system VM. The reason CS moves to Debian is, 1.Debian 6 has iptable checksum module 2.Debian 6 is stable If CentOS 6.* has iptable checksum module. I think you can. Anthony -Original Message- From: Donal Lafferty [mailto:donal.laffe...@citrix.com] Sent: Wednesday, December 05, 2012 10:16 AM To: cloudstack-dev@incubator.apache.org Subject: RE: CentOS System VM? CentOS has better support on Hyper-V -Original Message- From: Musayev, Ilya [mailto:imusa...@webmd.net] Sent: 05 December 2012 6:13 PM To: cloudstack-dev@incubator.apache.org Subject: RE: CentOS System VM? I guess you could, but why? -Original Message- From: Donal Lafferty [mailto:donal.laffe...@citrix.com] Sent: Wednesday, December 05, 2012 1:10 PM To: cloudstack-dev@incubator.apache.org Subject: CentOS System VM? Has anyone looked into building a system VM that runs on a CentOS distro?
Re: CentOS System VM?
I'm very interested in this. Sent from my iPhone On Dec 5, 2012, at 10:10 AM, Donal Lafferty donal.laffe...@citrix.com wrote: Has anyone looked into building a system VM that runs on a CentOS distro?
[jira] [Comment Edited] (CLOUDSTACK-585) DHCP entry provisioning is broken in the KVM agent
[ https://issues.apache.org/jira/browse/CLOUDSTACK-585?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13510659#comment-13510659 ] Wolfram Schlich edited comment on CLOUDSTACK-585 at 12/5/12 6:29 PM: - Sigh. That's the completely wrong approach IMHO. Fix crap where it emerges, not work around it afterwards. Can you at least *confirm* that the KVM Java agent is doing it wrong (adding spaces)? If you don't have a KVM node at hand, it might be hard, though ; was (Author: wschlich): Sigh. That's the completely wrong approach IMHO. Fix crap where it breaks, not work around it afterwards. Can you at least *confirm* that the KVM Java agent is doing it wrong (adding spaces)? If you don't have a KVM node at hand, it might be hard, though ; DHCP entry provisioning is broken in the KVM agent -- Key: CLOUDSTACK-585 URL: https://issues.apache.org/jira/browse/CLOUDSTACK-585 Project: CloudStack Issue Type: Bug Security Level: Public(Anyone can view this level - this is the default.) Components: Hypervisor Controller, KVM Environment: CloudStack 3.0.2 KVM agent running on Fedora 14 Reporter: Wolfram Schlich Labels: agent, kvm Attachments: 1-dhcp_entry.sh.log.patch, 2-dhcp_entry.sh.fix.patch When adding an instance to a routerVM DHCP configuration, it seems that the KVM agent calls /usr/lib64/cloud/agent/scripts/network/domr/dhcp_entry.sh with wrongly constructed command line arguments, making the script fail to add correct entries (specifically default router, DNS servers and static routes) for that instance to the routerVM's /etc/dhcphosts.txt + /etc/dhcpopts.txt. Especially adding a specific default gateway fails, so the routerVM will always announce itself as the default router, because the correct entry in /etc/dhcpopts specifying the gateway of the instance's default network as the gateway is missing. This is especially nasty for non-default/additional networks of an instance, messing up the default routing. Examples: Management server log entry: 2012-11-29 14:36:37,764 DEBUG [resource.virtualnetwork.VirtualRoutingResource] (agentRequest-Handler-3:null) Executing: /usr/lib64/cloud/agent/./scripts/network/domr/dhcp_entry.sh -r 169.254.0.122 -v 172.31.2.233 -m 06:b5:88:00:02:30 -n vmname -d 172.31.2.1 -N 172.31.2.201 Notice the double spaces before -d and -N (and the extra space at the EOL) -- WARNING: you have to view the source of this description to actually _see_ the double spaces!) After patching /usr/lib64/cloud/agent/scripts/network/domr/dhcp_entry.sh to do meaningful logging, it's clear that the script does not get called with -d, but -d instead (same for -N), so with an extra space before the dash. Thus, getopts fails to parse/recognize these two arguments correctly and passed empty values for $dfltrt and $dns to the /root/edithosts.sh being called on the routerVM. It's also clear that the CloudStack KVM Java agent calls it with the wrongly constructed command line, because if a shell would interpret this command line, it would just ignore the extra spaces itself. I've not been able to dig it down, but I somehow suspect that one of ./utils/src/com/cloud/utils/script/Script.java:protected String buildCommandLine(String[] command) { } ./utils/src/com/cloud/utils/script/Script.java:public String execute(OutputInterpreter interpreter) { } might mess up building the command line of the command that had been built by ./core/src/com/cloud/agent/resource/virtualnetwork/VirtualRoutingResource.java:protected synchronized Answer execute (final DhcpEntryCommand cmd) { } before. I've not tried 4.0.0 so far, thus I cannot say whether it might be affected or not. As a workaround, I've patched dhcp_entry.sh to re-evaluate the positional parameters using 'set -- ${@}' (will attach a patch, also one for logging). -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira
Re: CentOS System VM?
TBH Hyper-V synthetic drivers(modules) is supported in the mainline kernel. So the argument that CentOS 6.x has better support is moot. This assumes that the kernel version on the SSVM is at least 2.6.32. I ran Ubuntu Server 11.x and Centos 6.x on Hyper-V natively and just needed to load the kernel modules for the synthetic stuffs to work. Ancient example of getting the Hyper-V modules built/working on Debian 6.0 http://virtualisationandmanagement.wordpress.com/2010/11/02/debian-on-hyper-v-with-4-vcpu-support-and-syntetic-network/ On Wed, Dec 5, 2012 at 12:25 PM, Kelceydamage@bbits kel...@bbits.ca wrote: I'm very interested in this. Sent from my iPhone On Dec 5, 2012, at 10:10 AM, Donal Lafferty donal.laffe...@citrix.com wrote: Has anyone looked into building a system VM that runs on a CentOS distro?
Re: SSVM Network Configuration Issue
Anthony, The contents of that file are contained in the ssvm-zone1-diagnostics.log attached to my previous message (grep for cmdline. Thanks, -John On Dec 5, 2012, at 12:46 PM, Anthony Xu xuefei...@citrix.com wrote: Hi join, Try following, Set global configuration management.network.cidr to your management server CIDR, if this configuration is not available in UI, you can change it in DB directly. Restart management, Stop/Start SSVM and CPVM. And could you post cat /proc/cmdline in SSVM? Anthony -Original Message- From: John Burwell [mailto:jburw...@basho.com] Sent: Wednesday, December 05, 2012 9:11 AM To: cloudstack-dev@incubator.apache.org Subject: Re: SSVM Network Configuration Issue All, I was wondering if anyone else is experiencing this problem when using secondary storage on a devcloud-style VM with a host-only and NAT adapter. One aspect of this issue that seems interesting is that following route table from the SSVM: root@s-5-TEST:~# route Kernel IP routing table Destination Gateway Genmask Flags Metric RefUse Iface 10.0.3.2192.168.56.1255.255.255.255 UGH 0 00 eth1 10.0.3.0* 255.255.255.0 U 0 00 eth2 192.168.56.0* 255.255.255.0 U 0 00 eth1 192.168.56.0* 255.255.255.0 U 0 00 eth3 link-local * 255.255.0.0 U 0 00 eth0 default 10.0.3.20.0.0.0 UG0 00 eth2 In particular, the gateways for the management and guest networks do not match to the configuration provided to the management server (i.e. 10.0.3.2 is the gateway for the 10.0.3.0/24 network and 192.168.56.1 is the gateway for the 192.168.56.0/24 network). With this configuration, the SSVM has a socket connection to the management server, but is in alert state. Finally, when I remove the host-only NIC and use only a NAT adapter the SSVM's networking works as expecting leading me to believe that the segregated network configuration is at the root of the problem. Until I can get the networking on the SSVM configured, I am unable to complete the testing of the S3-backed Secondary Storage enhancement. Thank you for your help, -John On Dec 3, 2012, at 4:46 PM, John Burwell jburw...@basho.com wrote: All, I am setting up a multi-zone devcloud configuration on VirtualBox 4.2.4 using the Ubuntu 12.04.1 and Xen 4.1. I have configured the base management server VM (zone1) to serve as both the zone1, as well as, the management server (running MySql) with eth0 as a host-only adapter and a static IP of 192.168.56.15 and eth1 as a NAT adapter (see the attached zone1-interfaces file for the exact network configuration on the VM). The management and guest networks are configured as follows: Zone 1 Management: 192.168.56.100-149 gw 192.168.56.1 dns 10.0.3.2 (?) Guest: 10.0.3.200-10.0.3.220 gw 10.0.3.2 dns 8.8.8.8 Zone 2 Management: 192.168.56.150-200 gw 192.68.56.1 dns 10.0.3.2 (?) Guest: 10.0.3.221-240 gw 10.0.3.2 dns 8.8.8.8 The management server deploys and starts without error. I then populate the configuration it using the attached Marvin configuration file (zone1.devcloud.cfg) and restart the management server in order to allow the global configuration option changes to take effect. Following the restart, the CPVM and SSVM start without error. Unfortunately, they drop into alert status, and the SSVM is unable to connect outbound through the guest network (very important for my tests because I am testing S3-backed secondary storage). From the diagnostic checks I have performed on the management server and the SSVM, it appears that the daemon on the SSVM is connecting back to the management server. I have attached a set of diagnostic information from the management server (mgmtsvr-zone1-diagnostics.log) and SSVM server (ssvm-zone1-diagnostics.log) that includes the results of ifconfig, route, netstat and ping checks, as well as, other information (e.g. the contents of /var/cache/cloud/cmdline on the SSVM). Finally, I have attached the vmops log from the management server (vmops-zone1.log). What changes need to be made to management server configuration in order to start up an SSVM that can communicate with the secondary storage NFS volumes, management server, and connect to hosts on the Internet? Thanks for your help, -John ssvm-zone1-diagnostics.log vmops-zone1.tar.gz mgmtsvr-zone1-diagnostics.log zone1-interfaces zone1.devcloud.cfg
Re: CentOS System VM?
The sooner we decouple the systemVMs and allow preference in OS, and user building, we will be better off. Sent from my iPhone On Dec 5, 2012, at 10:38 AM, Jason Davis scr...@gmail.com wrote: TBH Hyper-V synthetic drivers(modules) is supported in the mainline kernel. So the argument that CentOS 6.x has better support is moot. This assumes that the kernel version on the SSVM is at least 2.6.32. I ran Ubuntu Server 11.x and Centos 6.x on Hyper-V natively and just needed to load the kernel modules for the synthetic stuffs to work. Ancient example of getting the Hyper-V modules built/working on Debian 6.0 http://virtualisationandmanagement.wordpress.com/2010/11/02/debian-on-hyper-v-with-4-vcpu-support-and-syntetic-network/ On Wed, Dec 5, 2012 at 12:25 PM, Kelceydamage@bbits kel...@bbits.ca wrote: I'm very interested in this. Sent from my iPhone On Dec 5, 2012, at 10:10 AM, Donal Lafferty donal.laffe...@citrix.com wrote: Has anyone looked into building a system VM that runs on a CentOS distro?
Re: Impact of tomcat CVE-2012-4534
Versions higher than 6.0.33 have problems with classloading the MySQL driver. On 12/5/12 6:20 AM, Gavin Lee gavin@gmail.com wrote: This vulnerability possibly causes denial of service. See below link: http://mail-archives.apache.org/mod_mbox/www-announce/201212.mbox/%3C50BE5 35a.9000...@apache.org%3E It was fixed in tomcat 6.0.36, but we recommand to use 6.0.33. Should we test a higher version and change the guide? -- Gavin
Re: SSVM Network Configuration Issue
Anthony, I apologize for forgetting to response to the part of your answer the first part of the question. I had set the management.network.cidr and host global settings to 192.168.0.0/24 and 192.168.56.18 respectively. Please see the zone1.devcloud.cfg Marvin configuration attached to my original email for the actual setting, as well as, the network configurations used when this problem occurs. Thanks, -John On Dec 5, 2012, at 12:46 PM, Anthony Xu xuefei...@citrix.com wrote: Hi join, Try following, Set global configuration management.network.cidr to your management server CIDR, if this configuration is not available in UI, you can change it in DB directly. Restart management, Stop/Start SSVM and CPVM. And could you post cat /proc/cmdline in SSVM? Anthony -Original Message- From: John Burwell [mailto:jburw...@basho.com] Sent: Wednesday, December 05, 2012 9:11 AM To: cloudstack-dev@incubator.apache.org Subject: Re: SSVM Network Configuration Issue All, I was wondering if anyone else is experiencing this problem when using secondary storage on a devcloud-style VM with a host-only and NAT adapter. One aspect of this issue that seems interesting is that following route table from the SSVM: root@s-5-TEST:~# route Kernel IP routing table Destination Gateway Genmask Flags Metric RefUse Iface 10.0.3.2192.168.56.1255.255.255.255 UGH 0 00 eth1 10.0.3.0* 255.255.255.0 U 0 00 eth2 192.168.56.0* 255.255.255.0 U 0 00 eth1 192.168.56.0* 255.255.255.0 U 0 00 eth3 link-local * 255.255.0.0 U 0 00 eth0 default 10.0.3.20.0.0.0 UG0 00 eth2 In particular, the gateways for the management and guest networks do not match to the configuration provided to the management server (i.e. 10.0.3.2 is the gateway for the 10.0.3.0/24 network and 192.168.56.1 is the gateway for the 192.168.56.0/24 network). With this configuration, the SSVM has a socket connection to the management server, but is in alert state. Finally, when I remove the host-only NIC and use only a NAT adapter the SSVM's networking works as expecting leading me to believe that the segregated network configuration is at the root of the problem. Until I can get the networking on the SSVM configured, I am unable to complete the testing of the S3-backed Secondary Storage enhancement. Thank you for your help, -John On Dec 3, 2012, at 4:46 PM, John Burwell jburw...@basho.com wrote: All, I am setting up a multi-zone devcloud configuration on VirtualBox 4.2.4 using the Ubuntu 12.04.1 and Xen 4.1. I have configured the base management server VM (zone1) to serve as both the zone1, as well as, the management server (running MySql) with eth0 as a host-only adapter and a static IP of 192.168.56.15 and eth1 as a NAT adapter (see the attached zone1-interfaces file for the exact network configuration on the VM). The management and guest networks are configured as follows: Zone 1 Management: 192.168.56.100-149 gw 192.168.56.1 dns 10.0.3.2 (?) Guest: 10.0.3.200-10.0.3.220 gw 10.0.3.2 dns 8.8.8.8 Zone 2 Management: 192.168.56.150-200 gw 192.68.56.1 dns 10.0.3.2 (?) Guest: 10.0.3.221-240 gw 10.0.3.2 dns 8.8.8.8 The management server deploys and starts without error. I then populate the configuration it using the attached Marvin configuration file (zone1.devcloud.cfg) and restart the management server in order to allow the global configuration option changes to take effect. Following the restart, the CPVM and SSVM start without error. Unfortunately, they drop into alert status, and the SSVM is unable to connect outbound through the guest network (very important for my tests because I am testing S3-backed secondary storage). From the diagnostic checks I have performed on the management server and the SSVM, it appears that the daemon on the SSVM is connecting back to the management server. I have attached a set of diagnostic information from the management server (mgmtsvr-zone1-diagnostics.log) and SSVM server (ssvm-zone1-diagnostics.log) that includes the results of ifconfig, route, netstat and ping checks, as well as, other information (e.g. the contents of /var/cache/cloud/cmdline on the SSVM). Finally, I have attached the vmops log from the management server (vmops-zone1.log). What changes need to be made to management server configuration in order to start up an SSVM that can communicate with the secondary storage NFS volumes, management server, and connect to hosts on the Internet? Thanks for your help, -John ssvm-zone1-diagnostics.log vmops-zone1.tar.gz mgmtsvr-zone1-diagnostics.log zone1-interfaces zone1.devcloud.cfg
RE: MEMORY_CONSTRAINT_VIOLATIONMemory limits must satisfy:
Hi Charles, This is a bug, have you filed a bug for this? Anthony -Original Message- From: Charles Moulliard [mailto:ch0...@gmail.com] Sent: Tuesday, December 04, 2012 8:52 AM To: cloudstack-dev Subject: Fwd: MEMORY_CONSTRAINT_VIOLATIONMemory limits must satisfy: Hi, How can I figure out this exception when I try to create a new instance using by example Centos 6.3 ISO (compute offering = 500MHertz, 500MB, local storage) ? WARN [xen.resource.CitrixResourceBase] (DirectAgent-36:) Catch Exception: class com.xensource.xenapi.Types$XenAPIException due to MEMORY_CONSTRAINT_VIOLATIONMemory limits must satisfy: static_min ? dynamic_min ? dynamic_max ? static_max MEMORY_CONSTRAINT_VIOLATIONMemory limits must satisfy: static_min ? dynamic_min ? dynamic_max ? static_max at com.xensource.xenapi.Types.checkResponse(Types.java:1732) at com.xensource.xenapi.Connection.dispatch(Connection.java:368) at com.cloud.hypervisor.xen.resource.XenServerConnectionPool$XenServerConn ection.dispatch(XenServerConnectionPool.java:909) at com.xensource.xenapi.VM.setMemoryStaticMin(VM.java:3450) at com.cloud.hypervisor.xen.resource.CitrixResourceBase.setMemory(CitrixRe sourceBase.java:3116) at com.cloud.hypervisor.xen.resource.CitrixResourceBase.createVmFromTempla te(CitrixResourceBase.java:985) at com.cloud.hypervisor.xen.resource.CitrixResourceBase.execute(CitrixReso urceBase.java:1278) at com.cloud.hypervisor.xen.resource.XcpOssResource.execute(XcpOssResource .java:142) at com.cloud.hypervisor.xen.resource.CitrixResourceBase.executeRequest(Cit rixResourceBase.java:497) at com.cloud.hypervisor.xen.resource.XcpOssResource.executeRequest(XcpOssR esource.java:136) at com.cloud.agent.manager.DirectAgentAttache$Task.run(DirectAgentAttache. java:191) at java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:439) at java.util.concurrent.FutureTask$Sync.innerRun(FutureTask.java:303) at java.util.concurrent.FutureTask.run(FutureTask.java:138) at java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.ac cess$301(ScheduledThreadPoolExecutor.java:98) at java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.ru n(ScheduledThreadPoolExecutor.java:206) at java.util.concurrent.ThreadPoolExecutor$Worker.runTask(ThreadPoolExecut or.java:886) at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.j ava:908) at java.lang.Thread.run(Thread.java:680) Regards, -- Charles Moulliard Apache Committer / Sr. Enterprise Architect (RedHat) Twitter : @cmoulliard | Blog : http://cmoulliard.blogspot.com -- Charles Moulliard Apache Committer / Sr. Enterprise Architect (RedHat) Twitter : @cmoulliard | Blog : http://cmoulliard.blogspot.com
RouterVM bug fix patch - PLS DISCUSS
I've previously started few threads on router vm issue. After the reboot of Router VM, the link local IP could not be reached until we initiate a ping request from within the Router VM. This inturn creates an ARP entry and network is functional from then on. You can see the details of my setup under this tutorial. https://cwiki.apache.org/confluence/display/CLOUDSTACK/CloudStack+Advanced+Network+Tutorial+-+Step+by+Step Example setup - I have 2 network - Link Local is untagged and Guest Network uses VLAN Tagging: Guest Network: 10.18.24.0/23 Guest Network GW: 10.18.24.1 Guest Router IP: 10.18.24.20 Link Local Network: 10.12.34.0/24 Link Local GW: 10.12.34.1 Link Local Router IP: 10.12.34.30 CloudStack Net: 10.15.32.0/23 CloudStack Net GW: 10.15.32.1 CloudStack IP: 10.15.32.141 When router VM powers up, based on variables that are passed via /var/cache/cloud/cmdline, the key variables for this patch are: *Localgw=10.12.34.1 *Gateway=10.18.24.1 *Mgmtcidr=10.15.32.0/23 With one of somewhat recent patches you will notice that Router VM will ping the Localgw and Geteway variables. This may fix some issue with local routing - however my CS Host and gateway is 2 hops away and therefore - the ARP table is not populated completely for inbound communication to function properly. Outbound from the RouterVM communication is fine. My proposed fix is to do a 2 second ping count to a new variable MGMTGW, derived via: MGMTGW=$(echo $Mgtmcidr | awk -F . '{print $1.$2.$3.1}' This will generate the gateway of CloudStack Net GW - however there is an assumption that your gateway always ends with .1 - which is probably typical 97%. Even is the ping to gateway fails because its not set to .1, in theory - the ARP table should be populated with proper MAC addresses and INBOUND communication should be functional after a 2 hop ping. The worst case user will get is a 2 second delay even if ping fails - no harm - so it's a very low risk solution. Please let me know if the proposed solution is acceptable, if so - I will push a patch into JIRA. Thanks ilya
RE: RouterVM bug fix patch - PLS DISCUSS
The ideal solution would be to push the IP of CloudStack core(s) in /var/cache/cloud/cmdline and ping would be initiated just to that host. This will avoid the gateway guessing all together. However, at this moment - no CS GW IP is passed and hence we take another non harmful route. -Original Message- From: Musayev, Ilya [mailto:imusa...@webmd.net] Sent: Wednesday, December 05, 2012 2:38 PM To: cloudstack-dev@incubator.apache.org Subject: RouterVM bug fix patch - PLS DISCUSS I've previously started few threads on router vm issue. After the reboot of Router VM, the link local IP could not be reached until we initiate a ping request from within the Router VM. This inturn creates an ARP entry and network is functional from then on. You can see the details of my setup under this tutorial. https://cwiki.apache.org/confluence/display/CLOUDSTACK/CloudStack+Advanced+Network+Tutorial+-+Step+by+Step Example setup - I have 2 network - Link Local is untagged and Guest Network uses VLAN Tagging: Guest Network: 10.18.24.0/23 Guest Network GW: 10.18.24.1 Guest Router IP: 10.18.24.20 Link Local Network: 10.12.34.0/24 Link Local GW: 10.12.34.1 Link Local Router IP: 10.12.34.30 CloudStack Net: 10.15.32.0/23 CloudStack Net GW: 10.15.32.1 CloudStack IP: 10.15.32.141 When router VM powers up, based on variables that are passed via /var/cache/cloud/cmdline, the key variables for this patch are: *Localgw=10.12.34.1 *Gateway=10.18.24.1 *Mgmtcidr=10.15.32.0/23 With one of somewhat recent patches you will notice that Router VM will ping the Localgw and Geteway variables. This may fix some issue with local routing - however my CS Host and gateway is 2 hops away and therefore - the ARP table is not populated completely for inbound communication to function properly. Outbound from the RouterVM communication is fine. My proposed fix is to do a 2 second ping count to a new variable MGMTGW, derived via: MGMTGW=$(echo $Mgtmcidr | awk -F . '{print $1.$2.$3.1}' This will generate the gateway of CloudStack Net GW - however there is an assumption that your gateway always ends with .1 - which is probably typical 97%. Even is the ping to gateway fails because its not set to .1, in theory - the ARP table should be populated with proper MAC addresses and INBOUND communication should be functional after a 2 hop ping. The worst case user will get is a 2 second delay even if ping fails - no harm - so it's a very low risk solution. Please let me know if the proposed solution is acceptable, if so - I will push a patch into JIRA. Thanks ilya
Re: SSVM Network Configuration Issue
Yes, see your cmdline. internaldns1=10.0.3.2, so it is forcing the use of management network to route to 10.0.3.2 for DNS. that's where the route is coming from. you will want to use something on your management net for internal DNS, or something other than that router. On Wed, Dec 5, 2012 at 11:59 AM, John Burwell jburw...@basho.com wrote: Anthony, I apologize for forgetting to response to the part of your answer the first part of the question. I had set the management.network.cidr and host global settings to 192.168.0.0/24 and 192.168.56.18 respectively. Please see the zone1.devcloud.cfg Marvin configuration attached to my original email for the actual setting, as well as, the network configurations used when this problem occurs. Thanks, -John On Dec 5, 2012, at 12:46 PM, Anthony Xu xuefei...@citrix.com wrote: Hi join, Try following, Set global configuration management.network.cidr to your management server CIDR, if this configuration is not available in UI, you can change it in DB directly. Restart management, Stop/Start SSVM and CPVM. And could you post cat /proc/cmdline in SSVM? Anthony -Original Message- From: John Burwell [mailto:jburw...@basho.com] Sent: Wednesday, December 05, 2012 9:11 AM To: cloudstack-dev@incubator.apache.org Subject: Re: SSVM Network Configuration Issue All, I was wondering if anyone else is experiencing this problem when using secondary storage on a devcloud-style VM with a host-only and NAT adapter. One aspect of this issue that seems interesting is that following route table from the SSVM: root@s-5-TEST:~# route Kernel IP routing table Destination Gateway Genmask Flags Metric RefUse Iface 10.0.3.2192.168.56.1255.255.255.255 UGH 0 00 eth1 10.0.3.0* 255.255.255.0 U 0 00 eth2 192.168.56.0* 255.255.255.0 U 0 00 eth1 192.168.56.0* 255.255.255.0 U 0 00 eth3 link-local * 255.255.0.0 U 0 00 eth0 default 10.0.3.20.0.0.0 UG0 00 eth2 In particular, the gateways for the management and guest networks do not match to the configuration provided to the management server (i.e. 10.0.3.2 is the gateway for the 10.0.3.0/24 network and 192.168.56.1 is the gateway for the 192.168.56.0/24 network). With this configuration, the SSVM has a socket connection to the management server, but is in alert state. Finally, when I remove the host-only NIC and use only a NAT adapter the SSVM's networking works as expecting leading me to believe that the segregated network configuration is at the root of the problem. Until I can get the networking on the SSVM configured, I am unable to complete the testing of the S3-backed Secondary Storage enhancement. Thank you for your help, -John On Dec 3, 2012, at 4:46 PM, John Burwell jburw...@basho.com wrote: All, I am setting up a multi-zone devcloud configuration on VirtualBox 4.2.4 using the Ubuntu 12.04.1 and Xen 4.1. I have configured the base management server VM (zone1) to serve as both the zone1, as well as, the management server (running MySql) with eth0 as a host-only adapter and a static IP of 192.168.56.15 and eth1 as a NAT adapter (see the attached zone1-interfaces file for the exact network configuration on the VM). The management and guest networks are configured as follows: Zone 1 Management: 192.168.56.100-149 gw 192.168.56.1 dns 10.0.3.2 (?) Guest: 10.0.3.200-10.0.3.220 gw 10.0.3.2 dns 8.8.8.8 Zone 2 Management: 192.168.56.150-200 gw 192.68.56.1 dns 10.0.3.2 (?) Guest: 10.0.3.221-240 gw 10.0.3.2 dns 8.8.8.8 The management server deploys and starts without error. I then populate the configuration it using the attached Marvin configuration file (zone1.devcloud.cfg) and restart the management server in order to allow the global configuration option changes to take effect. Following the restart, the CPVM and SSVM start without error. Unfortunately, they drop into alert status, and the SSVM is unable to connect outbound through the guest network (very important for my tests because I am testing S3-backed secondary storage). From the diagnostic checks I have performed on the management server and the SSVM, it appears that the daemon on the SSVM is connecting back to the management server. I have attached a set of diagnostic information from the management server (mgmtsvr-zone1-diagnostics.log) and SSVM server (ssvm-zone1-diagnostics.log) that includes the results of ifconfig, route, netstat and ping checks, as well as, other information (e.g. the contents of /var/cache/cloud/cmdline on the SSVM). Finally, I have attached the vmops log from the management server (vmops-zone1.log).
[jira] [Created] (CLOUDSTACK-587) MEMORY_CONSTRAINT_VIOLATIONMemory limits must satisfy:
Anthony Xu created CLOUDSTACK-587: - Summary: MEMORY_CONSTRAINT_VIOLATIONMemory limits must satisfy: Key: CLOUDSTACK-587 URL: https://issues.apache.org/jira/browse/CLOUDSTACK-587 Project: CloudStack Issue Type: Bug Security Level: Public (Anyone can view this level - this is the default.) Reporter: Anthony Xu Assignee: Anthony Xu How can I figure out this exception when I try to create a new instance using by example Centos 6.3 ISO (compute offering = 500MHertz, 500MB, local storage) ? WARN [xen.resource.CitrixResourceBase] (DirectAgent-36:) Catch Exception: class com.xensource.xenapi.Types$XenAPIException due to MEMORY_CONSTRAINT_VIOLATIONMemory limits must satisfy: static_min ? dynamic_min ? dynamic_max ? static_max MEMORY_CONSTRAINT_VIOLATIONMemory limits must satisfy: static_min ? dynamic_min ? dynamic_max ? static_max at com.xensource.xenapi.Types.checkResponse(Types.java:1732) at com.xensource.xenapi.Connection.dispatch(Connection.java:368) at com.cloud.hypervisor.xen.resource.XenServerConnectionPool$XenServerCon n ection.dispatch(XenServerConnectionPool.java:909) at com.xensource.xenapi.VM.setMemoryStaticMin(VM.java:3450) at com.cloud.hypervisor.xen.resource.CitrixResourceBase.setMemory(CitrixR e sourceBase.java:3116) at com.cloud.hypervisor.xen.resource.CitrixResourceBase.createVmFromTempl a te(CitrixResourceBase.java:985) at com.cloud.hypervisor.xen.resource.CitrixResourceBase.execute(CitrixRes o urceBase.java:1278) at com.cloud.hypervisor.xen.resource.XcpOssResource.execute(XcpOssResourc e .java:142) at com.cloud.hypervisor.xen.resource.CitrixResourceBase.executeRequest(Ci t rixResourceBase.java:497) at com.cloud.hypervisor.xen.resource.XcpOssResource.executeRequest(XcpOss R esource.java:136) at com.cloud.agent.manager.DirectAgentAttache$Task.run(DirectAgentAttache. java:191) at java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:439 ) at java.util.concurrent.FutureTask$Sync.innerRun(FutureTask.java:303) at java.util.concurrent.FutureTask.run(FutureTask.java:138) at java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.a c cess$301(ScheduledThreadPoolExecutor.java:98) at java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.r u n(ScheduledThreadPoolExecutor.java:206) at java.util.concurrent.ThreadPoolExecutor$Worker.runTask(ThreadPoolExecu t or.java:886) at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor. j ava:908) at java.lang.Thread.run(Thread.java:680) Regards, -- -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira
Re: Review Request: Changed the @Parameter for some VM group commands.
--- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/8133/#review14057 --- Ship it! Ship It! - Rohit Yadav On Nov. 20, 2012, 2:05 a.m., Fang Wang wrote: --- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/8133/ --- (Updated Nov. 20, 2012, 2:05 a.m.) Review request for cloudstack, Prachi Damle and Alex Huang. Description --- Part of the API refactoring work. This is the initial modification to the @Parameter annotation for some commands in the VM group. This addresses bug cloudstack-518. Diffs - api/src/com/cloud/api/Parameter.java 2da3c40 api/src/org/apache/cloudstack/user/api/vm/commands/AssignVMCmd.java PRE-CREATION api/src/org/apache/cloudstack/user/api/vm/commands/CreateVMGroupCmd.java PRE-CREATION api/src/org/apache/cloudstack/user/api/vm/commands/DeleteVMGroupCmd.java PRE-CREATION api/src/org/apache/cloudstack/user/api/vm/commands/DeployVMCmd.java PRE-CREATION api/src/org/apache/cloudstack/user/api/vm/commands/DestroySystemVmCmd.java PRE-CREATION Diff: https://reviews.apache.org/r/8133/diff/ Testing --- Thanks, Fang Wang
Re: Review Request: Changed the @Parameter for some VM group commands.
On Dec. 5, 2012, 7:53 p.m., Rohit Yadav wrote: Ship It! Committed on api_refactoring branch: commit ee0a4a41d0f8db2f895a1699150c60cac6facab8 Author: Fang Wang fang.w...@citrix.com Date: Wed Dec 5 11:50:23 2012 -0800 api_refactor: entity annotation and fix parameter - Add Entity annotation - Annotate few responses - Fix Parameter annotation to have entityType Signed-off-by: Rohit Yadav bhais...@apache.org - Rohit --- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/8133/#review14057 --- On Nov. 20, 2012, 2:05 a.m., Fang Wang wrote: --- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/8133/ --- (Updated Nov. 20, 2012, 2:05 a.m.) Review request for cloudstack, Prachi Damle and Alex Huang. Description --- Part of the API refactoring work. This is the initial modification to the @Parameter annotation for some commands in the VM group. This addresses bug cloudstack-518. Diffs - api/src/com/cloud/api/Parameter.java 2da3c40 api/src/org/apache/cloudstack/user/api/vm/commands/AssignVMCmd.java PRE-CREATION api/src/org/apache/cloudstack/user/api/vm/commands/CreateVMGroupCmd.java PRE-CREATION api/src/org/apache/cloudstack/user/api/vm/commands/DeleteVMGroupCmd.java PRE-CREATION api/src/org/apache/cloudstack/user/api/vm/commands/DeployVMCmd.java PRE-CREATION api/src/org/apache/cloudstack/user/api/vm/commands/DestroySystemVmCmd.java PRE-CREATION Diff: https://reviews.apache.org/r/8133/diff/ Testing --- Thanks, Fang Wang
[jira] [Created] (CLOUDSTACK-588) restoreVirtualMachine api functionality on the UI
Pranav Saxena created CLOUDSTACK-588: Summary: restoreVirtualMachine api functionality on the UI Key: CLOUDSTACK-588 URL: https://issues.apache.org/jira/browse/CLOUDSTACK-588 Project: CloudStack Issue Type: Bug Security Level: Public (Anyone can view this level - this is the default.) Components: UI Reporter: Pranav Saxena RestoreVirtualMachine api action not shown in UI. We have 2 apis in Cloudstack - restoreVirtualMachine and recoverVirtualMachine. But we show actions for only one of them in the UI actions. We should have one for restoreVirtualMachine api and call it Reset vm on the UI. Restore action can happen while VMs are in the Running and Stopped state and can be done by root admin /domain-admin and users. -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Assigned] (CLOUDSTACK-588) restoreVirtualMachine api functionality on the UI
[ https://issues.apache.org/jira/browse/CLOUDSTACK-588?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Pranav Saxena reassigned CLOUDSTACK-588: Assignee: Pranav Saxena restoreVirtualMachine api functionality on the UI - Key: CLOUDSTACK-588 URL: https://issues.apache.org/jira/browse/CLOUDSTACK-588 Project: CloudStack Issue Type: Bug Security Level: Public(Anyone can view this level - this is the default.) Components: UI Reporter: Pranav Saxena Assignee: Pranav Saxena RestoreVirtualMachine api action not shown in UI. We have 2 apis in Cloudstack - restoreVirtualMachine and recoverVirtualMachine. But we show actions for only one of them in the UI actions. We should have one for restoreVirtualMachine api and call it Reset vm on the UI. Restore action can happen while VMs are in the Running and Stopped state and can be done by root admin /domain-admin and users. -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira
qemu-img convert a qcow2 to a qcow2
Anyone know why we do a convert from qcow2 to qcow2 when we crete a volume or template? It seems slower than file copy, and it strips valuable compression which could speed up deployments significantly. Our qcow2 templates are compressed to about 1/3 size of uncompressed, and since the compression is read-only it doesn't really affect future write performance, and gives a slight performance gain to reads since less is read. createvolume.sh qemu_img convert -f qcow2 -O qcow2 managesnapshot.sh qemu_img convert -f qcow2 -O qcow2 createtmplt.sh qemu_img convert -f qcow2 -O qcow2 I'd suggest if we already know it's qcow2, and we want to create a qcow2 to copy the file. Objections?
[jira] [Commented] (CLOUDSTACK-536) remove citrix cloudpatform from 4.0 build - CloudStack is ASF project
[ https://issues.apache.org/jira/browse/CLOUDSTACK-536?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13510733#comment-13510733 ] Joe Brockmeier commented on CLOUDSTACK-536: --- As far as I can tell, these aren't displaying in ACS 4.0 - just present. Looks like it's been set up to have a workaround to call no-logo in the stylesheet. We should clean it up to get rid of these entirely, but I don't think it's crucial that they be removed in 4.0.1 if we don't have a patch by freeze. remove citrix cloudpatform from 4.0 build - CloudStack is ASF project - Key: CLOUDSTACK-536 URL: https://issues.apache.org/jira/browse/CLOUDSTACK-536 Project: CloudStack Issue Type: Bug Security Level: Public(Anyone can view this level - this is the default.) Components: UI Affects Versions: 4.0.0, 4.0.1 Environment: CS 4.0 from http://cloudstack.apt-get.eu/rhel/4.0/ The rpms are: cloud-utils-4.0.0-0.140.el6.4.0.x86_64 cloud-client-4.0.0-0.140.el6.4.0.x86_64 cloud-setup-4.0.0-0.140.el6.4.0.x86_64 cloud-scripts-4.0.0-0.140.el6.4.0.x86_64 cloud-deps-4.0.0-0.140.el6.4.0.x86_64 cloud-server-4.0.0-0.140.el6.4.0.x86_64 cloud-client-ui-4.0.0-0.140.el6.4.0.x86_64 cloud-python-4.0.0-0.140.el6.4.0.x86_64 cloud-core-4.0.0-0.140.el6.4.0.x86_64 cloud-usage-4.0.0-0.140.el6.4.0.x86_64 cloud-aws-api-4.0.0-0.140.el6.4.0.x86_64 Reporter: ilya musayev Assignee: Brian Federle Priority: Minor Fix For: 4.0.0, 4.0.1, 4.1.0 Please remove Citrix Logo and CloudPlaform from the UI page, files are (could be more): rpm/deb release: in /client/images/ logo-login.png citrix-logo-darkbg.png logo-cloudplatform.png and possibly others that may contain Citrix CloudPlatform mention. Also, please include the wording for CloudStack as Apache CloudStack in corresponding documentation and UI pages for CS ASF project. Thank you -ilya -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira
RE: [jira] [Commented] (CLOUDSTACK-536) remove citrix cloudpatform from 4.0 build - CloudStack is ASF project
Agreed, as these are not functionality impacting. -Original Message- From: Joe Brockmeier (JIRA) [mailto:j...@apache.org] Sent: Wednesday, December 05, 2012 3:30 PM To: cloudstack-dev@incubator.apache.org Subject: [jira] [Commented] (CLOUDSTACK-536) remove citrix cloudpatform from 4.0 build - CloudStack is ASF project [ https://issues.apache.org/jira/browse/CLOUDSTACK-536?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13510733#comment-13510733 ] Joe Brockmeier commented on CLOUDSTACK-536: --- As far as I can tell, these aren't displaying in ACS 4.0 - just present. Looks like it's been set up to have a workaround to call no-logo in the stylesheet. We should clean it up to get rid of these entirely, but I don't think it's crucial that they be removed in 4.0.1 if we don't have a patch by freeze. remove citrix cloudpatform from 4.0 build - CloudStack is ASF project - Key: CLOUDSTACK-536 URL: https://issues.apache.org/jira/browse/CLOUDSTACK-536 Project: CloudStack Issue Type: Bug Security Level: Public(Anyone can view this level - this is the default.) Components: UI Affects Versions: 4.0.0, 4.0.1 Environment: CS 4.0 from http://cloudstack.apt-get.eu/rhel/4.0/ The rpms are: cloud-utils-4.0.0-0.140.el6.4.0.x86_64 cloud-client-4.0.0-0.140.el6.4.0.x86_64 cloud-setup-4.0.0-0.140.el6.4.0.x86_64 cloud-scripts-4.0.0-0.140.el6.4.0.x86_64 cloud-deps-4.0.0-0.140.el6.4.0.x86_64 cloud-server-4.0.0-0.140.el6.4.0.x86_64 cloud-client-ui-4.0.0-0.140.el6.4.0.x86_64 cloud-python-4.0.0-0.140.el6.4.0.x86_64 cloud-core-4.0.0-0.140.el6.4.0.x86_64 cloud-usage-4.0.0-0.140.el6.4.0.x86_64 cloud-aws-api-4.0.0-0.140.el6.4.0.x86_64 Reporter: ilya musayev Assignee: Brian Federle Priority: Minor Fix For: 4.0.0, 4.0.1, 4.1.0 Please remove Citrix Logo and CloudPlaform from the UI page, files are (could be more): rpm/deb release: in /client/images/ logo-login.png citrix-logo-darkbg.png logo-cloudplatform.png and possibly others that may contain Citrix CloudPlatform mention. Also, please include the wording for CloudStack as Apache CloudStack in corresponding documentation and UI pages for CS ASF project. Thank you -ilya -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Commented] (CLOUDSTACK-536) remove citrix cloudpatform from 4.0 build - CloudStack is ASF project
[ https://issues.apache.org/jira/browse/CLOUDSTACK-536?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13510744#comment-13510744 ] Musayev, Ilya commented on CLOUDSTACK-536: -- Agreed, as these are not functionality impacting. remove citrix cloudpatform from 4.0 build - CloudStack is ASF project - Key: CLOUDSTACK-536 URL: https://issues.apache.org/jira/browse/CLOUDSTACK-536 Project: CloudStack Issue Type: Bug Security Level: Public(Anyone can view this level - this is the default.) Components: UI Affects Versions: 4.0.0, 4.0.1 Environment: CS 4.0 from http://cloudstack.apt-get.eu/rhel/4.0/ The rpms are: cloud-utils-4.0.0-0.140.el6.4.0.x86_64 cloud-client-4.0.0-0.140.el6.4.0.x86_64 cloud-setup-4.0.0-0.140.el6.4.0.x86_64 cloud-scripts-4.0.0-0.140.el6.4.0.x86_64 cloud-deps-4.0.0-0.140.el6.4.0.x86_64 cloud-server-4.0.0-0.140.el6.4.0.x86_64 cloud-client-ui-4.0.0-0.140.el6.4.0.x86_64 cloud-python-4.0.0-0.140.el6.4.0.x86_64 cloud-core-4.0.0-0.140.el6.4.0.x86_64 cloud-usage-4.0.0-0.140.el6.4.0.x86_64 cloud-aws-api-4.0.0-0.140.el6.4.0.x86_64 Reporter: ilya musayev Assignee: Brian Federle Priority: Minor Fix For: 4.0.0, 4.0.1, 4.1.0 Please remove Citrix Logo and CloudPlaform from the UI page, files are (could be more): rpm/deb release: in /client/images/ logo-login.png citrix-logo-darkbg.png logo-cloudplatform.png and possibly others that may contain Citrix CloudPlatform mention. Also, please include the wording for CloudStack as Apache CloudStack in corresponding documentation and UI pages for CS ASF project. Thank you -ilya -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira
Re: qemu-img convert a qcow2 to a qcow2
On 05-Dec-2012, at 12:17 PM, Marcus Sorensen shadow...@gmail.com wrote: Anyone know why we do a convert from qcow2 to qcow2 when we crete a volume or template? It seems slower than file copy, and it strips valuable compression which could speed up deployments significantly. Our qcow2 templates are compressed to about 1/3 size of uncompressed, and since the compression is read-only it doesn't really affect future write performance, and gives a slight performance gain to reads since less is read. createvolume.sh qemu_img convert -f qcow2 -O qcow2 managesnapshot.sh qemu_img convert -f qcow2 -O qcow2 createtmplt.sh qemu_img convert -f qcow2 -O qcow2 I'd suggest if we already know it's qcow2, and we want to create a qcow2 to copy the file. Objections? I don't know why we do that, but what the command is saying is that we want to convert an input image from qcow2 to keeping the same output format, -O qcow2 But yes, if the input format is already qcow2 and output is qcow2 we can just copy the file. Edison? Regards.
[jira] [Resolved] (CLOUDSTACK-587) MEMORY_CONSTRAINT_VIOLATIONMemory limits must satisfy:
[ https://issues.apache.org/jira/browse/CLOUDSTACK-587?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Rohit Yadav resolved CLOUDSTACK-587. Resolution: Fixed Fixed by Anthony on master: commit 40682fc43dc202b1dbffb4fd32388ad1a46f6e6a Author: anthony anth...@cloud.com Date: Wed Dec 5 05:24:46 2012 -0800 CLOUDSTACK-587: use setMemoryLimits instead MEMORY_CONSTRAINT_VIOLATIONMemory limits must satisfy: -- Key: CLOUDSTACK-587 URL: https://issues.apache.org/jira/browse/CLOUDSTACK-587 Project: CloudStack Issue Type: Bug Security Level: Public(Anyone can view this level - this is the default.) Reporter: Anthony Xu Assignee: Anthony Xu How can I figure out this exception when I try to create a new instance using by example Centos 6.3 ISO (compute offering = 500MHertz, 500MB, local storage) ? WARN [xen.resource.CitrixResourceBase] (DirectAgent-36:) Catch Exception: class com.xensource.xenapi.Types$XenAPIException due to MEMORY_CONSTRAINT_VIOLATIONMemory limits must satisfy: static_min ? dynamic_min ? dynamic_max ? static_max MEMORY_CONSTRAINT_VIOLATIONMemory limits must satisfy: static_min ? dynamic_min ? dynamic_max ? static_max at com.xensource.xenapi.Types.checkResponse(Types.java:1732) at com.xensource.xenapi.Connection.dispatch(Connection.java:368) at com.cloud.hypervisor.xen.resource.XenServerConnectionPool$XenServerCon n ection.dispatch(XenServerConnectionPool.java:909) at com.xensource.xenapi.VM.setMemoryStaticMin(VM.java:3450) at com.cloud.hypervisor.xen.resource.CitrixResourceBase.setMemory(CitrixR e sourceBase.java:3116) at com.cloud.hypervisor.xen.resource.CitrixResourceBase.createVmFromTempl a te(CitrixResourceBase.java:985) at com.cloud.hypervisor.xen.resource.CitrixResourceBase.execute(CitrixRes o urceBase.java:1278) at com.cloud.hypervisor.xen.resource.XcpOssResource.execute(XcpOssResourc e .java:142) at com.cloud.hypervisor.xen.resource.CitrixResourceBase.executeRequest(Ci t rixResourceBase.java:497) at com.cloud.hypervisor.xen.resource.XcpOssResource.executeRequest(XcpOss R esource.java:136) at com.cloud.agent.manager.DirectAgentAttache$Task.run(DirectAgentAttache. java:191) at java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:439 ) at java.util.concurrent.FutureTask$Sync.innerRun(FutureTask.java:303) at java.util.concurrent.FutureTask.run(FutureTask.java:138) at java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.a c cess$301(ScheduledThreadPoolExecutor.java:98) at java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.r u n(ScheduledThreadPoolExecutor.java:206) at java.util.concurrent.ThreadPoolExecutor$Worker.runTask(ThreadPoolExecu t or.java:886) at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor. j ava:908) at java.lang.Thread.run(Thread.java:680) Regards, -- -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Updated] (CLOUDSTACK-588) restoreVirtualMachine api functionality on the UI
[ https://issues.apache.org/jira/browse/CLOUDSTACK-588?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Pranav Saxena updated CLOUDSTACK-588: - Affects Version/s: 4.0.1 4.0.0 Fix Version/s: 4.0.1 Summary: restoreVirtualMachine api functionality on the UI (was: restoreVirtualMachine api functionality on the UI) restoreVirtualMachine api functionality on the UI -- Key: CLOUDSTACK-588 URL: https://issues.apache.org/jira/browse/CLOUDSTACK-588 Project: CloudStack Issue Type: Bug Security Level: Public(Anyone can view this level - this is the default.) Components: UI Affects Versions: 4.0.0, 4.0.1 Reporter: Pranav Saxena Assignee: Pranav Saxena Fix For: 4.0.1 RestoreVirtualMachine api action not shown in UI. We have 2 apis in Cloudstack - restoreVirtualMachine and recoverVirtualMachine. But we show actions for only one of them in the UI actions. We should have one for restoreVirtualMachine api and call it Reset vm on the UI. Restore action can happen while VMs are in the Running and Stopped state and can be done by root admin /domain-admin and users. -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira
Re: qemu-img convert a qcow2 to a qcow2
Eventually we'll probably want to support multiple versions, replacing one or both of the formats with variables, but even then we'll probably just want to copy if source format == destination format. On Wed, Dec 5, 2012 at 1:51 PM, Rohit Yadav rohit.ya...@citrix.com wrote: On 05-Dec-2012, at 12:17 PM, Marcus Sorensen shadow...@gmail.com wrote: Anyone know why we do a convert from qcow2 to qcow2 when we crete a volume or template? It seems slower than file copy, and it strips valuable compression which could speed up deployments significantly. Our qcow2 templates are compressed to about 1/3 size of uncompressed, and since the compression is read-only it doesn't really affect future write performance, and gives a slight performance gain to reads since less is read. createvolume.sh qemu_img convert -f qcow2 -O qcow2 managesnapshot.sh qemu_img convert -f qcow2 -O qcow2 createtmplt.sh qemu_img convert -f qcow2 -O qcow2 I'd suggest if we already know it's qcow2, and we want to create a qcow2 to copy the file. Objections? I don't know why we do that, but what the command is saying is that we want to convert an input image from qcow2 to keeping the same output format, -O qcow2 But yes, if the input format is already qcow2 and output is qcow2 we can just copy the file. Edison? Regards.
[jira] [Commented] (CLOUDSTACK-588) restoreVirtualMachine api functionality on the UI
[ https://issues.apache.org/jira/browse/CLOUDSTACK-588?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13510761#comment-13510761 ] David Nalley commented on CLOUDSTACK-588: - 4.0.1 is a bugfix version, please introduce new functionality into the next feature release (4.1.0) restoreVirtualMachine api functionality on the UI -- Key: CLOUDSTACK-588 URL: https://issues.apache.org/jira/browse/CLOUDSTACK-588 Project: CloudStack Issue Type: Bug Security Level: Public(Anyone can view this level - this is the default.) Components: UI Affects Versions: 4.0.0, 4.0.1 Reporter: Pranav Saxena Assignee: Pranav Saxena Fix For: 4.0.1 RestoreVirtualMachine api action not shown in UI. We have 2 apis in Cloudstack - restoreVirtualMachine and recoverVirtualMachine. But we show actions for only one of them in the UI actions. We should have one for restoreVirtualMachine api and call it Reset vm on the UI. Restore action can happen while VMs are in the Running and Stopped state and can be done by root admin /domain-admin and users. -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Updated] (CLOUDSTACK-588) restoreVirtualMachine api functionality on the UI
[ https://issues.apache.org/jira/browse/CLOUDSTACK-588?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Pranav Saxena updated CLOUDSTACK-588: - Fix Version/s: (was: 4.0.1) 4.1.0 restoreVirtualMachine api functionality on the UI -- Key: CLOUDSTACK-588 URL: https://issues.apache.org/jira/browse/CLOUDSTACK-588 Project: CloudStack Issue Type: Bug Security Level: Public(Anyone can view this level - this is the default.) Components: UI Affects Versions: 4.0.0, 4.0.1 Reporter: Pranav Saxena Assignee: Pranav Saxena Fix For: 4.1.0 RestoreVirtualMachine api action not shown in UI. We have 2 apis in Cloudstack - restoreVirtualMachine and recoverVirtualMachine. But we show actions for only one of them in the UI actions. We should have one for restoreVirtualMachine api and call it Reset vm on the UI. Restore action can happen while VMs are in the Running and Stopped state and can be done by root admin /domain-admin and users. -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Commented] (CLOUDSTACK-588) restoreVirtualMachine api functionality on the UI
[ https://issues.apache.org/jira/browse/CLOUDSTACK-588?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13510765#comment-13510765 ] Pranav Saxena commented on CLOUDSTACK-588: -- Sure David , done that . This functionality is supported by the API but wasn't supported by the UI .Hence I have added the functionality to support it from the UI as well . restoreVirtualMachine api functionality on the UI -- Key: CLOUDSTACK-588 URL: https://issues.apache.org/jira/browse/CLOUDSTACK-588 Project: CloudStack Issue Type: Bug Security Level: Public(Anyone can view this level - this is the default.) Components: UI Affects Versions: 4.0.0, 4.0.1 Reporter: Pranav Saxena Assignee: Pranav Saxena Fix For: 4.1.0 RestoreVirtualMachine api action not shown in UI. We have 2 apis in Cloudstack - restoreVirtualMachine and recoverVirtualMachine. But we show actions for only one of them in the UI actions. We should have one for restoreVirtualMachine api and call it Reset vm on the UI. Restore action can happen while VMs are in the Running and Stopped state and can be done by root admin /domain-admin and users. -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Commented] (CLOUDSTACK-588) restoreVirtualMachine api functionality on the UI
[ https://issues.apache.org/jira/browse/CLOUDSTACK-588?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13510766#comment-13510766 ] Pranav Saxena commented on CLOUDSTACK-588: -- Commit: http://git-wip-us.apache.org/repos/asf/incubator-cloudstack/commit/30dd34d2 restoreVirtualMachine api functionality on the UI -- Key: CLOUDSTACK-588 URL: https://issues.apache.org/jira/browse/CLOUDSTACK-588 Project: CloudStack Issue Type: Bug Security Level: Public(Anyone can view this level - this is the default.) Components: UI Affects Versions: 4.0.0, 4.0.1 Reporter: Pranav Saxena Assignee: Pranav Saxena Fix For: 4.1.0 RestoreVirtualMachine api action not shown in UI. We have 2 apis in Cloudstack - restoreVirtualMachine and recoverVirtualMachine. But we show actions for only one of them in the UI actions. We should have one for restoreVirtualMachine api and call it Reset vm on the UI. Restore action can happen while VMs are in the Running and Stopped state and can be done by root admin /domain-admin and users. -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Resolved] (CLOUDSTACK-588) restoreVirtualMachine api functionality on the UI
[ https://issues.apache.org/jira/browse/CLOUDSTACK-588?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Pranav Saxena resolved CLOUDSTACK-588. -- Resolution: Fixed restoreVirtualMachine api functionality on the UI -- Key: CLOUDSTACK-588 URL: https://issues.apache.org/jira/browse/CLOUDSTACK-588 Project: CloudStack Issue Type: Bug Security Level: Public(Anyone can view this level - this is the default.) Components: UI Affects Versions: 4.0.0, 4.0.1 Reporter: Pranav Saxena Assignee: Pranav Saxena Fix For: 4.1.0 RestoreVirtualMachine api action not shown in UI. We have 2 apis in Cloudstack - restoreVirtualMachine and recoverVirtualMachine. But we show actions for only one of them in the UI actions. We should have one for restoreVirtualMachine api and call it Reset vm on the UI. Restore action can happen while VMs are in the Running and Stopped state and can be done by root admin /domain-admin and users. -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Updated] (CLOUDSTACK-588) restoreVirtualMachine api functionality on the UI
[ https://issues.apache.org/jira/browse/CLOUDSTACK-588?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Pranav Saxena updated CLOUDSTACK-588: - Issue Type: Task (was: Bug) restoreVirtualMachine api functionality on the UI -- Key: CLOUDSTACK-588 URL: https://issues.apache.org/jira/browse/CLOUDSTACK-588 Project: CloudStack Issue Type: Task Security Level: Public(Anyone can view this level - this is the default.) Components: UI Affects Versions: 4.0.0, 4.0.1 Reporter: Pranav Saxena Assignee: Pranav Saxena Fix For: 4.1.0 RestoreVirtualMachine api action not shown in UI. We have 2 apis in Cloudstack - restoreVirtualMachine and recoverVirtualMachine. But we show actions for only one of them in the UI actions. We should have one for restoreVirtualMachine api and call it Reset vm on the UI. Restore action can happen while VMs are in the Running and Stopped state and can be done by root admin /domain-admin and users. -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira
RE: qemu-img convert a qcow2 to a qcow2
If the source image has dependence on other images, then we should use qemu-img convert , otherwise, cp is enough. -Original Message- From: Marcus Sorensen [mailto:shadow...@gmail.com] Sent: Wednesday, December 05, 2012 1:01 PM To: cloudstack-dev@incubator.apache.org Subject: Re: qemu-img convert a qcow2 to a qcow2 Eventually we'll probably want to support multiple versions, replacing one or both of the formats with variables, but even then we'll probably just want to copy if source format == destination format. On Wed, Dec 5, 2012 at 1:51 PM, Rohit Yadav rohit.ya...@citrix.com wrote: On 05-Dec-2012, at 12:17 PM, Marcus Sorensen shadow...@gmail.com wrote: Anyone know why we do a convert from qcow2 to qcow2 when we crete a volume or template? It seems slower than file copy, and it strips valuable compression which could speed up deployments significantly. Our qcow2 templates are compressed to about 1/3 size of uncompressed, and since the compression is read-only it doesn't really affect future write performance, and gives a slight performance gain to reads since less is read. createvolume.sh qemu_img convert -f qcow2 -O qcow2 managesnapshot.sh qemu_img convert -f qcow2 -O qcow2 createtmplt.sh qemu_img convert -f qcow2 -O qcow2 I'd suggest if we already know it's qcow2, and we want to create a qcow2 to copy the file. Objections? I don't know why we do that, but what the command is saying is that we want to convert an input image from qcow2 to keeping the same output format, -O qcow2 But yes, if the input format is already qcow2 and output is qcow2 we can just copy the file. Edison? Regards.
[jira] [Commented] (CLOUDSTACK-585) DHCP entry provisioning is broken in the KVM agent
[ https://issues.apache.org/jira/browse/CLOUDSTACK-585?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13510811#comment-13510811 ] Animesh Chaturvedi commented on CLOUDSTACK-585: --- I think it is caused by trailing space from the previous arguments, which can be fixed defensively by trimming the arguments in Java code and trimming the input at data entry DHCP entry provisioning is broken in the KVM agent -- Key: CLOUDSTACK-585 URL: https://issues.apache.org/jira/browse/CLOUDSTACK-585 Project: CloudStack Issue Type: Bug Security Level: Public(Anyone can view this level - this is the default.) Components: Hypervisor Controller, KVM Environment: CloudStack 3.0.2 KVM agent running on Fedora 14 Reporter: Wolfram Schlich Labels: agent, kvm Attachments: 1-dhcp_entry.sh.log.patch, 2-dhcp_entry.sh.fix.patch When adding an instance to a routerVM DHCP configuration, it seems that the KVM agent calls /usr/lib64/cloud/agent/scripts/network/domr/dhcp_entry.sh with wrongly constructed command line arguments, making the script fail to add correct entries (specifically default router, DNS servers and static routes) for that instance to the routerVM's /etc/dhcphosts.txt + /etc/dhcpopts.txt. Especially adding a specific default gateway fails, so the routerVM will always announce itself as the default router, because the correct entry in /etc/dhcpopts specifying the gateway of the instance's default network as the gateway is missing. This is especially nasty for non-default/additional networks of an instance, messing up the default routing. Examples: Management server log entry: 2012-11-29 14:36:37,764 DEBUG [resource.virtualnetwork.VirtualRoutingResource] (agentRequest-Handler-3:null) Executing: /usr/lib64/cloud/agent/./scripts/network/domr/dhcp_entry.sh -r 169.254.0.122 -v 172.31.2.233 -m 06:b5:88:00:02:30 -n vmname -d 172.31.2.1 -N 172.31.2.201 Notice the double spaces before -d and -N (and the extra space at the EOL) -- WARNING: you have to view the source of this description to actually _see_ the double spaces!) After patching /usr/lib64/cloud/agent/scripts/network/domr/dhcp_entry.sh to do meaningful logging, it's clear that the script does not get called with -d, but -d instead (same for -N), so with an extra space before the dash. Thus, getopts fails to parse/recognize these two arguments correctly and passed empty values for $dfltrt and $dns to the /root/edithosts.sh being called on the routerVM. It's also clear that the CloudStack KVM Java agent calls it with the wrongly constructed command line, because if a shell would interpret this command line, it would just ignore the extra spaces itself. I've not been able to dig it down, but I somehow suspect that one of ./utils/src/com/cloud/utils/script/Script.java:protected String buildCommandLine(String[] command) { } ./utils/src/com/cloud/utils/script/Script.java:public String execute(OutputInterpreter interpreter) { } might mess up building the command line of the command that had been built by ./core/src/com/cloud/agent/resource/virtualnetwork/VirtualRoutingResource.java:protected synchronized Answer execute (final DhcpEntryCommand cmd) { } before. I've not tried 4.0.0 so far, thus I cannot say whether it might be affected or not. As a workaround, I've patched dhcp_entry.sh to re-evaluate the positional parameters using 'set -- ${@}' (will attach a patch, also one for logging). -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Resolved] (CLOUDSTACK-536) remove citrix cloudpatform from 4.0 build - CloudStack is ASF project
[ https://issues.apache.org/jira/browse/CLOUDSTACK-536?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Brian Federle resolved CLOUDSTACK-536. -- Resolution: Fixed Committed to master: commit 9e90ff58e955c0908e14fa527995838f2a00b425 Author: Brian Federle brian.fede...@citrix.com Date: Wed Dec 5 13:18:25 2012 -0800 CLOUDSTACK-536: Remove legacy CloudPlatform files remove citrix cloudpatform from 4.0 build - CloudStack is ASF project - Key: CLOUDSTACK-536 URL: https://issues.apache.org/jira/browse/CLOUDSTACK-536 Project: CloudStack Issue Type: Bug Security Level: Public(Anyone can view this level - this is the default.) Components: UI Affects Versions: 4.0.0, 4.0.1 Environment: CS 4.0 from http://cloudstack.apt-get.eu/rhel/4.0/ The rpms are: cloud-utils-4.0.0-0.140.el6.4.0.x86_64 cloud-client-4.0.0-0.140.el6.4.0.x86_64 cloud-setup-4.0.0-0.140.el6.4.0.x86_64 cloud-scripts-4.0.0-0.140.el6.4.0.x86_64 cloud-deps-4.0.0-0.140.el6.4.0.x86_64 cloud-server-4.0.0-0.140.el6.4.0.x86_64 cloud-client-ui-4.0.0-0.140.el6.4.0.x86_64 cloud-python-4.0.0-0.140.el6.4.0.x86_64 cloud-core-4.0.0-0.140.el6.4.0.x86_64 cloud-usage-4.0.0-0.140.el6.4.0.x86_64 cloud-aws-api-4.0.0-0.140.el6.4.0.x86_64 Reporter: ilya musayev Assignee: Brian Federle Priority: Minor Fix For: 4.0.0, 4.0.1, 4.1.0 Please remove Citrix Logo and CloudPlaform from the UI page, files are (could be more): rpm/deb release: in /client/images/ logo-login.png citrix-logo-darkbg.png logo-cloudplatform.png and possibly others that may contain Citrix CloudPlatform mention. Also, please include the wording for CloudStack as Apache CloudStack in corresponding documentation and UI pages for CS ASF project. Thank you -ilya -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Comment Edited] (CLOUDSTACK-585) DHCP entry provisioning is broken in the KVM agent
[ https://issues.apache.org/jira/browse/CLOUDSTACK-585?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13510811#comment-13510811 ] Animesh Chaturvedi edited comment on CLOUDSTACK-585 at 12/5/12 10:01 PM: - I think it is caused by trailing space from the previous arguments, which can be fixed defensively by trimming the arguments in Java code and trimming the input at data entry. There is a discussion on dev-list to improve data validation and cases like this can be avoided was (Author: animeshc): I think it is caused by trailing space from the previous arguments, which can be fixed defensively by trimming the arguments in Java code and trimming the input at data entry DHCP entry provisioning is broken in the KVM agent -- Key: CLOUDSTACK-585 URL: https://issues.apache.org/jira/browse/CLOUDSTACK-585 Project: CloudStack Issue Type: Bug Security Level: Public(Anyone can view this level - this is the default.) Components: Hypervisor Controller, KVM Environment: CloudStack 3.0.2 KVM agent running on Fedora 14 Reporter: Wolfram Schlich Labels: agent, kvm Attachments: 1-dhcp_entry.sh.log.patch, 2-dhcp_entry.sh.fix.patch When adding an instance to a routerVM DHCP configuration, it seems that the KVM agent calls /usr/lib64/cloud/agent/scripts/network/domr/dhcp_entry.sh with wrongly constructed command line arguments, making the script fail to add correct entries (specifically default router, DNS servers and static routes) for that instance to the routerVM's /etc/dhcphosts.txt + /etc/dhcpopts.txt. Especially adding a specific default gateway fails, so the routerVM will always announce itself as the default router, because the correct entry in /etc/dhcpopts specifying the gateway of the instance's default network as the gateway is missing. This is especially nasty for non-default/additional networks of an instance, messing up the default routing. Examples: Management server log entry: 2012-11-29 14:36:37,764 DEBUG [resource.virtualnetwork.VirtualRoutingResource] (agentRequest-Handler-3:null) Executing: /usr/lib64/cloud/agent/./scripts/network/domr/dhcp_entry.sh -r 169.254.0.122 -v 172.31.2.233 -m 06:b5:88:00:02:30 -n vmname -d 172.31.2.1 -N 172.31.2.201 Notice the double spaces before -d and -N (and the extra space at the EOL) -- WARNING: you have to view the source of this description to actually _see_ the double spaces!) After patching /usr/lib64/cloud/agent/scripts/network/domr/dhcp_entry.sh to do meaningful logging, it's clear that the script does not get called with -d, but -d instead (same for -N), so with an extra space before the dash. Thus, getopts fails to parse/recognize these two arguments correctly and passed empty values for $dfltrt and $dns to the /root/edithosts.sh being called on the routerVM. It's also clear that the CloudStack KVM Java agent calls it with the wrongly constructed command line, because if a shell would interpret this command line, it would just ignore the extra spaces itself. I've not been able to dig it down, but I somehow suspect that one of ./utils/src/com/cloud/utils/script/Script.java:protected String buildCommandLine(String[] command) { } ./utils/src/com/cloud/utils/script/Script.java:public String execute(OutputInterpreter interpreter) { } might mess up building the command line of the command that had been built by ./core/src/com/cloud/agent/resource/virtualnetwork/VirtualRoutingResource.java:protected synchronized Answer execute (final DhcpEntryCommand cmd) { } before. I've not tried 4.0.0 so far, thus I cannot say whether it might be affected or not. As a workaround, I've patched dhcp_entry.sh to re-evaluate the positional parameters using 'set -- ${@}' (will attach a patch, also one for logging). -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Commented] (CLOUDSTACK-585) DHCP entry provisioning is broken in the KVM agent
[ https://issues.apache.org/jira/browse/CLOUDSTACK-585?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13510824#comment-13510824 ] Anthony Xu commented on CLOUDSTACK-585: --- Wolfram, Could you please check if there is trailing space in name, instance_name column of DB entry in vm_instance table for this VM? I suspect there is a trailing space. Anthony DHCP entry provisioning is broken in the KVM agent -- Key: CLOUDSTACK-585 URL: https://issues.apache.org/jira/browse/CLOUDSTACK-585 Project: CloudStack Issue Type: Bug Security Level: Public(Anyone can view this level - this is the default.) Components: Hypervisor Controller, KVM Environment: CloudStack 3.0.2 KVM agent running on Fedora 14 Reporter: Wolfram Schlich Labels: agent, kvm Attachments: 1-dhcp_entry.sh.log.patch, 2-dhcp_entry.sh.fix.patch When adding an instance to a routerVM DHCP configuration, it seems that the KVM agent calls /usr/lib64/cloud/agent/scripts/network/domr/dhcp_entry.sh with wrongly constructed command line arguments, making the script fail to add correct entries (specifically default router, DNS servers and static routes) for that instance to the routerVM's /etc/dhcphosts.txt + /etc/dhcpopts.txt. Especially adding a specific default gateway fails, so the routerVM will always announce itself as the default router, because the correct entry in /etc/dhcpopts specifying the gateway of the instance's default network as the gateway is missing. This is especially nasty for non-default/additional networks of an instance, messing up the default routing. Examples: Management server log entry: 2012-11-29 14:36:37,764 DEBUG [resource.virtualnetwork.VirtualRoutingResource] (agentRequest-Handler-3:null) Executing: /usr/lib64/cloud/agent/./scripts/network/domr/dhcp_entry.sh -r 169.254.0.122 -v 172.31.2.233 -m 06:b5:88:00:02:30 -n vmname -d 172.31.2.1 -N 172.31.2.201 Notice the double spaces before -d and -N (and the extra space at the EOL) -- WARNING: you have to view the source of this description to actually _see_ the double spaces!) After patching /usr/lib64/cloud/agent/scripts/network/domr/dhcp_entry.sh to do meaningful logging, it's clear that the script does not get called with -d, but -d instead (same for -N), so with an extra space before the dash. Thus, getopts fails to parse/recognize these two arguments correctly and passed empty values for $dfltrt and $dns to the /root/edithosts.sh being called on the routerVM. It's also clear that the CloudStack KVM Java agent calls it with the wrongly constructed command line, because if a shell would interpret this command line, it would just ignore the extra spaces itself. I've not been able to dig it down, but I somehow suspect that one of ./utils/src/com/cloud/utils/script/Script.java:protected String buildCommandLine(String[] command) { } ./utils/src/com/cloud/utils/script/Script.java:public String execute(OutputInterpreter interpreter) { } might mess up building the command line of the command that had been built by ./core/src/com/cloud/agent/resource/virtualnetwork/VirtualRoutingResource.java:protected synchronized Answer execute (final DhcpEntryCommand cmd) { } before. I've not tried 4.0.0 so far, thus I cannot say whether it might be affected or not. As a workaround, I've patched dhcp_entry.sh to re-evaluate the positional parameters using 'set -- ${@}' (will attach a patch, also one for logging). -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira
Connecting Netscaler VPX and CS4
What user permissions would I need? My Network Team would not give me admin priveleges - since I would not have admin in the production, so I need to know which perms are needed or commands used to create a custom perm set. Thanks ilya
[jira] [Commented] (CLOUDSTACK-335) KVM VPC load balancer not working
[ https://issues.apache.org/jira/browse/CLOUDSTACK-335?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13510892#comment-13510892 ] Joe Brockmeier commented on CLOUDSTACK-335: --- This has been pulled into 4.0.1 with commit 1217f5cd326c9a36d21b0136f76d77caceb09d84. KVM VPC load balancer not working - Key: CLOUDSTACK-335 URL: https://issues.apache.org/jira/browse/CLOUDSTACK-335 Project: CloudStack Issue Type: Bug Security Level: Public(Anyone can view this level - this is the default.) Components: KVM Affects Versions: pre-4.0.0, 4.0.0 Reporter: Marcus Sorensen Assignee: edison su Fix For: 4.0.0, 4.0.1, 4.1.0 Original Estimate: 4h Remaining Estimate: 4h The load balancer doesn't work in VPC for KVM. I've looked at the code and will fix this in the next day or two. We're just missing VPCLoadBalancerConfig, should be an easy fix. -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Commented] (CLOUDSTACK-576) PaaS Enablement: Composite Application Blueprints
[ https://issues.apache.org/jira/browse/CLOUDSTACK-576?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13510913#comment-13510913 ] Jie Feng commented on CLOUDSTACK-576: - Here is my take on Alex's questions: (1) I recommend something lightweight based on CloudStack concepts. We can use AWS CloudFormation's JS format as a reference starting point and make changes where needed. (2) For v1.0, at minimum, the ability to launch a set of VMs. The blueprint should capture VM deployment properties like template ID/network, startup order. (3) Scripts, Chef and Puppet can be used to install software and configure VMs after they are launched. You can already do this today within a running VM. In my opinion, deeper integration should be optional. This feature should be able to use CloudStack's credential and integrate into the CloudStack UI (through UI plugin or other means). Jie PaaS Enablement: Composite Application Blueprints - Key: CLOUDSTACK-576 URL: https://issues.apache.org/jira/browse/CLOUDSTACK-576 Project: CloudStack Issue Type: New Feature Security Level: Public(Anyone can view this level - this is the default.) Components: API, AWSAPI Reporter: Duncan Johnston-Watt Given the level of interest in CloudStack as a platform for private, hybrid and public cloud one of the gaps is support for composite/multi-tier application blueprints comparable to VMware vApp and CloudFormation templates without necessarily slavishly following either of them. This is almost certainly a new component. One that will leverage/enhance the API/AWSAPI components amongst others. -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Commented] (CLOUDSTACK-105) /tmp/stream-unix.####.###### stale sockets causing inodes to run out on Xenserver
[ https://issues.apache.org/jira/browse/CLOUDSTACK-105?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13510950#comment-13510950 ] Jason Bausewein commented on CLOUDSTACK-105: I am able to reproduce this issue consistently. I have a single xen host in a basic zone. I tracked it down to the following process creating the stream-unix.. files about every 10 seconds. root 8237 8223 0 09:59 ?00:00:00 ovs-vsctl add-br xapi0 Dec 5 09:59:15 xenserver1 ovs-vsctl: 1|vsctl|INFO|Called as ovs-vsctl add-br xapi0 Dec 5 09:59:15 xenserver1 ovs-vsctl: 2|stream_unix|ERR|/tmp/stream-unix.8237.0: connection to /var/run/openvswitch/db.sock failed: No such file or directory Dec 5 09:59:15 xenserver1 ovs-vsctl: 3|reconnect|WARN|unix:/var/run/openvswitch/db.sock: connection attempt failed (No such file or directory) Dec 5 09:59:16 xenserver1 ovs-vsctl: 4|stream_unix|ERR|/tmp/stream-unix.8237.1: connection to /var/run/openvswitch/db.sock failed: No such file or directory Dec 5 09:59:16 xenserver1 ovs-vsctl: 5|reconnect|WARN|unix:/var/run/openvswitch/db.sock: connection attempt failed (No such file or directory) Dec 5 09:59:18 xenserver1 ovs-vsctl: 6|stream_unix|ERR|/tmp/stream-unix.8237.2: connection to /var/run/openvswitch/db.sock failed: No such file or directory Dec 5 09:59:18 xenserver1 ovs-vsctl: 7|reconnect|WARN|unix:/var/run/openvswitch/db.sock: connection attempt failed (No such file or directory) I am not using openvswitch. Is there something I need to turn off? /tmp/stream-unix..## stale sockets causing inodes to run out on Xenserver - Key: CLOUDSTACK-105 URL: https://issues.apache.org/jira/browse/CLOUDSTACK-105 Project: CloudStack Issue Type: Bug Security Level: Public(Anyone can view this level - this is the default.) Components: XenServer Affects Versions: pre-4.0.0 Environment: Xenserver 6.0.2 Cloudstack 3.0.2 Reporter: Caleb Call Assignee: Devdeep Singh Fix For: 4.1.0 We came across an interesting issue in one of our clusters. We ran out of inodes on all of our cluster members (since when does this happen in 2012?). When this happened, it in turn made the / filesystem a read-only filesystem which in turn made all the hosts go in to emergency maintenance mode and as a result get marked down by Cloudstack. We found that it was caused by hundreds of thousands of stale socket files in /tmp named stream-unix..##. To resolve the issue, we had to delete those stale socket files (find /tmp -name *stream* -mtime +7 -exec rm -v {} \;), then kill and restart xapi, then correct the emergency maintenance mode. These hosts had only been up for 45 days before this issue occurred. In our scouring of the interwebs, the only other instance we've been able to find of this (or similar) happening is in the same setup we are currently running. Xenserver 6.0.2 with CS 3.0.2. Do these stream-unix sockets have anything to do with Cloudstack? I would think if this was a Xenserver issue (bug), there would be a lot more on the internet about this happening. For a temporary workaround, we've added a cronjob to cleanup these files but we'd really like to address the actual issue that's causing these sockets to become stale and not get cleaned-up. -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira
Re: CentOS System VM?
The current system vm is getting long in the tooth. I (or Rohit Yadav) will looking into building a wheezy-based systemvm that includes hyper-v drivers. Hopefully network throughput should be better as well when used with multiple cores. On 12/5/12 10:38 AM, Jason Davis scr...@gmail.com wrote: TBH Hyper-V synthetic drivers(modules) is supported in the mainline kernel. So the argument that CentOS 6.x has better support is moot. This assumes that the kernel version on the SSVM is at least 2.6.32. I ran Ubuntu Server 11.x and Centos 6.x on Hyper-V natively and just needed to load the kernel modules for the synthetic stuffs to work. Ancient example of getting the Hyper-V modules built/working on Debian 6.0 http://virtualisationandmanagement.wordpress.com/2010/11/02/debian-on-hype r-v-with-4-vcpu-support-and-syntetic-network/ On Wed, Dec 5, 2012 at 12:25 PM, Kelceydamage@bbits kel...@bbits.ca wrote: I'm very interested in this. Sent from my iPhone On Dec 5, 2012, at 10:10 AM, Donal Lafferty donal.laffe...@citrix.com wrote: Has anyone looked into building a system VM that runs on a CentOS distro?
Re: API refactoring work for 4.1.0
On Wed, Dec 05, 2012 at 03:15:11AM +0530, Andrew Bayer wrote: Seems to me that coming up with a general API validation test suite (using jclouds, Marvin, whatever) that verifies consistency of behavior across API versions would be a very handy thing to have. A. My idea is to use marvin to build a basic test suite to validate that none of the API contracts are broken. I might be able to get to this work going next week. There's a lot of APIs to go through though so I'll be calling out for help on the lists. I'll publish the first set of tests so others can pick up after me. On Tue, Dec 4, 2012 at 11:52 AM, David Nalley da...@gnsa.us wrote: Please go through the api fs proposal. Suggestions, proposals, feedback, flames? Sorry for not going through this earlier. The unit tests and testing in general concern me. So let me ask - what is the plan to test that the refactored API behaves the same as the current API? --David -- Prasanna.,
Re: Review Request: List API performance optimization part I.
--- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/8172/#review14081 --- Ship it! Please git pull latest api refactoring, the patch was big about 5-6k lines and huge merge conflicts, took me some time to merge them. Please check the code, in case I missed something. The build system is broken now, we need to fix it. Committed in api_refactoring; commit b0ce8fd4ff512c7a7da0a71671c4477492e854e7 Author: Min Chen min.c...@citrix.com Date: Wed Dec 5 16:18:38 2012 -0800 api: Optimize and improve api, db call perfomance This is part 1 of list API refactoring. Commands covered: listVmsCmd, listRoutersCmd Response covered: UserVmResponse, DomainRouterResponse. DB views created: user_vm_view, domain_router_view. Signed-off-by: Rohit Yadav bhais...@apache.org - Rohit Yadav On Nov. 29, 2012, 6:57 p.m., Min Chen wrote: --- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/8172/ --- (Updated Nov. 29, 2012, 6:57 p.m.) Review request for cloudstack and Prachi Damle. Description --- This is part 1 of list API refactoring. Commands covered: listVmsCmd, listRoutersCmd Response covered: UserVmResponse, DomainRouterResponse. DB views created: user_vm_view, domain_router_view. This addresses bug https://issues.apache.org/jira/browse/CLOUDSTACK-527. Diffs - api/src/com/cloud/api/ResponseGenerator.java 4e8fbd8 api/src/com/cloud/api/ResponseObject.java 2d08fb9 api/src/com/cloud/api/commands/ListRoutersCmd.java 8bf9ba8 api/src/com/cloud/api/commands/ListVMsCmd.java 2f6f988 api/src/com/cloud/api/response/BaseResponse.java e343a10 api/src/com/cloud/api/response/ControlledViewEntityResponse.java PRE-CREATION api/src/com/cloud/api/response/DomainRouterResponse.java d710aad api/src/com/cloud/api/response/NicResponse.java 69d5c31 api/src/com/cloud/api/response/UserVmResponse.java f74c072 api/src/com/cloud/api/view/vo/ControlledViewEntity.java PRE-CREATION api/src/com/cloud/api/view/vo/DomainRouterJoinVO.java PRE-CREATION api/src/com/cloud/api/view/vo/UserVmJoinVO.java PRE-CREATION api/src/com/cloud/server/ManagementService.java 7532cae api/src/com/cloud/vm/UserVmService.java 98d02db api/test/src/com/cloud/api/commands/test/ListRoutersCmdTest.java PRE-CREATION api/test/src/com/cloud/api/commands/test/ListVmsCmdTest.java PRE-CREATION server/src/com/cloud/api/ApiDBUtils.java 3b5f634 server/src/com/cloud/api/ApiResponseHelper.java ebe8415 server/src/com/cloud/api/ApiServer.java a5c9ea5 server/src/com/cloud/api/response/ApiResponseSerializer.java 4be5dfa server/src/com/cloud/configuration/DefaultComponentLibrary.java ef61044 server/src/com/cloud/server/ManagementServerImpl.java 117be57 server/src/com/cloud/user/AccountManager.java 90a34ad server/src/com/cloud/user/AccountManagerImpl.java f595478 server/src/com/cloud/vm/UserVmManager.java 4ce9bfe server/src/com/cloud/vm/UserVmManagerImpl.java 687f521 server/src/com/cloud/vm/dao/DomainRouterJoinDao.java PRE-CREATION server/src/com/cloud/vm/dao/DomainRouterJoinDaoImpl.java PRE-CREATION server/src/com/cloud/vm/dao/UserVmJoinDao.java PRE-CREATION server/src/com/cloud/vm/dao/UserVmJoinDaoImpl.java PRE-CREATION server/test/com/cloud/keystore/KeystoreTest.java e0e2126 server/test/com/cloud/user/MockAccountManagerImpl.java 08234fd server/test/com/cloud/vm/MockUserVmManagerImpl.java 35ee139 setup/db/create-schema.sql fff084e utils/src/com/cloud/utils/db/GenericDaoBase.java 8d5cb96 Diff: https://reviews.apache.org/r/8172/diff/ Testing --- Create a performance unit test class to test the performance time. Thanks, Min Chen
[jira] [Created] (CLOUDSTACK-589) Juniper SRX does not depends on f5 iControl jar
Hiroaki Kawai created CLOUDSTACK-589: Summary: Juniper SRX does not depends on f5 iControl jar Key: CLOUDSTACK-589 URL: https://issues.apache.org/jira/browse/CLOUDSTACK-589 Project: CloudStack Issue Type: Bug Security Level: Public (Anyone can view this level - this is the default.) Components: Network Devices Affects Versions: 4.0.0, 4.0.1, 4.1.0 Reporter: Hiroaki Kawai In plugins / network-elements / juniper-srx / pom.xml, there's a dependency element which says srx requires icontrol. But the code does not use icontrol at all, and it is safe to remove this dependency from here. -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Updated] (CLOUDSTACK-589) Juniper SRX does not depends on f5 iControl jar
[ https://issues.apache.org/jira/browse/CLOUDSTACK-589?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Hiroaki Kawai updated CLOUDSTACK-589: - Attachment: srx_pom.patch Juniper SRX does not depends on f5 iControl jar --- Key: CLOUDSTACK-589 URL: https://issues.apache.org/jira/browse/CLOUDSTACK-589 Project: CloudStack Issue Type: Bug Security Level: Public(Anyone can view this level - this is the default.) Components: Network Devices Affects Versions: 4.0.0, 4.0.1, 4.1.0 Reporter: Hiroaki Kawai Labels: maven, patch Attachments: srx_pom.patch In plugins / network-elements / juniper-srx / pom.xml, there's a dependency element which says srx requires icontrol. But the code does not use icontrol at all, and it is safe to remove this dependency from here. -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira
RE: [Discuss] Style Guide for Apache CloudStack Documentation ?
Thank you Animesh. This should help us. -Original Message- From: Animesh Chaturvedi [mailto:animesh.chaturv...@citrix.com] Sent: Wednesday, December 05, 2012 12:08 AM To: cloudstack-dev@incubator.apache.org Subject: RE: [Discuss] Style Guide for Apache CloudStack Documentation ? Radhika The Struts style guide has more details http://struts.apache.org/2.2.3/docs/documentation-style-guide.html and we can borrow some tips from there. One other thing I think will help users is quick link to Single Page Html. This is useful because people often want to search for a specific word or phrase across the entire documentation. Generally, they already know what they're looking for; they just can't remember what section it's in. The separate-page style is useful for those who already know what section they need, or who want to read the entire documentation from front to back in sequence. But this is not the most common way documentation is accessed. Thanks Animesh -Original Message- From: Radhika Puthiyetath [mailto:radhika.puthiyet...@citrix.com] Sent: Tuesday, December 04, 2012 6:11 AM To: cloudstack-dev@incubator.apache.org Subject: [Discuss] Style Guide for Apache CloudStack Documentation ? Hello Community, Any thoughts on the need of a style guide for the CloudStack Documentation ? First of all, do we need one-If yes, do we need to create afresh , or follow the existing style guides within Apache projects, such as Maven. Here is the link: http://maven.apache.org/guides/development/guide-documentation-style.html. Please educate me on the Apache way for Documentation style. Alternatively, we can always follow what XenServer documentation is following. Thanks -Radhika PS: I do understand that enhancing the documentation/ fixing the doc defect is the first priority of the doc contributors.
Re: CentOS System VM?
Granted both debian wheezy and centos/rhel 6, run the same major kernel version 2.6.32 i386 and hopefully same glibc library (need to confirm) - it should be really easy to port the code over without needing to recompile anything. I don't believe we do anything overly complex - within application code - that would glue components to specific OS. Applications like apache, dns masq, haproxy, sshd and dhcp can be stock versions of what OS vendor released. This is my observations so far and I will give it a shot when time allows to confirm. If someone knows of reason why this would fail, please let me know so I don't waste my time:) but I am fairly optimistic. Thanks Ilya Kelceydamage@bbits kel...@bbits.ca wrote: This sounds great, however I am hoping for updated wiki on how to create your own system vm(distro). Maybe one day Sent from my iPhone On Dec 5, 2012, at 5:04 PM, Chiradeep Vittal chiradeep.vit...@citrix.com wrote: The current system vm is getting long in the tooth. I (or Rohit Yadav) will looking into building a wheezy-based systemvm that includes hyper-v drivers. Hopefully network throughput should be better as well when used with multiple cores. On 12/5/12 10:38 AM, Jason Davis scr...@gmail.com wrote: TBH Hyper-V synthetic drivers(modules) is supported in the mainline kernel. So the argument that CentOS 6.x has better support is moot. This assumes that the kernel version on the SSVM is at least 2.6.32. I ran Ubuntu Server 11.x and Centos 6.x on Hyper-V natively and just needed to load the kernel modules for the synthetic stuffs to work. Ancient example of getting the Hyper-V modules built/working on Debian 6.0 http://virtualisationandmanagement.wordpress.com/2010/11/02/debian-on-hype r-v-with-4-vcpu-support-and-syntetic-network/ On Wed, Dec 5, 2012 at 12:25 PM, Kelceydamage@bbits kel...@bbits.ca wrote: I'm very interested in this. Sent from my iPhone On Dec 5, 2012, at 10:10 AM, Donal Lafferty donal.laffe...@citrix.com wrote: Has anyone looked into building a system VM that runs on a CentOS distro?
[jira] [Assigned] (CLOUDSTACK-488) can't register awsapi user
[ https://issues.apache.org/jira/browse/CLOUDSTACK-488?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Rajesh Battala reassigned CLOUDSTACK-488: - Assignee: Likitha Shetty can't register awsapi user -- Key: CLOUDSTACK-488 URL: https://issues.apache.org/jira/browse/CLOUDSTACK-488 Project: CloudStack Issue Type: Bug Security Level: Public(Anyone can view this level - this is the default.) Components: AWSAPI Affects Versions: 4.1.0 Environment: OSX 10.8.1 mysql55 maven3 devcloud Reporter: sebastien goasguen Assignee: Likitha Shetty Running devcloud host-only environment as described in: https://cwiki.apache.org/CLOUDSTACK/cloudstack-devcloud-environment-setup.html Run the awsapi server with: mvn -pl :cloud-awsapi jetty:run Then tried to register a user: air-2:setup sebastiengoasguen$ python ./cloudstack-aws-api-register -a 2GmlMc59ItbOE4EVbwVIqyO25nmZWOrNWRhWi-AvqCmNZ9H6FBXmpTycts5KnV6CmeCaz9yZjdnovURNWSoDuw -s GKsUGz_TvdN5U7cLNQoIElmwQI7AVnFcGK0Wo7SdMkJevjmJ5TN7theFnpVqKYlGAs7n_6gODYvtIOcKZLifbw -c ~/.ssh/cert.pem -u http://localhost:7080/awsapi/ User registration failed with http error code: 401 Jetty logs show: EC2RestServlet.doGetOrPost: javax.servlet.forward.query_string: AWSAccessKeyId=2GmlMc59ItbOE4EVbwVIqyO25nmZWOrNWRhWi-AvqCmNZ9H6FBXmpTycts5KnV6CmeCaz9yZjdnovURNWSoDuwAction=SetUserKeysSignatureMethod=HmacSHA1SignatureVersion=2Timestamp=2012-11-15T11:10:12ZVersion=2010-11-15accesskey=2GmlMc59ItbOE4EVbwVIqyO25nmZWOrNWRhWi-AvqCmNZ9H6FBXmpTycts5KnV6CmeCaz9yZjdnovURNWSoDuwsecretkey=GKsUGz_TvdN5U7cLNQoIElmwQI7AVnFcGK0Wo7SdMkJevjmJ5TN7theFnpVqKYlGAs7n_6gODYvtIOcKZLifbwSignature=lOgFvzNy0t2%2FCBzfKZB3W84lx38%3D Nov 15, 2012 11:10:13 AM com.cloud.bridge.service.EC2RestServlet logRequest INFO: EC2 Request method: GET Nov 15, 2012 11:10:13 AM com.cloud.bridge.service.EC2RestServlet logRequest INFO: Request contextPath: /awsapi Nov 15, 2012 11:10:13 AM com.cloud.bridge.service.EC2RestServlet logRequest INFO: Request pathInfo: / Nov 15, 2012 11:10:13 AM com.cloud.bridge.service.EC2RestServlet logRequest INFO: Request pathTranslated: /Users/sebastiengoasguen/Documents/incubator-cloudstack/awsapi/target/cloud-awsapi-4.1.0-SNAPSHOT Nov 15, 2012 11:10:13 AM com.cloud.bridge.service.EC2RestServlet logRequest INFO: Request queryString: AWSAccessKeyId=2GmlMc59ItbOE4EVbwVIqyO25nmZWOrNWRhWi-AvqCmNZ9H6FBXmpTycts5KnV6CmeCaz9yZjdnovURNWSoDuwAction=SetUserKeysSignatureMethod=HmacSHA1SignatureVersion=2Timestamp=2012-11-15T11:10:12ZVersion=2010-11-15accesskey=2GmlMc59ItbOE4EVbwVIqyO25nmZWOrNWRhWi-AvqCmNZ9H6FBXmpTycts5KnV6CmeCaz9yZjdnovURNWSoDuwsecretkey=GKsUGz_TvdN5U7cLNQoIElmwQI7AVnFcGK0Wo7SdMkJevjmJ5TN7theFnpVqKYlGAs7n_6gODYvtIOcKZLifbwSignature=lOgFvzNy0t2%2FCBzfKZB3W84lx38%3D Nov 15, 2012 11:10:13 AM com.cloud.bridge.service.EC2RestServlet logRequest INFO: Request requestURI: /awsapi/rest/AmazonEC2/ Nov 15, 2012 11:10:13 AM com.cloud.bridge.service.EC2RestServlet logRequest INFO: Request requestURL: http://localhost:7080/awsapi/rest/AmazonEC2/ Nov 15, 2012 11:10:13 AM com.cloud.bridge.service.EC2RestServlet logRequest INFO: Request servletPath: /rest/AmazonEC2 Nov 15, 2012 11:10:13 AM com.cloud.bridge.service.EC2RestServlet logRequest INFO: Request header Accept-Encoding:identity Nov 15, 2012 11:10:13 AM com.cloud.bridge.service.EC2RestServlet logRequest INFO: Request header Host:localhost:7080 Nov 15, 2012 11:10:13 AM com.cloud.bridge.service.EC2RestServlet logRequest INFO: Request header Connection:close Nov 15, 2012 11:10:13 AM com.cloud.bridge.service.EC2RestServlet logRequest INFO: Request header User-Agent:Python-urllib/2.7 Nov 15, 2012 11:10:13 AM com.cloud.bridge.service.EC2RestServlet logRequest INFO: Request parameter Action:SetUserKeys Nov 15, 2012 11:10:13 AM com.cloud.bridge.service.EC2RestServlet logRequest INFO: Request parameter SignatureMethod:HmacSHA1 Nov 15, 2012 11:10:13 AM com.cloud.bridge.service.EC2RestServlet logRequest INFO: Request parameter AWSAccessKeyId:2GmlMc59ItbOE4EVbwVIqyO25nmZWOrNWRhWi-AvqCmNZ9H6FBXmpTycts5KnV6CmeCaz9yZjdnovURNWSoDuw Nov 15, 2012 11:10:13 AM com.cloud.bridge.service.EC2RestServlet logRequest INFO: Request parameter Version:2010-11-15 Nov 15, 2012 11:10:13 AM com.cloud.bridge.service.EC2RestServlet logRequest INFO: Request parameter SignatureVersion:2 Nov 15, 2012 11:10:13 AM com.cloud.bridge.service.EC2RestServlet logRequest INFO: Request parameter Signature:lOgFvzNy0t2/CBzfKZB3W84lx38= Nov 15, 2012 11:10:13 AM com.cloud.bridge.service.EC2RestServlet logRequest INFO: Request parameter secretkey:GKsUGz_TvdN5U7cLNQoIElmwQI7AVnFcGK0Wo7SdMkJevjmJ5TN7theFnpVqKYlGAs7n_6gODYvtIOcKZLifbw Nov 15, 2012 11:10:13 AM com.cloud.bridge.service.EC2RestServlet logRequest INFO: Request parameter
[jira] [Commented] (CLOUDSTACK-488) can't register awsapi user
[ https://issues.apache.org/jira/browse/CLOUDSTACK-488?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13511206#comment-13511206 ] Rajesh Battala commented on CLOUDSTACK-488: --- Likitha had fixed, she will commit it ASF. aswsapi db property is missing so awsaspi service is not able to connect to the db. can't register awsapi user -- Key: CLOUDSTACK-488 URL: https://issues.apache.org/jira/browse/CLOUDSTACK-488 Project: CloudStack Issue Type: Bug Security Level: Public(Anyone can view this level - this is the default.) Components: AWSAPI Affects Versions: 4.1.0 Environment: OSX 10.8.1 mysql55 maven3 devcloud Reporter: sebastien goasguen Running devcloud host-only environment as described in: https://cwiki.apache.org/CLOUDSTACK/cloudstack-devcloud-environment-setup.html Run the awsapi server with: mvn -pl :cloud-awsapi jetty:run Then tried to register a user: air-2:setup sebastiengoasguen$ python ./cloudstack-aws-api-register -a 2GmlMc59ItbOE4EVbwVIqyO25nmZWOrNWRhWi-AvqCmNZ9H6FBXmpTycts5KnV6CmeCaz9yZjdnovURNWSoDuw -s GKsUGz_TvdN5U7cLNQoIElmwQI7AVnFcGK0Wo7SdMkJevjmJ5TN7theFnpVqKYlGAs7n_6gODYvtIOcKZLifbw -c ~/.ssh/cert.pem -u http://localhost:7080/awsapi/ User registration failed with http error code: 401 Jetty logs show: EC2RestServlet.doGetOrPost: javax.servlet.forward.query_string: AWSAccessKeyId=2GmlMc59ItbOE4EVbwVIqyO25nmZWOrNWRhWi-AvqCmNZ9H6FBXmpTycts5KnV6CmeCaz9yZjdnovURNWSoDuwAction=SetUserKeysSignatureMethod=HmacSHA1SignatureVersion=2Timestamp=2012-11-15T11:10:12ZVersion=2010-11-15accesskey=2GmlMc59ItbOE4EVbwVIqyO25nmZWOrNWRhWi-AvqCmNZ9H6FBXmpTycts5KnV6CmeCaz9yZjdnovURNWSoDuwsecretkey=GKsUGz_TvdN5U7cLNQoIElmwQI7AVnFcGK0Wo7SdMkJevjmJ5TN7theFnpVqKYlGAs7n_6gODYvtIOcKZLifbwSignature=lOgFvzNy0t2%2FCBzfKZB3W84lx38%3D Nov 15, 2012 11:10:13 AM com.cloud.bridge.service.EC2RestServlet logRequest INFO: EC2 Request method: GET Nov 15, 2012 11:10:13 AM com.cloud.bridge.service.EC2RestServlet logRequest INFO: Request contextPath: /awsapi Nov 15, 2012 11:10:13 AM com.cloud.bridge.service.EC2RestServlet logRequest INFO: Request pathInfo: / Nov 15, 2012 11:10:13 AM com.cloud.bridge.service.EC2RestServlet logRequest INFO: Request pathTranslated: /Users/sebastiengoasguen/Documents/incubator-cloudstack/awsapi/target/cloud-awsapi-4.1.0-SNAPSHOT Nov 15, 2012 11:10:13 AM com.cloud.bridge.service.EC2RestServlet logRequest INFO: Request queryString: AWSAccessKeyId=2GmlMc59ItbOE4EVbwVIqyO25nmZWOrNWRhWi-AvqCmNZ9H6FBXmpTycts5KnV6CmeCaz9yZjdnovURNWSoDuwAction=SetUserKeysSignatureMethod=HmacSHA1SignatureVersion=2Timestamp=2012-11-15T11:10:12ZVersion=2010-11-15accesskey=2GmlMc59ItbOE4EVbwVIqyO25nmZWOrNWRhWi-AvqCmNZ9H6FBXmpTycts5KnV6CmeCaz9yZjdnovURNWSoDuwsecretkey=GKsUGz_TvdN5U7cLNQoIElmwQI7AVnFcGK0Wo7SdMkJevjmJ5TN7theFnpVqKYlGAs7n_6gODYvtIOcKZLifbwSignature=lOgFvzNy0t2%2FCBzfKZB3W84lx38%3D Nov 15, 2012 11:10:13 AM com.cloud.bridge.service.EC2RestServlet logRequest INFO: Request requestURI: /awsapi/rest/AmazonEC2/ Nov 15, 2012 11:10:13 AM com.cloud.bridge.service.EC2RestServlet logRequest INFO: Request requestURL: http://localhost:7080/awsapi/rest/AmazonEC2/ Nov 15, 2012 11:10:13 AM com.cloud.bridge.service.EC2RestServlet logRequest INFO: Request servletPath: /rest/AmazonEC2 Nov 15, 2012 11:10:13 AM com.cloud.bridge.service.EC2RestServlet logRequest INFO: Request header Accept-Encoding:identity Nov 15, 2012 11:10:13 AM com.cloud.bridge.service.EC2RestServlet logRequest INFO: Request header Host:localhost:7080 Nov 15, 2012 11:10:13 AM com.cloud.bridge.service.EC2RestServlet logRequest INFO: Request header Connection:close Nov 15, 2012 11:10:13 AM com.cloud.bridge.service.EC2RestServlet logRequest INFO: Request header User-Agent:Python-urllib/2.7 Nov 15, 2012 11:10:13 AM com.cloud.bridge.service.EC2RestServlet logRequest INFO: Request parameter Action:SetUserKeys Nov 15, 2012 11:10:13 AM com.cloud.bridge.service.EC2RestServlet logRequest INFO: Request parameter SignatureMethod:HmacSHA1 Nov 15, 2012 11:10:13 AM com.cloud.bridge.service.EC2RestServlet logRequest INFO: Request parameter AWSAccessKeyId:2GmlMc59ItbOE4EVbwVIqyO25nmZWOrNWRhWi-AvqCmNZ9H6FBXmpTycts5KnV6CmeCaz9yZjdnovURNWSoDuw Nov 15, 2012 11:10:13 AM com.cloud.bridge.service.EC2RestServlet logRequest INFO: Request parameter Version:2010-11-15 Nov 15, 2012 11:10:13 AM com.cloud.bridge.service.EC2RestServlet logRequest INFO: Request parameter SignatureVersion:2 Nov 15, 2012 11:10:13 AM com.cloud.bridge.service.EC2RestServlet logRequest INFO: Request parameter Signature:lOgFvzNy0t2/CBzfKZB3W84lx38= Nov 15, 2012 11:10:13 AM com.cloud.bridge.service.EC2RestServlet logRequest INFO: Request parameter