[
https://issues.apache.org/jira/browse/CLOUDSTACK-8485?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14991479#comment-14991479
]
ASF GitHub Bot commented on CLOUDSTACK-8485:
--------------------------------------------
Github user koushik-das commented on a diff in the pull request:
https://github.com/apache/cloudstack/pull/1021#discussion_r43996820
--- Diff: server/src/com/cloud/api/ApiResponseGsonHelper.java ---
@@ -27,30 +29,40 @@
import com.google.gson.GsonBuilder;
/**
- * The ApiResonseGsonHelper is different from ApiGsonHelper - it
registeres one more adapter for String type required for api response encoding
+ * The ApiResonseGsonHelper is different from ApiGsonHelper - it registers
one more adapter for String type required for api response encoding
--- End diff --
I have to read more on the gson changes you are referring to. But I think
we shouldn't mix multiple things with this. The primary purpose of this PR is
to fix the perf. issue in list API responses.
There may be some rework if gson is upgraded but it is better to do that
when gson upgrade work is done.
> listAPIs are taking too long to return results
> ----------------------------------------------
>
> Key: CLOUDSTACK-8485
> URL: https://issues.apache.org/jira/browse/CLOUDSTACK-8485
> Project: CloudStack
> Issue Type: Bug
> Security Level: Public(Anyone can view this level - this is the
> default.)
> Components: Management Server
> Affects Versions: 4.5.1, 4.6.0
> Reporter: Sowmya Krishnan
> Assignee: Koushik Das
> Fix For: 4.6.0
>
>
> listAPIs are taking significantly longer than before (4.2.x)
> I tried out few listAPI calls using a simulator set up with ~ 10K VMs and 8K
> Routers. Here are few results:
> listVirtualMachines is taking > 25 sec to return with pagesize set to 50.
> This is in comparison to 2 sec in earlier cases such as 4.2.
> listVolumes with pagesize = 1000 took more than 10 mins and finally times out.
> Further observations show that there are also lot of slow queries being
> logged in catalina.out and in MySQL slow query logs. I am not sure if this
> could be the reason for DB performance getting impacted in turn causing an
> impact on listAPIs too.
> Here's a sample of slow queries from catalina.out:
> Mon May 11 07:31:15 UTC 2015 INFO: Profiler Event: [SLOW QUERY] at
> org.apache.commons.dbcp.DelegatingPreparedStatement.executeQuery(DelegatingPreparedStatement.java:96)
> duration: 3305 ms, connection-id: 637759, statement-id: 3218312,
> resultset-id: 0, message: Slow query (exceeded 2,000 ms, duration: 3,305
> ms):SELECT user_vm_details.id, user_vm_details.vm_id, user_vm_details.name,
> user_vm_details.value, user_vm_details.display FROM user_vm_details WHERE
> user_vm_details.vm_id = 9117Mon May 11 07:31:15 UTC 2015 INFO: Profiler
> Event: [SLOW QUERY] at
> org.apache.commons.dbcp.DelegatingPreparedStatement.executeQuery(DelegatingPreparedStatement.java:96)
> duration: 3305 ms, connection-id: 637843, statement-id: 3218311,
> resultset-id: 0, message: Slow query (exceeded 2,000 ms, duration: 3,305
> ms):SELECT host.id, host.disconnected, host.name, host.status, host.type,
> host.private_ip_address, host.private_mac_address, host.private_netmask,
> host.public_netmask, host.public_ip_address, host.public_mac_address,
> host.storage_ip_address, host.cluster_id, host.storage_netmask,
> host.storage_mac_address, host.storage_ip_address_2, host.storage_netmask_2,
> host.storage_mac_address_2, host.hypervisor_type, host.proxy_port,
> host.resource, host.fs_type, host.available, host.setup, host.resource_state,
> host.hypervisor_version, host.update_count, host.uuid, host.data_center_id,
> host.pod_id, host.cpu_sockets, host.cpus, host.url, host.speed, host.ram,
> host.parent, host.guid, host.capabilities, host.total_size, host.last_ping,
> host.mgmt_server_id, host.dom0_memory, host.version, host.created,
> host.removed FROM host WHERE host.id = 345 AND host.removed IS NULL
> Mon May 11 07:31:17 UTC 2015 INFO: Profiler Event: [SLOW QUERY] at
> org.apache.commons.dbcp.DelegatingPreparedStatement.executeQuery(DelegatingPreparedStatement.java:96)
> duration: 6458 ms, connection-id: 637623, statement-id: 3218243,
> resultset-id: 0, message: Slow query (exceeded 2,000 ms, duration: 6,458
> ms):SELECT storage_pool_host_ref.host_id FROM storage_pool_host_ref INNER
> JOIN host ON storage_pool_host_ref.host_id=host.id WHERE
> storage_pool_host_ref.pool_id = 197 AND (host.status = 'Up' AND
> host.resource_state = 'Enabled' )Mon May 11 07:31:17 UTC 2015 INFO: Profiler
> Event: [SLOW QUERY] at
> org.apache.commons.dbcp.DelegatingPreparedStatement.executeQuery(DelegatingPreparedStatement.java:96)
> duration: 2402 ms, connection-id: 637754, statement-id: 3218371,
> resultset-id: 0, message: Slow query (exceeded 2,000 ms, duration: 2,402
> ms):SELECT host.id, host.disconnected, host.name, host.status, host.type,
> host.private_ip_address, host.private_mac_address, host.private_netmask,
> host.public_netmask, host.public_ip_address, host.public_mac_address,
> host.storage_ip_address, host.cluster_id, host.storage_netmask,
> host.storage_mac_address, host.storage_ip_address_2, host.storage_netmask_2,
> host.storage_mac_address_2, host.hypervisor_type, host.proxy_port,
> host.resource, host.fs_type, host.available, host.setup, host.resource_state,
> host.hypervisor_version, host.update_count, host.uuid, host.data_center_id,
> host.pod_id, host.cpu_sockets, host.cpus, host.url, host.speed, host.ram,
> host.parent, host.guid, host.capabilities, host.total_size, host.last_ping,
> host.mgmt_server_id, host.dom0_memory, host.version, host.created,
> host.removed FROM host WHERE host.id = 669 AND host.removed IS NULL
> The following is from MySQL slow query log:
> SELECT vm_instance.id, vm_instance.name, vm_instance.vnc_password,
> vm_instance.proxy_id, vm_instance.proxy_assign_time, vm_instance.state,
> vm_instance.private_ip_address, vm_instance.instance_name,
> vm_instance.vm_template_id, vm_instance.guest_os_id, vm_instance.host_id,
> vm_instance.last_host_id, vm_instance.pod_id,
> vm_instance.private_mac_address, vm_instance.data_center_id,
> vm_instance.vm_type, vm_instance.ha_enabled, vm_instance.display_vm,
> vm_instance.limit_cpu_use, vm_instance.update_count, vm_instance.created,
> vm_instance.removed, vm_instance.update_time, vm_instance.domain_id,
> vm_instance.account_id, vm_instance.service_offering_id,
> vm_instance.reservation_id, vm_instance.hypervisor_type,
> vm_instance.dynamically_scalable, vm_instance.uuid,
> vm_instance.disk_offering_id, vm_instance.power_state,
> vm_instance.power_state_update_time, vm_instance.power_state_update_count,
> vm_instance.power_host FROM vm_instance WHERE vm_instance.instance_name =
> _binary'r-16916-VM' AND vm_instance.removed IS NULL ORDER BY RAND() LIMIT
> 1;# User@Host: cloud[cloud] @ x3 [10.81.28.128] Id: 637881# Query_time:
> 8.193784 Lock_time: 0.000107 Rows_sent: 1 Rows_examined: 19935SET
> timestamp=1431329557;
> SET timestamp=1431329601;SELECT host.id, host.disconnected, host.name,
> host.status, host.type, host.private_ip_address, host.private_mac_address,
> host.private_netmask, host.public_netmask, host.public_ip_address,
> host.public_mac_address, host.storage_ip_address, host.cluster_id,
> host.storage_netmask, host.storage_mac_address, host.storage_ip_address_2,
> host.storage_netmask_2, host.storage_mac_address_2, host.hypervisor_type,
> host.proxy_port, host.resource, host.fs_type, host.available, host.setup,
> host.resource_state, host.hypervisor_version, host.update_count, host.uuid,
> host.data_center_id, host.pod_id, host.cpu_sockets, host.cpus, host.url,
> host.speed, host.ram, host.parent, host.guid, host.capabilities,
> host.total_size, host.last_ping, host.mgmt_server_id, host.dom0_memory,
> host.version, host.created, host.removed FROM host WHERE host.id = 913 AND
> host.removed IS NULL;# User@Host: cloud[cloud] @ x3 [10.81.28.128] Id:
> 637865# Query_time: 5.861241 Lock_time: 0.000139 Rows_sent: 1
> Rows_examined: 1
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
