Two bugs

classic Classic list List threaded Threaded
11 messages Options
Reply | Threaded
Open this post in threaded view
|

Two bugs

Kevin Howe

Hi Apollo developers, 

We experiencing a couple of bugs when deploying Apollo for WormBase. The attached movies should hopefully explain what's going on. 

In the first, after first logging in, turning on a track results in two copies of the track appearing in the main window; turning it off again results in one of the copies disappearing, but trying to turn the track off completely has no effect (the button keeps slipping back to the on position). Shift-reload fixes this for a time, but the problem often later occurs again later in the session. 

In the second, we are trying to promote a long RNASeq read (from a BAM file) into the user annotations track. Nothing happens, and it appears to block other edit operations (promotion of smaller reads and deletes). A shift-reload reenables the ability to do these small edits, but the attempt to the promote the longer read blocks consistently. 

A third problem (not in the movies) is that we regularly encounter sessions where promoting features to the user-annotations track works, but nothing appears until you reload the page. 

We are using the the Apollo 2.x docker container (https://github.com/GMOD/docker-apollo) to create a local service, and then creating a WAR file which is deployed on EMBL-EBI externally-facing servers using Tomcat (EMBL-EBI do not support docker for externally-facing services yet). And as you can see from the videos, we are using the Neat Features JBrowse plug-in. 

We can provide more details if necessary - and even a test account for you to have a play. But to start things off: has anyone seen problems like this before? 

Many thanks,

Kevin Howe
WormBase





This list is for the Apollo Annotation Editing Tool. Info at http://genomearchitect.org/
If you wish to unsubscribe from the Apollo List: 1. From the address with which you subscribed to the list, send a message to [hidden email] | 2. In the subject line of your email type: unsubscribe apollo | 3. Leave the message body blank.

Reply | Threaded
Open this post in threaded view
|

Re: Two bugs

nathandunn

I haven’t seen these effects on master, though there were some effects with the older Track Panel if you had lots of tracks that has since been fixed.  My suspicion is that something is crashing on the client and affecting all of the downstream behaviors.

A couple of quick questions:

1 - Can you provide any console log errors?
2 - Would it be possible for you to try running off of master as an experiment and see if you can repeat the experiment?

gmod/apollo:latest

If you are using another branch from the docker hub (other than 2.0.6), you can replace the WEBAPOLLO_VERSION with  07fed213ea9b122d68efc85c1fe114c7c3948f79 in the Docker file to upgrade the Apollo version, otherwise the “latest” tag will be your best bet.

Nathan

> On Sep 26, 2017, at 3:31 AM, Kevin Howe <[hidden email]> wrote:
>
>
> Hi Apollo developers,
>
> We experiencing a couple of bugs when deploying Apollo for WormBase. The attached movies should hopefully explain what's going on.
>
> In the first, after first logging in, turning on a track results in two copies of the track appearing in the main window; turning it off again results in one of the copies disappearing, but trying to turn the track off completely has no effect (the button keeps slipping back to the on position). Shift-reload fixes this for a time, but the problem often later occurs again later in the session.
>
> In the second, we are trying to promote a long RNASeq read (from a BAM file) into the user annotations track. Nothing happens, and it appears to block other edit operations (promotion of smaller reads and deletes). A shift-reload reenables the ability to do these small edits, but the attempt to the promote the longer read blocks consistently.
>
> A third problem (not in the movies) is that we regularly encounter sessions where promoting features to the user-annotations track works, but nothing appears until you reload the page.
>
> We are using the the Apollo 2.x docker container (https://github.com/GMOD/docker-apollo) to create a local service, and then creating a WAR file which is deployed on EMBL-EBI externally-facing servers using Tomcat (EMBL-EBI do not support docker for externally-facing services yet). And as you can see from the videos, we are using the Neat Features JBrowse plug-in.
>
> We can provide more details if necessary - and even a test account for you to have a play. But to start things off: has anyone seen problems like this before?
>
> Many thanks,
>
> Kevin Howe
> WormBase
>
> ​
>  apollo_bugs.cannot_promote_long_BAM_read.mov​​
>  apollo_bugs.double_track_load.mov​
>
>
>
> This list is for the Apollo Annotation Editing Tool. Info at http://genomearchitect.org/
> If you wish to unsubscribe from the Apollo List: 1. From the address with which you subscribed to the list, send a message to [hidden email] | 2. In the subject line of your email type: unsubscribe apollo | 3. Leave the message body blank.
>




This list is for the Apollo Annotation Editing Tool. Info at http://genomearchitect.org/
If you wish to unsubscribe from the Apollo List: 1. From the address with which you subscribed to the list, send a message to [hidden email] | 2. In the subject line of your email type: unsubscribe apollo | 3. Leave the message body blank.

Reply | Threaded
Open this post in threaded view
|

Re: Two bugs

Scott Cain
In reply to this post by Kevin Howe
Hi Kevin,

The thing with two identical tracks going on and then only one going away when turning off I've seen before in JBrowse.  I think it results from a configuration problem. I think it may happen when the "key" value in the track config isn't unique, but I can't remember for sure.

Scott


On Tue, Sep 26, 2017 at 6:31 AM, Kevin Howe <[hidden email]> wrote:

Hi Apollo developers, 

We experiencing a couple of bugs when deploying Apollo for WormBase. The attached movies should hopefully explain what's going on. 

In the first, after first logging in, turning on a track results in two copies of the track appearing in the main window; turning it off again results in one of the copies disappearing, but trying to turn the track off completely has no effect (the button keeps slipping back to the on position). Shift-reload fixes this for a time, but the problem often later occurs again later in the session. 

In the second, we are trying to promote a long RNASeq read (from a BAM file) into the user annotations track. Nothing happens, and it appears to block other edit operations (promotion of smaller reads and deletes). A shift-reload reenables the ability to do these small edits, but the attempt to the promote the longer read blocks consistently. 

A third problem (not in the movies) is that we regularly encounter sessions where promoting features to the user-annotations track works, but nothing appears until you reload the page. 

We are using the the Apollo 2.x docker container (https://github.com/GMOD/docker-apollo) to create a local service, and then creating a WAR file which is deployed on EMBL-EBI externally-facing servers using Tomcat (EMBL-EBI do not support docker for externally-facing services yet). And as you can see from the videos, we are using the Neat Features JBrowse plug-in. 

We can provide more details if necessary - and even a test account for you to have a play. But to start things off: has anyone seen problems like this before? 

Many thanks,

Kevin Howe
WormBase





This list is for the Apollo Annotation Editing Tool. Info at http://genomearchitect.org/
If you wish to unsubscribe from the Apollo List: 1. From the address with which you subscribed to the list, send a message to [hidden email] | 2. In the subject line of your email type: unsubscribe apollo | 3. Leave the message body blank.





--
------------------------------------------------------------------------
Scott Cain, Ph. D.                                   scott at scottcain dot net
GMOD Coordinator (http://gmod.org/)                     216-392-3087
Ontario Institute for Cancer Research




This list is for the Apollo Annotation Editing Tool. Info at http://genomearchitect.org/
If you wish to unsubscribe from the Apollo List: 1. From the address with which you subscribed to the list, send a message to [hidden email] | 2. In the subject line of your email type: unsubscribe apollo | 3. Leave the message body blank.

Reply | Threaded
Open this post in threaded view
|

Re: Two bugs

Kevin Howe
In reply to this post by nathandunn

Hi Nathan,

We are able to reproduce the failure-to-promote-from-BAM problem on master (from today). Upon inspection, we are seeing the below in the server logs. 

Thanks,

Kevin


09-Oct-2017 15:57:05.359 SEVERE [https-openssl-nio-8443-exec-8] org.apache.coyote.AbstractProtocol$ConnectionHandler.process Error reading request, ignored
 java.lang.IllegalStateException: Unexpected read state [CLOSING]
    at org.apache.tomcat.websocket.server.WsFrameServer.notifyDataAvailable(WsFrameServer.java:163)
    at org.apache.tomcat.websocket.server.WsHttpUpgradeHandler.upgradeDispatch(WsHttpUpgradeHandler.java:148)
    at org.apache.coyote.http11.upgrade.UpgradeProcessorInternal.dispatch(UpgradeProcessorInternal.java:54)
    at org.apache.coyote.AbstractProcessorLight.process(AbstractProcessorLight.java:53)
    at org.apache.coyote.AbstractProtocol$ConnectionHandler.process(AbstractProtocol.java:868)
    at org.apache.tomcat.util.net.NioEndpoint$SocketProcessor.doRun(NioEndpoint.java:1459)
    at org.apache.tomcat.util.net.SocketProcessorBase.run(SocketProcessorBase.java:49)
    at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1149)
    at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:624)
    at org.apache.tomcat.util.threads.TaskThread$WrappingRunnable.run(TaskThread.java:61)
    at java.lang.Thread.run(Thread.java:748)



On 26 September 2017 at 17:02, Nathan Dunn <[hidden email]> wrote:

I haven’t seen these effects on master, though there were some effects with the older Track Panel if you had lots of tracks that has since been fixed.  My suspicion is that something is crashing on the client and affecting all of the downstream behaviors.

A couple of quick questions:

1 - Can you provide any console log errors?
2 - Would it be possible for you to try running off of master as an experiment and see if you can repeat the experiment?

gmod/apollo:latest

If you are using another branch from the docker hub (other than 2.0.6), you can replace the WEBAPOLLO_VERSION with  07fed213ea9b122d68efc85c1fe114c7c3948f79 in the Docker file to upgrade the Apollo version, otherwise the “latest” tag will be your best bet.

Nathan

> On Sep 26, 2017, at 3:31 AM, Kevin Howe <[hidden email]> wrote:
>
>
> Hi Apollo developers,
>
> We experiencing a couple of bugs when deploying Apollo for WormBase. The attached movies should hopefully explain what's going on.
>
> In the first, after first logging in, turning on a track results in two copies of the track appearing in the main window; turning it off again results in one of the copies disappearing, but trying to turn the track off completely has no effect (the button keeps slipping back to the on position). Shift-reload fixes this for a time, but the problem often later occurs again later in the session.
>
> In the second, we are trying to promote a long RNASeq read (from a BAM file) into the user annotations track. Nothing happens, and it appears to block other edit operations (promotion of smaller reads and deletes). A shift-reload reenables the ability to do these small edits, but the attempt to the promote the longer read blocks consistently.
>
> A third problem (not in the movies) is that we regularly encounter sessions where promoting features to the user-annotations track works, but nothing appears until you reload the page.
>
> We are using the the Apollo 2.x docker container (https://github.com/GMOD/docker-apollo) to create a local service, and then creating a WAR file which is deployed on EMBL-EBI externally-facing servers using Tomcat (EMBL-EBI do not support docker for externally-facing services yet). And as you can see from the videos, we are using the Neat Features JBrowse plug-in.
>
> We can provide more details if necessary - and even a test account for you to have a play. But to start things off: has anyone seen problems like this before?
>
> Many thanks,
>
> Kevin Howe
> WormBase
>
> ​
>  apollo_bugs.cannot_promote_long_BAM_read.mov​​
>  apollo_bugs.double_track_load.mov​
>
>
>
> This list is for the Apollo Annotation Editing Tool. Info at http://genomearchitect.org/
> If you wish to unsubscribe from the Apollo List: 1. From the address with which you subscribed to the list, send a message to [hidden email] | 2. In the subject line of your email type: unsubscribe apollo | 3. Leave the message body blank.
>





This list is for the Apollo Annotation Editing Tool. Info at http://genomearchitect.org/
If you wish to unsubscribe from the Apollo List: 1. From the address with which you subscribed to the list, send a message to [hidden email] | 2. In the subject line of your email type: unsubscribe apollo | 3. Leave the message body blank.







This list is for the Apollo Annotation Editing Tool. Info at http://genomearchitect.org/
If you wish to unsubscribe from the Apollo List: 1. From the address with which you subscribed to the list, send a message to [hidden email] | 2. In the subject line of your email type: unsubscribe apollo | 3. Leave the message body blank.

Reply | Threaded
Open this post in threaded view
|

Re: Two bugs

nathandunn

It looks like there is a problem with web sockets, but I am unsure if this is your error.  It wouldn’t be unusual to get that on a refresh, for example when logging in so this is probably a red herring.


1 - Can you post the javascript console when loading the page and when adding the long BAM track?   Feel free to send these to me directly through slack as well if anything is private.  The network console might be good as well. 
2 - A full tomcat log when you add a BAM would be helpful, but it looks like you aren’t seeing much, which means its probably another error.
3 - Can you add shorter BAMs from that same track?
4 - Can you add BAMs from other tracks of varying lengths?

Thanks,

Nathan


On Oct 9, 2017, at 9:56 AM, Kevin Howe <[hidden email]> wrote:


Hi Nathan,

We are able to reproduce the failure-to-promote-from-BAM problem on master (from today). Upon inspection, we are seeing the below in the server logs. 

Thanks,

Kevin


09-Oct-2017 15:57:05.359 SEVERE [https-openssl-nio-8443-exec-8] org.apache.coyote.AbstractProtocol$ConnectionHandler.process Error reading request, ignored
 java.lang.IllegalStateException: Unexpected read state [CLOSING]
    at org.apache.tomcat.websocket.server.WsFrameServer.notifyDataAvailable(WsFrameServer.java:163)
    at org.apache.tomcat.websocket.server.WsHttpUpgradeHandler.upgradeDispatch(WsHttpUpgradeHandler.java:148)
    at org.apache.coyote.http11.upgrade.UpgradeProcessorInternal.dispatch(UpgradeProcessorInternal.java:54)
    at org.apache.coyote.AbstractProcessorLight.process(AbstractProcessorLight.java:53)
    at org.apache.coyote.AbstractProtocol$ConnectionHandler.process(AbstractProtocol.java:868)
    at org.apache.tomcat.util.net.NioEndpoint$SocketProcessor.doRun(NioEndpoint.java:1459)
    at org.apache.tomcat.util.net.SocketProcessorBase.run(SocketProcessorBase.java:49)
    at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1149)
    at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:624)
    at org.apache.tomcat.util.threads.TaskThread$WrappingRunnable.run(TaskThread.java:61)
    at java.lang.Thread.run(Thread.java:748)



On 26 September 2017 at 17:02, Nathan Dunn <[hidden email]> wrote:

I haven’t seen these effects on master, though there were some effects with the older Track Panel if you had lots of tracks that has since been fixed.  My suspicion is that something is crashing on the client and affecting all of the downstream behaviors.

A couple of quick questions:

1 - Can you provide any console log errors?
2 - Would it be possible for you to try running off of master as an experiment and see if you can repeat the experiment?

gmod/apollo:latest

If you are using another branch from the docker hub (other than 2.0.6), you can replace the WEBAPOLLO_VERSION with  07fed213ea9b122d68efc85c1fe114c7c3948f79 in the Docker file to upgrade the Apollo version, otherwise the “latest” tag will be your best bet.

Nathan

> On Sep 26, 2017, at 3:31 AM, Kevin Howe <[hidden email]> wrote:
>
>
> Hi Apollo developers,
>
> We experiencing a couple of bugs when deploying Apollo for WormBase. The attached movies should hopefully explain what's going on.
>
> In the first, after first logging in, turning on a track results in two copies of the track appearing in the main window; turning it off again results in one of the copies disappearing, but trying to turn the track off completely has no effect (the button keeps slipping back to the on position). Shift-reload fixes this for a time, but the problem often later occurs again later in the session.
>
> In the second, we are trying to promote a long RNASeq read (from a BAM file) into the user annotations track. Nothing happens, and it appears to block other edit operations (promotion of smaller reads and deletes). A shift-reload reenables the ability to do these small edits, but the attempt to the promote the longer read blocks consistently.
>
> A third problem (not in the movies) is that we regularly encounter sessions where promoting features to the user-annotations track works, but nothing appears until you reload the page.
>
> We are using the the Apollo 2.x docker container (https://github.com/GMOD/docker-apollo) to create a local service, and then creating a WAR file which is deployed on EMBL-EBI externally-facing servers using Tomcat (EMBL-EBI do not support docker for externally-facing services yet). And as you can see from the videos, we are using the Neat Features JBrowse plug-in.
>
> We can provide more details if necessary - and even a test account for you to have a play. But to start things off: has anyone seen problems like this before?
>
> Many thanks,
>
> Kevin Howe
> WormBase
>
> ​
>  apollo_bugs.cannot_promote_long_BAM_read.mov​​
>  apollo_bugs.double_track_load.mov​
>
>
>
> This list is for the Apollo Annotation Editing Tool. Info at http://genomearchitect.org/
> If you wish to unsubscribe from the Apollo List: 1. From the address with which you subscribed to the list, send a message to [hidden email] | 2. In the subject line of your email type: unsubscribe apollo | 3. Leave the message body blank.
>





This list is for the Apollo Annotation Editing Tool. Info at http://genomearchitect.org/
If you wish to unsubscribe from the Apollo List: 1. From the address with which you subscribed to the list, send a message to [hidden email] | 2. In the subject line of your email type: unsubscribe apollo | 3. Leave the message body blank.






This list is for the Apollo Annotation Editing Tool. Info at http://genomearchitect.org/
If you wish to unsubscribe from the Apollo List: 1. From the address with which you subscribed to the list, send a message to [hidden email] | 2. In the subject line of your email type: unsubscribe apollo | 3. Leave the message body blank.






This list is for the Apollo Annotation Editing Tool. Info at http://genomearchitect.org/
If you wish to unsubscribe from the Apollo List: 1. From the address with which you subscribed to the list, send a message to [hidden email] | 2. In the subject line of your email type: unsubscribe apollo | 3. Leave the message body blank.

Reply | Threaded
Open this post in threaded view
|

Re: Two bugs

Kevin Howe

Hi Nathan,

The exception is reliably reproduced every time we try to promote this one read from the BAM file. For some reason, the action closes the web socket connection, and all subsequent attempts to push anything into the user-annotations track fail (until you reload the page). 

Promoting shorter reads first generally works. We've had various problems promoting BAM reads since setting Apollo, but many of them are not reliably reproducible. This particular one is. It may not explain all of our problems, but it's a start!

I'll set up an account for you and send you the details off-line. That way you examine the javascript and network consoles yourself (rather than me guessing at what you might be looking for). 

Thanks,

Kevin


 

On 10 October 2017 at 00:55, Nathan Dunn <[hidden email]> wrote:

It looks like there is a problem with web sockets, but I am unsure if this is your error.  It wouldn’t be unusual to get that on a refresh, for example when logging in so this is probably a red herring.


1 - Can you post the javascript console when loading the page and when adding the long BAM track?   Feel free to send these to me directly through slack as well if anything is private.  The network console might be good as well. 
2 - A full tomcat log when you add a BAM would be helpful, but it looks like you aren’t seeing much, which means its probably another error.
3 - Can you add shorter BAMs from that same track?
4 - Can you add BAMs from other tracks of varying lengths?

Thanks,

Nathan


On Oct 9, 2017, at 9:56 AM, Kevin Howe <[hidden email]> wrote:


Hi Nathan,

We are able to reproduce the failure-to-promote-from-BAM problem on master (from today). Upon inspection, we are seeing the below in the server logs. 

Thanks,

Kevin


09-Oct-2017 15:57:05.359 SEVERE [https-openssl-nio-8443-exec-8] org.apache.coyote.AbstractProtocol$ConnectionHandler.process Error reading request, ignored
 java.lang.IllegalStateException: Unexpected read state [CLOSING]
    at org.apache.tomcat.websocket.server.WsFrameServer.notifyDataAvailable(WsFrameServer.java:163)
    at org.apache.tomcat.websocket.server.WsHttpUpgradeHandler.upgradeDispatch(WsHttpUpgradeHandler.java:148)
    at org.apache.coyote.http11.upgrade.UpgradeProcessorInternal.dispatch(UpgradeProcessorInternal.java:54)
    at org.apache.coyote.AbstractProcessorLight.process(AbstractProcessorLight.java:53)
    at org.apache.coyote.AbstractProtocol$ConnectionHandler.process(AbstractProtocol.java:868)
    at org.apache.tomcat.util.net.NioEndpoint$SocketProcessor.doRun(NioEndpoint.java:1459)
    at org.apache.tomcat.util.net.SocketProcessorBase.run(SocketProcessorBase.java:49)
    at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1149)
    at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:624)
    at org.apache.tomcat.util.threads.TaskThread$WrappingRunnable.run(TaskThread.java:61)
    at java.lang.Thread.run(Thread.java:748)



On 26 September 2017 at 17:02, Nathan Dunn <[hidden email]> wrote:

I haven’t seen these effects on master, though there were some effects with the older Track Panel if you had lots of tracks that has since been fixed.  My suspicion is that something is crashing on the client and affecting all of the downstream behaviors.

A couple of quick questions:

1 - Can you provide any console log errors?
2 - Would it be possible for you to try running off of master as an experiment and see if you can repeat the experiment?

gmod/apollo:latest

If you are using another branch from the docker hub (other than 2.0.6), you can replace the WEBAPOLLO_VERSION with  07fed213ea9b122d68efc85c1fe114c7c3948f79 in the Docker file to upgrade the Apollo version, otherwise the “latest” tag will be your best bet.

Nathan

> On Sep 26, 2017, at 3:31 AM, Kevin Howe <[hidden email]> wrote:
>
>
> Hi Apollo developers,
>
> We experiencing a couple of bugs when deploying Apollo for WormBase. The attached movies should hopefully explain what's going on.
>
> In the first, after first logging in, turning on a track results in two copies of the track appearing in the main window; turning it off again results in one of the copies disappearing, but trying to turn the track off completely has no effect (the button keeps slipping back to the on position). Shift-reload fixes this for a time, but the problem often later occurs again later in the session.
>
> In the second, we are trying to promote a long RNASeq read (from a BAM file) into the user annotations track. Nothing happens, and it appears to block other edit operations (promotion of smaller reads and deletes). A shift-reload reenables the ability to do these small edits, but the attempt to the promote the longer read blocks consistently.
>
> A third problem (not in the movies) is that we regularly encounter sessions where promoting features to the user-annotations track works, but nothing appears until you reload the page.
>
> We are using the the Apollo 2.x docker container (https://github.com/GMOD/docker-apollo) to create a local service, and then creating a WAR file which is deployed on EMBL-EBI externally-facing servers using Tomcat (EMBL-EBI do not support docker for externally-facing services yet). And as you can see from the videos, we are using the Neat Features JBrowse plug-in.
>
> We can provide more details if necessary - and even a test account for you to have a play. But to start things off: has anyone seen problems like this before?
>
> Many thanks,
>
> Kevin Howe
> WormBase
>
> ​
>  apollo_bugs.cannot_promote_long_BAM_read.mov​​
>  apollo_bugs.double_track_load.mov​
>
>
>
> This list is for the Apollo Annotation Editing Tool. Info at http://genomearchitect.org/
> If you wish to unsubscribe from the Apollo List: 1. From the address with which you subscribed to the list, send a message to [hidden email] | 2. In the subject line of your email type: unsubscribe apollo | 3. Leave the message body blank.
>





This list is for the Apollo Annotation Editing Tool. Info at http://genomearchitect.org/
If you wish to unsubscribe from the Apollo List: 1. From the address with which you subscribed to the list, send a message to [hidden email] | 2. In the subject line of your email type: unsubscribe apollo | 3. Leave the message body blank.






This list is for the Apollo Annotation Editing Tool. Info at http://genomearchitect.org/
If you wish to unsubscribe from the Apollo List: 1. From the address with which you subscribed to the list, send a message to [hidden email] | 2. In the subject line of your email type: unsubscribe apollo | 3. Leave the message body blank.






This list is for the Apollo Annotation Editing Tool. Info at http://genomearchitect.org/
If you wish to unsubscribe from the Apollo List: 1. From the address with which you subscribed to the list, send a message to [hidden email] | 2. In the subject line of your email type: unsubscribe apollo | 3. Leave the message body blank.







This list is for the Apollo Annotation Editing Tool. Info at http://genomearchitect.org/
If you wish to unsubscribe from the Apollo List: 1. From the address with which you subscribed to the list, send a message to [hidden email] | 2. In the subject line of your email type: unsubscribe apollo | 3. Leave the message body blank.

Reply | Threaded
Open this post in threaded view
|

Re: Two bugs

nathandunn
Excellent thanks.

If you could provide server details as well and how you setup the proxy that might also be useful. 

Nathan

On Oct 10, 2017, at 8:09 AM, Kevin Howe <[hidden email]> wrote:


Hi Nathan,

The exception is reliably reproduced every time we try to promote this one read from the BAM file. For some reason, the action closes the web socket connection, and all subsequent attempts to push anything into the user-annotations track fail (until you reload the page). 

Promoting shorter reads first generally works. We've had various problems promoting BAM reads since setting Apollo, but many of them are not reliably reproducible. This particular one is. It may not explain all of our problems, but it's a start!

I'll set up an account for you and send you the details off-line. That way you examine the javascript and network consoles yourself (rather than me guessing at what you might be looking for). 

Thanks,

Kevin


 

On 10 October 2017 at 00:55, Nathan Dunn <[hidden email]> wrote:

It looks like there is a problem with web sockets, but I am unsure if this is your error.  It wouldn’t be unusual to get that on a refresh, for example when logging in so this is probably a red herring.


1 - Can you post the javascript console when loading the page and when adding the long BAM track?   Feel free to send these to me directly through slack as well if anything is private.  The network console might be good as well. 
2 - A full tomcat log when you add a BAM would be helpful, but it looks like you aren’t seeing much, which means its probably another error.
3 - Can you add shorter BAMs from that same track?
4 - Can you add BAMs from other tracks of varying lengths?

Thanks,

Nathan


On Oct 9, 2017, at 9:56 AM, Kevin Howe <[hidden email]> wrote:


Hi Nathan,

We are able to reproduce the failure-to-promote-from-BAM problem on master (from today). Upon inspection, we are seeing the below in the server logs. 

Thanks,

Kevin


09-Oct-2017 15:57:05.359 SEVERE [https-openssl-nio-8443-exec-8] org.apache.coyote.AbstractProtocol$ConnectionHandler.process Error reading request, ignored
 java.lang.IllegalStateException: Unexpected read state [CLOSING]
    at org.apache.tomcat.websocket.server.WsFrameServer.notifyDataAvailable(WsFrameServer.java:163)
    at org.apache.tomcat.websocket.server.WsHttpUpgradeHandler.upgradeDispatch(WsHttpUpgradeHandler.java:148)
    at org.apache.coyote.http11.upgrade.UpgradeProcessorInternal.dispatch(UpgradeProcessorInternal.java:54)
    at org.apache.coyote.AbstractProcessorLight.process(AbstractProcessorLight.java:53)
    at org.apache.coyote.AbstractProtocol$ConnectionHandler.process(AbstractProtocol.java:868)
    at org.apache.tomcat.util.net.NioEndpoint$SocketProcessor.doRun(NioEndpoint.java:1459)
    at org.apache.tomcat.util.net.SocketProcessorBase.run(SocketProcessorBase.java:49)
    at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1149)
    at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:624)
    at org.apache.tomcat.util.threads.TaskThread$WrappingRunnable.run(TaskThread.java:61)
    at java.lang.Thread.run(Thread.java:748)



On 26 September 2017 at 17:02, Nathan Dunn <[hidden email]> wrote:

I haven’t seen these effects on master, though there were some effects with the older Track Panel if you had lots of tracks that has since been fixed.  My suspicion is that something is crashing on the client and affecting all of the downstream behaviors.

A couple of quick questions:

1 - Can you provide any console log errors?
2 - Would it be possible for you to try running off of master as an experiment and see if you can repeat the experiment?

gmod/apollo:latest

If you are using another branch from the docker hub (other than 2.0.6), you can replace the WEBAPOLLO_VERSION with  07fed213ea9b122d68efc85c1fe114c7c3948f79 in the Docker file to upgrade the Apollo version, otherwise the “latest” tag will be your best bet.

Nathan

> On Sep 26, 2017, at 3:31 AM, Kevin Howe <[hidden email]> wrote:
>
>
> Hi Apollo developers,
>
> We experiencing a couple of bugs when deploying Apollo for WormBase. The attached movies should hopefully explain what's going on.
>
> In the first, after first logging in, turning on a track results in two copies of the track appearing in the main window; turning it off again results in one of the copies disappearing, but trying to turn the track off completely has no effect (the button keeps slipping back to the on position). Shift-reload fixes this for a time, but the problem often later occurs again later in the session.
>
> In the second, we are trying to promote a long RNASeq read (from a BAM file) into the user annotations track. Nothing happens, and it appears to block other edit operations (promotion of smaller reads and deletes). A shift-reload reenables the ability to do these small edits, but the attempt to the promote the longer read blocks consistently.
>
> A third problem (not in the movies) is that we regularly encounter sessions where promoting features to the user-annotations track works, but nothing appears until you reload the page.
>
> We are using the the Apollo 2.x docker container (https://github.com/GMOD/docker-apollo) to create a local service, and then creating a WAR file which is deployed on EMBL-EBI externally-facing servers using Tomcat (EMBL-EBI do not support docker for externally-facing services yet). And as you can see from the videos, we are using the Neat Features JBrowse plug-in.
>
> We can provide more details if necessary - and even a test account for you to have a play. But to start things off: has anyone seen problems like this before?
>
> Many thanks,
>
> Kevin Howe
> WormBase
>
> ​
>  apollo_bugs.cannot_promote_long_BAM_read.mov​​
>  apollo_bugs.double_track_load.mov​
>
>
>
> This list is for the Apollo Annotation Editing Tool. Info at http://genomearchitect.org/
> If you wish to unsubscribe from the Apollo List: 1. From the address with which you subscribed to the list, send a message to [hidden email] | 2. In the subject line of your email type: unsubscribe apollo | 3. Leave the message body blank.
>





This list is for the Apollo Annotation Editing Tool. Info at http://genomearchitect.org/
If you wish to unsubscribe from the Apollo List: 1. From the address with which you subscribed to the list, send a message to [hidden email] | 2. In the subject line of your email type: unsubscribe apollo | 3. Leave the message body blank.






This list is for the Apollo Annotation Editing Tool. Info at http://genomearchitect.org/
If you wish to unsubscribe from the Apollo List: 1. From the address with which you subscribed to the list, send a message to [hidden email] | 2. In the subject line of your email type: unsubscribe apollo | 3. Leave the message body blank.






This list is for the Apollo Annotation Editing Tool. Info at http://genomearchitect.org/
If you wish to unsubscribe from the Apollo List: 1. From the address with which you subscribed to the list, send a message to [hidden email] | 2. In the subject line of your email type: unsubscribe apollo | 3. Leave the message body blank.






This list is for the Apollo Annotation Editing Tool. Info at http://genomearchitect.org/
If you wish to unsubscribe from the Apollo List: 1. From the address with which you subscribed to the list, send a message to [hidden email] | 2. In the subject line of your email type: unsubscribe apollo | 3. Leave the message body blank.





This list is for the Apollo Annotation Editing Tool. Info at http://genomearchitect.org/
If you wish to unsubscribe from the Apollo List: 1. From the address with which you subscribed to the list, send a message to [hidden email] | 2. In the subject line of your email type: unsubscribe apollo | 3. Leave the message body blank.

Reply | Threaded
Open this post in threaded view
|

Re: Two bugs

nathandunn
Okay.  I opened an issue here. 


I was actually able to add the larger BAM reads on the second or third try.   Looks like the socket is being closed on the server-side and not re-opened again for a particular client.  No errors on the client or the server except as noted.   

Some possibilities / things to try (will follow-up in the GH issue):

1 - using a newer docker-apollo version, which points to a newer Apollo version.  We made some further improvements

I’m downloading the data you sent me and will retest locally. 

Nathan


On Oct 10, 2017, at 8:17 AM, Lbl <[hidden email]> wrote:

Excellent thanks.

If you could provide server details as well and how you setup the proxy that might also be useful. 

Nathan

On Oct 10, 2017, at 8:09 AM, Kevin Howe <[hidden email]> wrote:


Hi Nathan,

The exception is reliably reproduced every time we try to promote this one read from the BAM file. For some reason, the action closes the web socket connection, and all subsequent attempts to push anything into the user-annotations track fail (until you reload the page). 

Promoting shorter reads first generally works. We've had various problems promoting BAM reads since setting Apollo, but many of them are not reliably reproducible. This particular one is. It may not explain all of our problems, but it's a start!

I'll set up an account for you and send you the details off-line. That way you examine the javascript and network consoles yourself (rather than me guessing at what you might be looking for). 

Thanks,

Kevin


 

On 10 October 2017 at 00:55, Nathan Dunn <[hidden email]> wrote:

It looks like there is a problem with web sockets, but I am unsure if this is your error.  It wouldn’t be unusual to get that on a refresh, for example when logging in so this is probably a red herring.


1 - Can you post the javascript console when loading the page and when adding the long BAM track?   Feel free to send these to me directly through slack as well if anything is private.  The network console might be good as well. 
2 - A full tomcat log when you add a BAM would be helpful, but it looks like you aren’t seeing much, which means its probably another error.
3 - Can you add shorter BAMs from that same track?
4 - Can you add BAMs from other tracks of varying lengths?

Thanks,

Nathan


On Oct 9, 2017, at 9:56 AM, Kevin Howe <[hidden email]> wrote:


Hi Nathan,

We are able to reproduce the failure-to-promote-from-BAM problem on master (from today). Upon inspection, we are seeing the below in the server logs. 

Thanks,

Kevin


09-Oct-2017 15:57:05.359 SEVERE [https-openssl-nio-8443-exec-8] org.apache.coyote.AbstractProtocol$ConnectionHandler.process Error reading request, ignored
 java.lang.IllegalStateException: Unexpected read state [CLOSING]
    at org.apache.tomcat.websocket.server.WsFrameServer.notifyDataAvailable(WsFrameServer.java:163)
    at org.apache.tomcat.websocket.server.WsHttpUpgradeHandler.upgradeDispatch(WsHttpUpgradeHandler.java:148)
    at org.apache.coyote.http11.upgrade.UpgradeProcessorInternal.dispatch(UpgradeProcessorInternal.java:54)
    at org.apache.coyote.AbstractProcessorLight.process(AbstractProcessorLight.java:53)
    at org.apache.coyote.AbstractProtocol$ConnectionHandler.process(AbstractProtocol.java:868)
    at org.apache.tomcat.util.net.NioEndpoint$SocketProcessor.doRun(NioEndpoint.java:1459)
    at org.apache.tomcat.util.net.SocketProcessorBase.run(SocketProcessorBase.java:49)
    at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1149)
    at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:624)
    at org.apache.tomcat.util.threads.TaskThread$WrappingRunnable.run(TaskThread.java:61)
    at java.lang.Thread.run(Thread.java:748)



On 26 September 2017 at 17:02, Nathan Dunn <[hidden email]> wrote:

I haven’t seen these effects on master, though there were some effects with the older Track Panel if you had lots of tracks that has since been fixed.  My suspicion is that something is crashing on the client and affecting all of the downstream behaviors.

A couple of quick questions:

1 - Can you provide any console log errors?
2 - Would it be possible for you to try running off of master as an experiment and see if you can repeat the experiment?

gmod/apollo:latest

If you are using another branch from the docker hub (other than 2.0.6), you can replace the WEBAPOLLO_VERSION with  07fed213ea9b122d68efc85c1fe114c7c3948f79 in the Docker file to upgrade the Apollo version, otherwise the “latest” tag will be your best bet.

Nathan

> On Sep 26, 2017, at 3:31 AM, Kevin Howe <[hidden email]> wrote:
>
>
> Hi Apollo developers,
>
> We experiencing a couple of bugs when deploying Apollo for WormBase. The attached movies should hopefully explain what's going on.
>
> In the first, after first logging in, turning on a track results in two copies of the track appearing in the main window; turning it off again results in one of the copies disappearing, but trying to turn the track off completely has no effect (the button keeps slipping back to the on position). Shift-reload fixes this for a time, but the problem often later occurs again later in the session.
>
> In the second, we are trying to promote a long RNASeq read (from a BAM file) into the user annotations track. Nothing happens, and it appears to block other edit operations (promotion of smaller reads and deletes). A shift-reload reenables the ability to do these small edits, but the attempt to the promote the longer read blocks consistently.
>
> A third problem (not in the movies) is that we regularly encounter sessions where promoting features to the user-annotations track works, but nothing appears until you reload the page.
>
> We are using the the Apollo 2.x docker container (https://github.com/GMOD/docker-apollo) to create a local service, and then creating a WAR file which is deployed on EMBL-EBI externally-facing servers using Tomcat (EMBL-EBI do not support docker for externally-facing services yet). And as you can see from the videos, we are using the Neat Features JBrowse plug-in.
>
> We can provide more details if necessary - and even a test account for you to have a play. But to start things off: has anyone seen problems like this before?
>
> Many thanks,
>
> Kevin Howe
> WormBase
>
> ​
>  apollo_bugs.cannot_promote_long_BAM_read.mov​​
>  apollo_bugs.double_track_load.mov​
>
>
>
> This list is for the Apollo Annotation Editing Tool. Info at http://genomearchitect.org/
> If you wish to unsubscribe from the Apollo List: 1. From the address with which you subscribed to the list, send a message to [hidden email] | 2. In the subject line of your email type: unsubscribe apollo | 3. Leave the message body blank.
>





This list is for the Apollo Annotation Editing Tool. Info at http://genomearchitect.org/
If you wish to unsubscribe from the Apollo List: 1. From the address with which you subscribed to the list, send a message to [hidden email] | 2. In the subject line of your email type: unsubscribe apollo | 3. Leave the message body blank.






This list is for the Apollo Annotation Editing Tool. Info at http://genomearchitect.org/
If you wish to unsubscribe from the Apollo List: 1. From the address with which you subscribed to the list, send a message to [hidden email] | 2. In the subject line of your email type: unsubscribe apollo | 3. Leave the message body blank.






This list is for the Apollo Annotation Editing Tool. Info at http://genomearchitect.org/
If you wish to unsubscribe from the Apollo List: 1. From the address with which you subscribed to the list, send a message to [hidden email] | 2. In the subject line of your email type: unsubscribe apollo | 3. Leave the message body blank.






This list is for the Apollo Annotation Editing Tool. Info at http://genomearchitect.org/
If you wish to unsubscribe from the Apollo List: 1. From the address with which you subscribed to the list, send a message to [hidden email] | 2. In the subject line of your email type: unsubscribe apollo | 3. Leave the message body blank.






This list is for the Apollo Annotation Editing Tool. Info at http://genomearchitect.org/
If you wish to unsubscribe from the Apollo List: 1. From the address with which you subscribed to the list, send a message to [hidden email] | 2. In the subject line of your email type: unsubscribe apollo | 3. Leave the message body blank.

Reply | Threaded
Open this post in threaded view
|

Re: Two bugs

Michael Paulini
Hi Nathan,

I just tried the master branch, and as far as the error is concerned, it behaves the same.

M

On 10/10/17 20:58, Nathan Dunn wrote:
Okay.  I opened an issue here. 


I was actually able to add the larger BAM reads on the second or third try.   Looks like the socket is being closed on the server-side and not re-opened again for a particular client.  No errors on the client or the server except as noted.   

Some possibilities / things to try (will follow-up in the GH issue):

1 - using a newer docker-apollo version, which points to a newer Apollo version.  We made some further improvements

I’m downloading the data you sent me and will retest locally. 

Nathan


On Oct 10, 2017, at 8:17 AM, Lbl <[hidden email]> wrote:

Excellent thanks.

If you could provide server details as well and how you setup the proxy that might also be useful. 

Nathan

On Oct 10, 2017, at 8:09 AM, Kevin Howe <[hidden email]> wrote:


Hi Nathan,

The exception is reliably reproduced every time we try to promote this one read from the BAM file. For some reason, the action closes the web socket connection, and all subsequent attempts to push anything into the user-annotations track fail (until you reload the page). 

Promoting shorter reads first generally works. We've had various problems promoting BAM reads since setting Apollo, but many of them are not reliably reproducible. This particular one is. It may not explain all of our problems, but it's a start!

I'll set up an account for you and send you the details off-line. That way you examine the javascript and network consoles yourself (rather than me guessing at what you might be looking for). 

Thanks,

Kevin


 

On 10 October 2017 at 00:55, Nathan Dunn <[hidden email]> wrote:

It looks like there is a problem with web sockets, but I am unsure if this is your error.  It wouldn’t be unusual to get that on a refresh, for example when logging in so this is probably a red herring.


1 - Can you post the javascript console when loading the page and when adding the long BAM track?   Feel free to send these to me directly through slack as well if anything is private.  The network console might be good as well. 
2 - A full tomcat log when you add a BAM would be helpful, but it looks like you aren’t seeing much, which means its probably another error.
3 - Can you add shorter BAMs from that same track?
4 - Can you add BAMs from other tracks of varying lengths?

Thanks,

Nathan


On Oct 9, 2017, at 9:56 AM, Kevin Howe <[hidden email]> wrote:


Hi Nathan,

We are able to reproduce the failure-to-promote-from-BAM problem on master (from today). Upon inspection, we are seeing the below in the server logs. 

Thanks,

Kevin


09-Oct-2017 15:57:05.359 SEVERE [https-openssl-nio-8443-exec-8] org.apache.coyote.AbstractProtocol$ConnectionHandler.process Error reading request, ignored
 java.lang.IllegalStateException: Unexpected read state [CLOSING]
    at org.apache.tomcat.websocket.server.WsFrameServer.notifyDataAvailable(WsFrameServer.java:163)
    at org.apache.tomcat.websocket.server.WsHttpUpgradeHandler.upgradeDispatch(WsHttpUpgradeHandler.java:148)
    at org.apache.coyote.http11.upgrade.UpgradeProcessorInternal.dispatch(UpgradeProcessorInternal.java:54)
    at org.apache.coyote.AbstractProcessorLight.process(AbstractProcessorLight.java:53)
    at org.apache.coyote.AbstractProtocol$ConnectionHandler.process(AbstractProtocol.java:868)
    at org.apache.tomcat.util.net.NioEndpoint$SocketProcessor.doRun(NioEndpoint.java:1459)
    at org.apache.tomcat.util.net.SocketProcessorBase.run(SocketProcessorBase.java:49)
    at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1149)
    at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:624)
    at org.apache.tomcat.util.threads.TaskThread$WrappingRunnable.run(TaskThread.java:61)
    at java.lang.Thread.run(Thread.java:748)



On 26 September 2017 at 17:02, Nathan Dunn <[hidden email]> wrote:

I haven’t seen these effects on master, though there were some effects with the older Track Panel if you had lots of tracks that has since been fixed.  My suspicion is that something is crashing on the client and affecting all of the downstream behaviors.

A couple of quick questions:

1 - Can you provide any console log errors?
2 - Would it be possible for you to try running off of master as an experiment and see if you can repeat the experiment?

gmod/apollo:latest

If you are using another branch from the docker hub (other than 2.0.6), you can replace the WEBAPOLLO_VERSION with  07fed213ea9b122d68efc85c1fe114c7c3948f79 in the Docker file to upgrade the Apollo version, otherwise the “latest” tag will be your best bet.

Nathan

> On Sep 26, 2017, at 3:31 AM, Kevin Howe <[hidden email]> wrote:
>
>
> Hi Apollo developers,
>
> We experiencing a couple of bugs when deploying Apollo for WormBase. The attached movies should hopefully explain what's going on.
>
> In the first, after first logging in, turning on a track results in two copies of the track appearing in the main window; turning it off again results in one of the copies disappearing, but trying to turn the track off completely has no effect (the button keeps slipping back to the on position). Shift-reload fixes this for a time, but the problem often later occurs again later in the session.
>
> In the second, we are trying to promote a long RNASeq read (from a BAM file) into the user annotations track. Nothing happens, and it appears to block other edit operations (promotion of smaller reads and deletes). A shift-reload reenables the ability to do these small edits, but the attempt to the promote the longer read blocks consistently.
>
> A third problem (not in the movies) is that we regularly encounter sessions where promoting features to the user-annotations track works, but nothing appears until you reload the page.
>
> We are using the the Apollo 2.x docker container (https://github.com/GMOD/docker-apollo) to create a local service, and then creating a WAR file which is deployed on EMBL-EBI externally-facing servers using Tomcat (EMBL-EBI do not support docker for externally-facing services yet). And as you can see from the videos, we are using the Neat Features JBrowse plug-in.
>
> We can provide more details if necessary - and even a test account for you to have a play. But to start things off: has anyone seen problems like this before?
>
> Many thanks,
>
> Kevin Howe
> WormBase
>
> ​
>  apollo_bugs.cannot_promote_long_BAM_read.mov​​
>  apollo_bugs.double_track_load.mov​
>
>
>
> This list is for the Apollo Annotation Editing Tool. Info at http://genomearchitect.org/
> If you wish to unsubscribe from the Apollo List: 1. From the address with which you subscribed to the list, send a message to [hidden email] | 2. In the subject line of your email type: unsubscribe apollo | 3. Leave the message body blank.
>





This list is for the Apollo Annotation Editing Tool. Info at http://genomearchitect.org/
If you wish to unsubscribe from the Apollo List: 1. From the address with which you subscribed to the list, send a message to [hidden email] | 2. In the subject line of your email type: unsubscribe apollo | 3. Leave the message body blank.






This list is for the Apollo Annotation Editing Tool. Info at http://genomearchitect.org/
If you wish to unsubscribe from the Apollo List: 1. From the address with which you subscribed to the list, send a message to [hidden email] | 2. In the subject line of your email type: unsubscribe apollo | 3. Leave the message body blank.






This list is for the Apollo Annotation Editing Tool. Info at http://genomearchitect.org/
If you wish to unsubscribe from the Apollo List: 1. From the address with which you subscribed to the list, send a message to [hidden email] | 2. In the subject line of your email type: unsubscribe apollo | 3. Leave the message body blank.






This list is for the Apollo Annotation Editing Tool. Info at http://genomearchitect.org/
If you wish to unsubscribe from the Apollo List: 1. From the address with which you subscribed to the list, send a message to [hidden email] | 2. In the subject line of your email type: unsubscribe apollo | 3. Leave the message body blank.





This list is for the Apollo Annotation Editing Tool. Info at http://genomearchitect.org/
If you wish to unsubscribe from the Apollo List: 1. From the address with which you subscribed to the list, send a message to [hidden email] | 2. In the subject line of your email type: unsubscribe apollo | 3. Leave the message body blank. 






This list is for the Apollo Annotation Editing Tool. Info at http://genomearchitect.org/
If you wish to unsubscribe from the Apollo List: 1. From the address with which you subscribed to the list, send a message to [hidden email] | 2. In the subject line of your email type: unsubscribe apollo | 3. Leave the message body blank.

Reply | Threaded
Open this post in threaded view
|

Re: Two bugs

nathandunn

Thanks for testing the update.  It looks like its either the web socket frame size is too small (a change to the tomcat XML in docker-apollo) or something other detail in the BAM read is causing the call to the server to fragment into bad JSON.   We should get this resolved this week.

Nathan


On Oct 11, 2017, at 2:26 AM, Michael Paulini <[hidden email]> wrote:

Hi Nathan,

I just tried the master branch, and as far as the error is concerned, it behaves the same.

M

On 10/10/17 20:58, Nathan Dunn wrote:
Okay.  I opened an issue here. 


I was actually able to add the larger BAM reads on the second or third try.   Looks like the socket is being closed on the server-side and not re-opened again for a particular client.  No errors on the client or the server except as noted.   

Some possibilities / things to try (will follow-up in the GH issue):

1 - using a newer docker-apollo version, which points to a newer Apollo version.  We made some further improvements

I’m downloading the data you sent me and will retest locally. 

Nathan


On Oct 10, 2017, at 8:17 AM, Lbl <[hidden email]> wrote:

Excellent thanks.

If you could provide server details as well and how you setup the proxy that might also be useful. 

Nathan

On Oct 10, 2017, at 8:09 AM, Kevin Howe <[hidden email]> wrote:


Hi Nathan,

The exception is reliably reproduced every time we try to promote this one read from the BAM file. For some reason, the action closes the web socket connection, and all subsequent attempts to push anything into the user-annotations track fail (until you reload the page). 

Promoting shorter reads first generally works. We've had various problems promoting BAM reads since setting Apollo, but many of them are not reliably reproducible. This particular one is. It may not explain all of our problems, but it's a start!

I'll set up an account for you and send you the details off-line. That way you examine the javascript and network consoles yourself (rather than me guessing at what you might be looking for). 

Thanks,

Kevin


 

On 10 October 2017 at 00:55, Nathan Dunn <[hidden email]> wrote:

It looks like there is a problem with web sockets, but I am unsure if this is your error.  It wouldn’t be unusual to get that on a refresh, for example when logging in so this is probably a red herring.


1 - Can you post the javascript console when loading the page and when adding the long BAM track?   Feel free to send these to me directly through slack as well if anything is private.  The network console might be good as well. 
2 - A full tomcat log when you add a BAM would be helpful, but it looks like you aren’t seeing much, which means its probably another error.
3 - Can you add shorter BAMs from that same track?
4 - Can you add BAMs from other tracks of varying lengths?

Thanks,

Nathan


On Oct 9, 2017, at 9:56 AM, Kevin Howe <[hidden email]> wrote:


Hi Nathan,

We are able to reproduce the failure-to-promote-from-BAM problem on master (from today). Upon inspection, we are seeing the below in the server logs. 

Thanks,

Kevin


09-Oct-2017 15:57:05.359 SEVERE [https-openssl-nio-8443-exec-8] org.apache.coyote.AbstractProtocol$ConnectionHandler.process Error reading request, ignored
 java.lang.IllegalStateException: Unexpected read state [CLOSING]
    at org.apache.tomcat.websocket.server.WsFrameServer.notifyDataAvailable(WsFrameServer.java:163)
    at org.apache.tomcat.websocket.server.WsHttpUpgradeHandler.upgradeDispatch(WsHttpUpgradeHandler.java:148)
    at org.apache.coyote.http11.upgrade.UpgradeProcessorInternal.dispatch(UpgradeProcessorInternal.java:54)
    at org.apache.coyote.AbstractProcessorLight.process(AbstractProcessorLight.java:53)
    at org.apache.coyote.AbstractProtocol$ConnectionHandler.process(AbstractProtocol.java:868)
    at org.apache.tomcat.util.net.NioEndpoint$SocketProcessor.doRun(NioEndpoint.java:1459)
    at org.apache.tomcat.util.net.SocketProcessorBase.run(SocketProcessorBase.java:49)
    at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1149)
    at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:624)
    at org.apache.tomcat.util.threads.TaskThread$WrappingRunnable.run(TaskThread.java:61)
    at java.lang.Thread.run(Thread.java:748)



On 26 September 2017 at 17:02, Nathan Dunn <[hidden email]> wrote:

I haven’t seen these effects on master, though there were some effects with the older Track Panel if you had lots of tracks that has since been fixed.  My suspicion is that something is crashing on the client and affecting all of the downstream behaviors.

A couple of quick questions:

1 - Can you provide any console log errors?
2 - Would it be possible for you to try running off of master as an experiment and see if you can repeat the experiment?

gmod/apollo:latest

If you are using another branch from the docker hub (other than 2.0.6), you can replace the WEBAPOLLO_VERSION with  07fed213ea9b122d68efc85c1fe114c7c3948f79 in the Docker file to upgrade the Apollo version, otherwise the “latest” tag will be your best bet.

Nathan

> On Sep 26, 2017, at 3:31 AM, Kevin Howe <[hidden email]> wrote:
>
>
> Hi Apollo developers,
>
> We experiencing a couple of bugs when deploying Apollo for WormBase. The attached movies should hopefully explain what's going on.
>
> In the first, after first logging in, turning on a track results in two copies of the track appearing in the main window; turning it off again results in one of the copies disappearing, but trying to turn the track off completely has no effect (the button keeps slipping back to the on position). Shift-reload fixes this for a time, but the problem often later occurs again later in the session.
>
> In the second, we are trying to promote a long RNASeq read (from a BAM file) into the user annotations track. Nothing happens, and it appears to block other edit operations (promotion of smaller reads and deletes). A shift-reload reenables the ability to do these small edits, but the attempt to the promote the longer read blocks consistently.
>
> A third problem (not in the movies) is that we regularly encounter sessions where promoting features to the user-annotations track works, but nothing appears until you reload the page.
>
> We are using the the Apollo 2.x docker container (https://github.com/GMOD/docker-apollo) to create a local service, and then creating a WAR file which is deployed on EMBL-EBI externally-facing servers using Tomcat (EMBL-EBI do not support docker for externally-facing services yet). And as you can see from the videos, we are using the Neat Features JBrowse plug-in.
>
> We can provide more details if necessary - and even a test account for you to have a play. But to start things off: has anyone seen problems like this before?
>
> Many thanks,
>
> Kevin Howe
> WormBase
>
> ​
>  apollo_bugs.cannot_promote_long_BAM_read.mov​​
>  apollo_bugs.double_track_load.mov​
>
>
>
> This list is for the Apollo Annotation Editing Tool. Info at http://genomearchitect.org/
> If you wish to unsubscribe from the Apollo List: 1. From the address with which you subscribed to the list, send a message to [hidden email] | 2. In the subject line of your email type: unsubscribe apollo | 3. Leave the message body blank.
>





This list is for the Apollo Annotation Editing Tool. Info at http://genomearchitect.org/
If you wish to unsubscribe from the Apollo List: 1. From the address with which you subscribed to the list, send a message to [hidden email] | 2. In the subject line of your email type: unsubscribe apollo | 3. Leave the message body blank.






This list is for the Apollo Annotation Editing Tool. Info at http://genomearchitect.org/
If you wish to unsubscribe from the Apollo List: 1. From the address with which you subscribed to the list, send a message to [hidden email] | 2. In the subject line of your email type: unsubscribe apollo | 3. Leave the message body blank.






This list is for the Apollo Annotation Editing Tool. Info at http://genomearchitect.org/
If you wish to unsubscribe from the Apollo List: 1. From the address with which you subscribed to the list, send a message to [hidden email] | 2. In the subject line of your email type: unsubscribe apollo | 3. Leave the message body blank.






This list is for the Apollo Annotation Editing Tool. Info at http://genomearchitect.org/
If you wish to unsubscribe from the Apollo List: 1. From the address with which you subscribed to the list, send a message to [hidden email] | 2. In the subject line of your email type: unsubscribe apollo | 3. Leave the message body blank.




This list is for the Apollo Annotation Editing Tool. Info at http://genomearchitect.org/
If you wish to unsubscribe from the Apollo List: 1. From the address with which you subscribed to the list, send a message to [hidden email] | 2. In the subject line of your email type: unsubscribe apollo | 3. Leave the message body blank. 





This list is for the Apollo Annotation Editing Tool. Info at http://genomearchitect.org/
If you wish to unsubscribe from the Apollo List: 1. From the address with which you subscribed to the list, send a message to [hidden email] | 2. In the subject line of your email type: unsubscribe apollo | 3. Leave the message body blank.






This list is for the Apollo Annotation Editing Tool. Info at http://genomearchitect.org/
If you wish to unsubscribe from the Apollo List: 1. From the address with which you subscribed to the list, send a message to [hidden email] | 2. In the subject line of your email type: unsubscribe apollo | 3. Leave the message body blank.

Reply | Threaded
Open this post in threaded view
|

Re: Two bugs

nathandunn
In reply to this post by nathandunn

Kevin,

We think we have a solution for this.  The problem, as you guessed, was that some reads have a ton of exons in them and the frame is split by the default tomcat buffer.  



To handle the additional density of sub features, we have to increased the websocket size and potentially increase memory to correspond for the additional data.

The most recent version of Apollo (which is building on docker) automatically sets the websocket size to a sufficient amount as seen in Deepak’s PR: https://github.com/GMOD/Apollo/pull/1785.

There is additional documentation in the sample files that indicate how to up the corresponding memory (handling the additional exons and determining overlap requires more memory): https://genomearchitect.readthedocs.io/en/latest/Setup.html#memory-configuration  if you run into memory problems (OutOfMemory / Exceed HeapSize errors in tomcat logs):


I’m not sure what your installation size is, but for your example data I think we needed 4 GB max memory.  If its a larger project, maybe 8GB might work better.

Nathan


On Oct 9, 2017, at 4:55 PM, Nathan Dunn <[hidden email]> wrote:


It looks like there is a problem with web sockets, but I am unsure if this is your error.  It wouldn’t be unusual to get that on a refresh, for example when logging in so this is probably a red herring.


1 - Can you post the javascript console when loading the page and when adding the long BAM track?   Feel free to send these to me directly through slack as well if anything is private.  The network console might be good as well. 
2 - A full tomcat log when you add a BAM would be helpful, but it looks like you aren’t seeing much, which means its probably another error.
3 - Can you add shorter BAMs from that same track?
4 - Can you add BAMs from other tracks of varying lengths?

Thanks,

Nathan


On Oct 9, 2017, at 9:56 AM, Kevin Howe <[hidden email]> wrote:


Hi Nathan,

We are able to reproduce the failure-to-promote-from-BAM problem on master (from today). Upon inspection, we are seeing the below in the server logs. 

Thanks,

Kevin


09-Oct-2017 15:57:05.359 SEVERE [https-openssl-nio-8443-exec-8] org.apache.coyote.AbstractProtocol$ConnectionHandler.process Error reading request, ignored
 java.lang.IllegalStateException: Unexpected read state [CLOSING]
    at org.apache.tomcat.websocket.server.WsFrameServer.notifyDataAvailable(WsFrameServer.java:163)
    at org.apache.tomcat.websocket.server.WsHttpUpgradeHandler.upgradeDispatch(WsHttpUpgradeHandler.java:148)
    at org.apache.coyote.http11.upgrade.UpgradeProcessorInternal.dispatch(UpgradeProcessorInternal.java:54)
    at org.apache.coyote.AbstractProcessorLight.process(AbstractProcessorLight.java:53)
    at org.apache.coyote.AbstractProtocol$ConnectionHandler.process(AbstractProtocol.java:868)
    at org.apache.tomcat.util.net.NioEndpoint$SocketProcessor.doRun(NioEndpoint.java:1459)
    at org.apache.tomcat.util.net.SocketProcessorBase.run(SocketProcessorBase.java:49)
    at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1149)
    at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:624)
    at org.apache.tomcat.util.threads.TaskThread$WrappingRunnable.run(TaskThread.java:61)
    at java.lang.Thread.run(Thread.java:748)



On 26 September 2017 at 17:02, Nathan Dunn <[hidden email]> wrote:

I haven’t seen these effects on master, though there were some effects with the older Track Panel if you had lots of tracks that has since been fixed.  My suspicion is that something is crashing on the client and affecting all of the downstream behaviors.

A couple of quick questions:

1 - Can you provide any console log errors?
2 - Would it be possible for you to try running off of master as an experiment and see if you can repeat the experiment?

gmod/apollo:latest

If you are using another branch from the docker hub (other than 2.0.6), you can replace the WEBAPOLLO_VERSION with  07fed213ea9b122d68efc85c1fe114c7c3948f79 in the Docker file to upgrade the Apollo version, otherwise the “latest” tag will be your best bet.

Nathan

> On Sep 26, 2017, at 3:31 AM, Kevin Howe <[hidden email]> wrote:
>
>
> Hi Apollo developers,
>
> We experiencing a couple of bugs when deploying Apollo for WormBase. The attached movies should hopefully explain what's going on.
>
> In the first, after first logging in, turning on a track results in two copies of the track appearing in the main window; turning it off again results in one of the copies disappearing, but trying to turn the track off completely has no effect (the button keeps slipping back to the on position). Shift-reload fixes this for a time, but the problem often later occurs again later in the session.
>
> In the second, we are trying to promote a long RNASeq read (from a BAM file) into the user annotations track. Nothing happens, and it appears to block other edit operations (promotion of smaller reads and deletes). A shift-reload reenables the ability to do these small edits, but the attempt to the promote the longer read blocks consistently.
>
> A third problem (not in the movies) is that we regularly encounter sessions where promoting features to the user-annotations track works, but nothing appears until you reload the page.
>
> We are using the the Apollo 2.x docker container (https://github.com/GMOD/docker-apollo) to create a local service, and then creating a WAR file which is deployed on EMBL-EBI externally-facing servers using Tomcat (EMBL-EBI do not support docker for externally-facing services yet). And as you can see from the videos, we are using the Neat Features JBrowse plug-in.
>
> We can provide more details if necessary - and even a test account for you to have a play. But to start things off: has anyone seen problems like this before?
>
> Many thanks,
>
> Kevin Howe
> WormBase
>
> ​
>  apollo_bugs.cannot_promote_long_BAM_read.mov​​
>  apollo_bugs.double_track_load.mov​
>
>
>
> This list is for the Apollo Annotation Editing Tool. Info at http://genomearchitect.org/
> If you wish to unsubscribe from the Apollo List: 1. From the address with which you subscribed to the list, send a message to [hidden email] | 2. In the subject line of your email type: unsubscribe apollo | 3. Leave the message body blank.
>





This list is for the Apollo Annotation Editing Tool. Info at http://genomearchitect.org/
If you wish to unsubscribe from the Apollo List: 1. From the address with which you subscribed to the list, send a message to [hidden email] | 2. In the subject line of your email type: unsubscribe apollo | 3. Leave the message body blank.






This list is for the Apollo Annotation Editing Tool. Info at http://genomearchitect.org/
If you wish to unsubscribe from the Apollo List: 1. From the address with which you subscribed to the list, send a message to [hidden email] | 2. In the subject line of your email type: unsubscribe apollo | 3. Leave the message body blank.







This list is for the Apollo Annotation Editing Tool. Info at http://genomearchitect.org/
If you wish to unsubscribe from the Apollo List: 1. From the address with which you subscribed to the list, send a message to [hidden email] | 2. In the subject line of your email type: unsubscribe apollo | 3. Leave the message body blank.