IGV stopped working properly on all computers at my laboratory

[Ryan Pfeiffer]

I was wondering if anyone was experiencing any problems with IGV not functioning properly.  A few days ago all of our computers were working great on IGV, however today, one will not work at all, some will load and browse slowly but do not show the color coding in the coverage track for mixed reads, nor the nucleotide, amino  acid, or gene information for the reference.  Our IT department has been working on this all day trying to make progress, but even a clean install on a fresh OS doesn’t work properly.  attached is a screenshot of what I am seeing.

[James Robinson]

Hi,

It looks like something has changed in your network environment and you are having problems accessing the reference sequence.   Perhaps a new firewall or proxy?    For example, for that assembly and chr 7 the sequence would be accessed from this url:  http://igv.broadinstitute.org/genomes/seq/b37/7.txt.

Another potential cause could be a change in proxy or firewall that would filter out "range byte" request headers from outgoing http requests.

There might be more information on specific errors in your igv.log file.  If you attach it I will have a look.   

On Fri, Oct 14, 2016 at 11:57 AM, Ryan Pfeiffer <pfeiff...@gmail.com> wrote:

I was wondering if anyone was experiencing any problems with IGV not functioning properly.  A few days ago all of our computers were working great on IGV, however today, one will not work at all, some will load and browse slowly but do not show the color coding in the coverage track for mixed reads, nor the nucleotide, amino  acid, or gene information for the reference.  Our IT department has been working on this all day trying to make progress, but even a clean install on a fresh OS doesn’t work properly.  attached is a screenshot of what I am seeing.

--

---
You received this message because you are subscribed to the Google Groups "igv-help" group.
To unsubscribe from this group and stop receiving emails from it, send an email to igv-help+unsubscribe@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/igv-help/a22ca561-33fd-4bf0-aa0c-e3b8a0f5b800%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Xian Fan]

Dear Jim,

I have the similar yet different problem. Last night everything is working fine. But today the IGV cannot get to the particular position. Here's the error message:

Loading resource, path /home/xfan3/plot/small_bam/IL_1_64541948_IL.sorted.bam

INFO [2016-10-14 14:47:04,703]  [HttpUtils.java:839] [AWT-EventQueue-0]  Range-byte test failed -- Host: igvdata.broadinstitute.org does not support range-byte requests or there is a problem with client network environment.

What does that mean? Thanks.

-Xian

On Friday, October 14, 2016 at 2:09:11 PM UTC-5, Jim Robinson wrote:

Hi,

It looks like something has changed in your network environment and you are having problems accessing the reference sequence.   Perhaps a new firewall or proxy?    For example, for that assembly and chr 7 the sequence would be accessed from this url:  http://igv.broadinstitute.org/genomes/seq/b37/7.txt.

Another potential cause could be a change in proxy or firewall that would filter out "range byte" request headers from outgoing http requests.

There might be more information on specific errors in your igv.log file.  If you attach it I will have a look.   

On Fri, Oct 14, 2016 at 11:57 AM, Ryan Pfeiffer <pfeiff...@gmail.com> wrote:

I was wondering if anyone was experiencing any problems with IGV not functioning properly.  A few days ago all of our computers were working great on IGV, however today, one will not work at all, some will load and browse slowly but do not show the color coding in the coverage track for mixed reads, nor the nucleotide, amino  acid, or gene information for the reference.  Our IT department has been working on this all day trying to make progress, but even a clean install on a fresh OS doesn’t work properly.  attached is a screenshot of what I am seeing.

--

---
You received this message because you are subscribed to the Google Groups "igv-help" group.

To unsubscribe from this group and stop receiving emails from it, send an email to igv-help+u...@googlegroups.com.

[Ryan Pfeiffer]

Jim,

i found the log file and attached it.  We are getting the error " [HttpUtils.java:834]  Range-byte test failed -- Host: igv.broadinstitute.org does not support range-byte requests or there is a problem with client network environment." 

 

However, i see we have been getting this on a regular basis and it has been working for me on a regular basis so i dont think this is related.

i spoke with IT and they said nothing should have changed in our network, firewall, or proxy recently.

please have a look and let me know if you see anything or have any ideas.

Thanks

Ryan

On Friday, October 14, 2016 at 3:09:11 PM UTC-4, Jim Robinson wrote:

Hi,

It looks like something has changed in your network environment and you are having problems accessing the reference sequence.   Perhaps a new firewall or proxy?    For example, for that assembly and chr 7 the sequence would be accessed from this url:  http://igv.broadinstitute.org/genomes/seq/b37/7.txt.

Another potential cause could be a change in proxy or firewall that would filter out "range byte" request headers from outgoing http requests.

There might be more information on specific errors in your igv.log file.  If you attach it I will have a look.   

On Fri, Oct 14, 2016 at 11:57 AM, Ryan Pfeiffer <pfeiff...@gmail.com> wrote:

I was wondering if anyone was experiencing any problems with IGV not functioning properly.  A few days ago all of our computers were working great on IGV, however today, one will not work at all, some will load and browse slowly but do not show the color coding in the coverage track for mixed reads, nor the nucleotide, amino  acid, or gene information for the reference.  Our IT department has been working on this all day trying to make progress, but even a clean install on a fresh OS doesn’t work properly.  attached is a screenshot of what I am seeing.

--

---
You received this message because you are subscribed to the Google Groups "igv-help" group.

To unsubscribe from this group and stop receiving emails from it, send an email to igv-help+u...@googlegroups.com.

[Xian Fan]

Dear Jim,

An update: I found that an old version (2.3.32) installed on the same server is working fine. The version that I am trying to use but have the range-byte error is the one supporting "consensus" of Pacbio reads. Would that be a host problem on broad, as Ryan also has the same error message? 

-Xian

[James Robinson]

Which genome assembly?  

It most often means there is a proxy in your local environment that is stripping request headers, specifically "Range" headers, from outgoing request.  The "Squid" proxy is notorious for this.   However if you tell me the assembly and what locus you were trying to access I'll test from here.

[Xian Fan]

Thanks so much Jim. It is human genome build38. It looks no locus works, but what I tried are (chr1:64541948, chr1:71992591, chr20:100000-101000). Thank you!

-Xian

[James Robinson]

I can't reproduce your problem.   What version of IGV are you using?  Could you attach your igv.log file?

[Xian Fan]

Hi Jim,

It is a develop version downloaded from http://software.broadinstitute.org/software/igv/download_snapshot (IGV_Snapshot.zip).  Attached is igv.log.

Thanks!

-Xian

[James Robinson]

Well, there isn't any problem on the server that I can detect.    Could you try shutting down, deleting the file  "hg38.genome" in igv/genomes, then restarting and try again?

What sort of network environment are you in?   

--

---
You received this message because you are subscribed to the Google Groups "igv-help" group.

To unsubscribe from this group and stop receiving emails from it, send an email to igv-help+unsubscribe@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/igv-help/14934d14-e084-4f66-b787-fde12f5dd025%40googlegroups.com.

[Xian Fan]

Thanks Jim! I tried remaking the hg38.genome from the fasta file in igv, and it works now. It's so kind of you to give me this suggestion. I really appreciate it!

-Xian

On Friday, October 14, 2016 at 5:52:40 PM UTC-5, Jim Robinson wrote:

Well, there isn't any problem on the server that I can detect.    Could you try shutting down, deleting the file  "hg38.genome" in igv/genomes, then restarting and try again?

What sort of network environment are you in?   

On Fri, Oct 14, 2016 at 3:35 PM, Xian Fan <xianf...@gmail.com> wrote:

Hi Jim,

It is a develop version downloaded from http://software.broadinstitute.org/software/igv/download_snapshot (IGV_Snapshot.zip).  Attached is igv.log.

Thanks!

-Xian

On Friday, October 14, 2016 at 5:19:30 PM UTC-5, Jim Robinson wrote:

I can't reproduce your problem.   What version of IGV are you using?  Could you attach your igv.log file?

On Fri, Oct 14, 2016 at 2:31 PM, Xian Fan <xianf...@gmail.com> wrote:

Thanks so much Jim. It is human genome build38. It looks no locus works, but what I tried are (chr1:64541948, chr1:71992591, chr20:100000-101000). Thank you!

-Xian

--

---
You received this message because you are subscribed to the Google Groups "igv-help" group.

To unsubscribe from this group and stop receiving emails from it, send an email to igv-help+u...@googlegroups.com.

[Ryan Pfeiffer]

Jim,

So far we have deleted the genome (37) and reloaded it.  We have tried re-installing the IGV on windows 10 computers as well as an old windows 7 computer.  We have completely turned off the entire laboratory firewall temporarily.  

However, we are still in the situation where we have no nucleotides or amino acids showing up for the reference (only empty boxes), as well as no coloring of the coverage track for the mixed reads.  if we click on the location it shows the mixed bases, and if we click on the amino acid location, it gives us the gene, exon, amino acid number etc.  it is also running extremely slow, with a long delay after providing the search coordinates.

Do you have any suggestions on what else we should try?

Thanks for your help

Ryan

[James Robinson]

Hi,

These symptoms all point to a network issue, IGV is not able to load the sequence.  Did you try loading http://igv.broadinstitute.org/genomes/seq/b37/7.txt in a web browser?  

Can you attach an igv.log file,  you'll find it in an igv folder under the users home directory.

One way to resolve this is setup a locallly hosted genome on your own server.   Users would then load it from the Genomes menu by URL.  If you want to pursue that route let me know,  there are some manual steps to create one but its not difficult.

Jim

[Anh Le]

Hi Jim,

I am also having similar issues with IGV where I cannot zoom into the nucleotide level to see the reads.  On the Java console I get same error of Range-byte test failed (attached file of the Java readout).  I'm currently using version 2.3.82 IGV and Java version 8 Update 101. The genome assembly I'm using is human hg19.  Your solution is to setup a locally hosted genome.  How do I go about doing this for human hg19?

Thanks

[James Robinson]

Could you please attach your igv.log file?  Its in your igv folder, under your home directory.

[Anh Le]

Hi Jim,

Attached is the log file

[James Robinson]

OK, to everyone having this problem could you try updating IGV to the just released version, 2.3.83,  and report the results back?  Some of you might have opened a git issue as well.  

--

---
You received this message because you are subscribed to the Google Groups "igv-help" group.

To unsubscribe from this group and stop receiving emails from it, send an email to igv-help+unsubscribe@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/igv-help/3ebc1be4-1357-4580-a648-40d17dd1fcd9%40googlegroups.com.

[Anh Le]

Hello Jim,

Updating to the new version 2.3.83 corrected the issue.  It's running normal again.

Thanks for your assistance

Anh

On Tuesday, October 18, 2016 at 1:57:03 AM UTC-6, Jim Robinson wrote:

OK, to everyone having this problem could you try updating IGV to the just released version, 2.3.83,  and report the results back?  Some of you might have opened a git issue as well.  

On Mon, Oct 17, 2016 at 3:53 PM, Anh Le <daffyw...@gmail.com> wrote:

Hi Jim,

Attached is the log file

On Monday, October 17, 2016 at 4:35:18 PM UTC-6, Jim Robinson wrote:

Could you please attach your igv.log file?  Its in your igv folder, under your home directory.

--

---
You received this message because you are subscribed to the Google Groups "igv-help" group.

To unsubscribe from this group and stop receiving emails from it, send an email to igv-help+u...@googlegroups.com.

[James Robinson]

Your welcome.  However, I should let you know that I suspect it actually hides an issue on your end.    If you look in your igv.log file I think you will find the "range byte tests" are still failing.    Range-byte headers are added to http requests from IGV to fetch just those portions of larger files,  in this case the sequence fasta,  that are required.  Without them the entire file would be loaded.     Some proxies remove these headers so the server never sees them.   To prevent the entire file being transferred to your computer we do a test on a small file to detect this condition.   This is where the error message comes from.  If that test fails a back-up strategy is followed that uses a web service running at the Broad that takes range parameters, rather than the header.    It was this web service that changed and required the new release.

So my recommendation would be to find what is removing Range headers from your outgoing http requests and change the setting so it does not do that.   The "Squid" proxy will do this for example,  but probably others as well.

Best

Jim

To unsubscribe from this group and stop receiving emails from it, send an email to igv-help+unsubscribe@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/igv-help/7c73875f-95e7-45e8-9297-97aa9f11e86c%40googlegroups.com.

[Anh Le]

Hi Jim,

How would I go about finding out about the Range headers as per your recommendation?

Anh

[James Robinson]

Its a question for your IT department.   It is not uncommon, particularly in hospital environments, for all outgoing http/https requests to pass through a proxy.  The proxies often have settings to remove headers that are not on a "whitelist",  sometimes Range is not allowed through simply because its not a common header and not on the list.  Changing the settings to allow it through will fix the problem.   Some departments have whitelisted our servers to allow everything through, which also fixes the problem.   The current solution, which relies on the Broad webservice,  will work but there is a performance penalty.

You will know when this is resolved when you stop seeing the "Range byte test" failure errors in the igv.log.