[chef] Re: cookbook upload internal server error

[Noah Kantrowitz < >]

We ran into exactly this issue on Friday and just had to `chef-server-ctl 
restart` and delete the cranky hashes in the bookshelf folder (not sure if 
the latter was required).

--Noah

On Feb 23, 2014, at 5:51 PM, 
<
 >
 wrote:

> I just spun up a Chef Server community AMI on EC2.  It's running 11.0.10 and
> I've setup a Chef workstation with 11.10.4 along with knife.  When I run 
> "knife
> cookbook upload" (I tried both --all and a basic one like iptables), I get:
> 
> Uploading iptables     [0.12.0]
> ERROR: internal server error
> Response: internal service error
> 
> On the server side, I run "sudo chef-server-ctl tail" and I see this:
> 
> ...
> 2014-02-24_01:29:10.31425 =ERROR REPORT==== 24-Feb-2014::01:29:10 ===
> 2014-02-24_01:29:10.31426 Checking presence of checksum:
> <<"1f3ac5801386aa2b22e4cbb9877fb431">> for org
> <<"00000000000000000000000000000000">> from bucket "bookshelf" has taken 
> longer
> than 5000 ms
> 2014-02-24_01:29:10.31768
> 2014-02-24_01:29:10.31770 =ERROR REPORT==== 24-Feb-2014::01:29:10 ===
> 2014-02-24_01:29:10.31771 webmachine error:
> path="/sandboxes/000000000000c0ddfbf55f671a4b55be"
> 2014-02-24_01:29:10.31772 {error,
> 2014-02-24_01:29:10.31773     {throw,
> 2014-02-24_01:29:10.31773       {checksum_check_error,13},
> 2014-02-24_01:29:10.31774      
> [{chef_wm_named_sandbox,validate_checksums_uploaded,2,
> 2014-02-24_01:29:10.31775           
> [{file,"src/chef_wm_named_sandbox.erl"},{line,144}]},
> 2014-02-24_01:29:10.31776        {chef_wm_named_sandbox,from_json,2,
> 2014-02-24_01:29:10.31776           
> [{file,"src/chef_wm_named_sandbox.erl"},{line,99}]},
> 2014-02-24_01:29:10.31777        {webmachine_resource,resource_call,3,
> 2014-02-24_01:29:10.31778           
> [{file,"src/webmachine_resource.erl"},{line,166}]},
> 2014-02-24_01:29:10.31778        {webmachine_resource,do,3,
> 2014-02-24_01:29:10.31779           
> [{file,"src/webmachine_resource.erl"},{line,125}]},
> 2014-02-24_01:29:10.31780        {webmachine_decision_core,resource_call,1,
> 2014-02-24_01:29:10.31781           
> [{file,"src/webmachine_decision_core.erl"},{line,48}]},
> 2014-02-24_01:29:10.31781        {webmachine_decision_core,accept_helper,0,
> 2014-02-24_01:29:10.31782           
> [{file,"src/webmachine_decision_core.erl"},{line,583}]},
> 2014-02-24_01:29:10.31783        {webmachine_decision_core,decision,1,
> 2014-02-24_01:29:10.31784           
> [{file,"src/webmachine_decision_core.erl"},{line,489}]},
> 2014-02-24_01:29:10.31784        {webmachine_decision_core,handle_request,2,
> 2014-02-24_01:29:10.31786           
> [{file,"src/webmachine_decision_core.erl"},{line,33}]}]}}
> 
> The ellipsis (...) is actually that first "ERROR REPORT" repeated several
> times, I'm guessing for each piece of the cookbook.  I also see this:
> 
> ==> /var/log/chef-server/erchef/erchef.log.1 <==
> 2014-02-24T01:29:10Z 
> 
 
>  ERR req_id=StLBKjUP42+NHX+GoRa6zQ==;
> status=500; method=PUT; path=/sandboxes/000000000000c0ddfbf55f671a4b55be;
> user=yoway; msg=[]; req_time=5064; rdbms_time=3; rdbms_count=3; 
> s3_time=5039;
> s3_count=1
> 
> That s3_time is just over the 5000ms time in the first error, so I'm 
> wondering
> if this has something to do with s3, although I'm not using s3 afaik.
> 
> Googling these error messages returns discussions about how chef-server
> redirects api requests across its components using fqdn's.  Here are a 
> couple
> suggestions I found and tried:
> 
> 1/ set hostname to public dns name, add public dns name to /etc/hosts,
> reconfigure chef-server
> 2/ use ip address instead of fqdn in /etc/chef-server/chef-server.rb
> 3/ Set s3_url_ttl to 3600 in /etc/chef-server/chef-server.rb
> 
> None of these approaches worked for me.  Any help is appreciated.  Thank 
> you.

Attachment: signature.asc
Description: Message signed with OpenPGP using GPGMail