Uploaded image for project: 'Jira Data Center'
  1. Jira Data Center
  2. JRASERVER-33376

Chrome caching issue causes CSS and JavaScript assets to not load

      Chrome seems to have some issues where certain assets which are cached have their mime type clobbered and set to text/plain when retrieved (304 response). This seems to have started rearing its head in March and we are having all kinds of issues with Tempo. The Tempo folks think this should be brought to Atlassian's attention, I imagine because it is affected by which Headers (X-Content-Type-Options: nosniff) are set on certain requests.

      The end result is certain assets are retrieved from Chrome's cache with the wrong mime type which then causes Chrome's JS engine to barf or the CSS to not be parsed. (See the Tempo issue I reference above for screen caps and the errors the Javascript engine throws.)

        1. DashboardRefresh.jpg
          DashboardRefresh.jpg
          61 kB
        2. broken.png
          broken.png
          173 kB
        3. broken.mov
          7.09 MB

            [JRASERVER-33376] Chrome caching issue causes CSS and JavaScript assets to not load

            Jeany added a comment - - edited

            We observe this issue after updating our Jira to LTS version 9.4.19 on our productive system. 
            In contrast, it does not occur in the staging environment.

            For the Gadget Sprint Burndown Chart it also appears with firefox.

            Jeany added a comment - - edited We observe this issue after updating our Jira to LTS version 9.4.19 on our productive system.  In contrast, it does not occur in the staging environment. For the Gadget Sprint Burndown Chart it also appears with firefox.

            Hi scottschulthess,

            We have actually never found an exact set of reproduction steps/configuration to reproduce this reliably.

            Therefore, could you please contact us at https://support.atlassian.com so we can assign an engineer to work with you to try to troubleshoot this further and get to a definite root cause and configuration/reproduction steps?

            Thanks in advance

            Regards,

            Oswaldo Hernández.
            JIRA Bugmaster.
            [Atlassian].

            Oswaldo Hernandez (Inactive) added a comment - Hi scottschulthess , We have actually never found an exact set of reproduction steps/configuration to reproduce this reliably. Therefore, could you please contact us at https://support.atlassian.com so we can assign an engineer to work with you to try to troubleshoot this further and get to a definite root cause and configuration/reproduction steps? Thanks in advance Regards, Oswaldo Hernández. JIRA Bugmaster. [Atlassian] .

            Scott Schulthess added a comment - - edited

            For priority reference, this affects many of the users in our company on a regular basis and is quite confusing and costs us a lot of time, as it's a non intuitive bug with a non intuitive workaround.

            Also, on our system, the issue is only reproducible on chrome (and so far, only on OSX though we have far less linux users) and over Cisco AnyConnect VPN. (Haven't tried any other VPN software).

            Scott Schulthess added a comment - - edited For priority reference, this affects many of the users in our company on a regular basis and is quite confusing and costs us a lot of time, as it's a non intuitive bug with a non intuitive workaround. Also, on our system, the issue is only reproducible on chrome (and so far, only on OSX though we have far less linux users) and over Cisco AnyConnect VPN. (Haven't tried any other VPN software).

            I have noticed an increase of reports from users the last few weeks of this issue.

            Evangelos Vrocharis added a comment - I have noticed an increase of reports from users the last few weeks of this issue.

            MattS added a comment -

            The bug linked to by Oswaldo appears to have been closed long ago. Perhaps this is regression

            MattS added a comment - The bug linked to by Oswaldo appears to have been closed long ago. Perhaps this is regression

            Sam Caldwell (Inactive) added a comment - - edited

            CST has observed this error in JIRA 7.1.0 with Service Desk 3.1.0 (Jira Data Center) when using Chrome.

            The JDC cluster is behind NGINX with gzip turned on at the app server layer.

            Sam Caldwell (Inactive) added a comment - - edited CST has observed this error in JIRA 7.1.0 with Service Desk 3.1.0 (Jira Data Center) when using Chrome. The JDC cluster is behind NGINX with gzip turned on at the app server layer.

            Hi jitendra.daswani,

            This is ultimately caused by a bug in chrome at https://code.google.com/p/chromium/issues/detail?id=246875

            According to some people here enabling Gzip compression under general configuration seemed to work around the bug, however I have not been able to confirm this myself. Could you give that a try and report back?

            Otherwise, could you please create a support ticket at https://support.atlassian.com so we can investigate further? The credentials should be the same as for this site (https://jira.atlassian.com).

            Regards,

            Oswaldo Hernández.
            JIRA Bugmaster.
            [Atlassian].

            Oswaldo Hernandez (Inactive) added a comment - Hi jitendra.daswani , This is ultimately caused by a bug in chrome at https://code.google.com/p/chromium/issues/detail?id=246875 According to some people here enabling Gzip compression under general configuration seemed to work around the bug, however I have not been able to confirm this myself. Could you give that a try and report back? Otherwise, could you please create a support ticket at https://support.atlassian.com so we can investigate further? The credentials should be the same as for this site ( https://jira.atlassian.com ). Regards, Oswaldo Hernández. JIRA Bugmaster. [Atlassian] .

            Hi David Dixon,
            We are also running Tomcat 7 through IIS 7.5 on Windows Server , with SQL Server 2012 for the database (on a separate server) and facing same chrome loading issue.

            Have you got any solution??

            jitendra daswani added a comment - Hi David Dixon, We are also running Tomcat 7 through IIS 7.5 on Windows Server , with SQL Server 2012 for the database (on a separate server) and facing same chrome loading issue. Have you got any solution??

            Further information: the Chrome bug tracking the 304 on CSS assets issue is at https://code.google.com/p/chromium/issues/detail?id=246875

            From browsing the bug report it seems that the fix has been merged and it should be in Chrome 30/31.

            Oswaldo Hernandez (Inactive) added a comment - Further information: the Chrome bug tracking the 304 on CSS assets issue is at https://code.google.com/p/chromium/issues/detail?id=246875 From browsing the bug report it seems that the fix has been merged and it should be in Chrome 30/31.

            Olivier,
            enabling GZIP compression working for me, thanks!

            Тимофеев Кирилл added a comment - Olivier, enabling GZIP compression working for me, thanks!

              Unassigned Unassigned
              7da6d7ffe6d1 Graham Bakay
              Affected customers:
              17 This affects my team
              Watchers:
              26 Start watching this issue

                Created:
                Updated: