Allura Platform Instability


We’re currently experiencing poor performance on the Allura platform which powers many of the tools on the SourceForge site. Our teams are working to get this fixed and back to normal as soon as they can. Until then, any pages served by the Allura platform are generally timing out and will fail to load. Some pages will occasionally load, but currently, they are mostly failing. Among others, the affected tools include Tickets, Forums, Wiki, Blog, and Code browsers.

Additionally, permission checks for writes to code repositories also interface with the Allura platform, so this may also cause code writes to intermittently fail with permission errors. Operations that don’t require a permission check (ie. read-only operations) are unaffected.

The File Release System is unaffected by this issue.

UPDATE: Project Icons and Screenshots are currently disabled as we continue to work on this issue.

UPDATE 2: Stability is greatly improved, however, our teams continue to work on this issue to make sure the root cause is addressed so it doesn’t recur. Project Icons and screenshots are also re-enabled.

UPDATE 3: We are now in a normal operating state, this incident is considered resolved.

Chris Tsai, Support

6 Responses to “Allura Platform Instability”

  1. loclator Feb 18, 2014 at 10:08 am #

    sourceforge Ta foda

  2. PilavlMuhterem Feb 18, 2014 at 11:02 am #

    sourceforge sfnet_ops

  3. ilg Feb 25, 2014 at 8:56 am #

    After these problems my account is mostly functional, except I cannot access trackers past the first page. for example: which ends with: Error 500 We’re sorry but we weren’t able to process this request.

  4. Myles Feb 28, 2014 at 11:56 am #

    I am glad it is fixed, that is why i am seeing a lot of errors everywhere! YAY 😀

  5. Dave Watson Mar 2, 2014 at 4:40 pm #

    I’m getting the same problem with Apache Open Office GB, yey the blog says it’s fixed and didn’t apply to file downloads anyway. It’s been hapening for several days downloads fail imediately or within a couple of Mb. Anyone actualy looked at this?

  6. ctsai Mar 3, 2014 at 10:56 am #

    ilg  That was an unrelated issue. This has also been fixed.