Layer7 API Management

 View Only
  • 1.  Weird Gateway behaviour with GMU

    Posted 16 days ago

    Hi everyone,

    I'm having an issue migrating to a gateway instance of mine.

    I'm not able to GMU to it from another machine, however I'm able to GMU when running the GMU on the GW itself.

    We have figured out its something to do at the instance itself since doing the migration from another machine in the same subnet goes direct (not through any LB or FW)

    The normal service of APIs works fine

    Any help / direction would be greatly appreciated.

    Thanks Everyone!

  • 2.  RE: Weird Gateway behaviour with GMU

    Broadcom Employee
    Posted 16 days ago

    You did not really provide specific details regarding your configuration. Would help if you include details regarding the error message you are receiving when running GMU remotely?

    Is your LB  port mapping from 443 to 8443?
    What port are you specifying when calling GMU?

    if not specifying a port then GMU defaults to 8443 and if you LB is port mapping this will cause the issue.

  • 3.  RE: Weird Gateway behaviour with GMU

    Posted 15 days ago
    Hi Barry,

    Apologies for the lack of details.

    Attached are the GMU logs and GW logs.

    For this test we are calling from a jumphost directly to the GW

    We are calling on port 8443

    This issue only occurs when calling the migrateIn command, other commands (migrateOut) is working fine.


    GW.log   7 KB 1 version
    gmu.log   2 KB 1 version

  • 4.  RE: Weird Gateway behaviour with GMU

    Posted 15 days ago

    Have you tried re-publishing the REST Management service to get the latest version.
    Having done that, try increasing the value of the restman.request.message.maxSize cluster property if it still fails.

    Details Here:

  • 5.  RE: Weird Gateway behaviour with GMU

    Posted 15 days ago
    Hi David,

    Thanks for the suggestions.

    I've re-published the REST management service but am still hitting the error.

    For the restman.request.message.maxSIze, I've left it as the default of 50MB as the bundle size is much smaller than that (100kb)

    Any other suggestions that I should try?

  • 6.  RE: Weird Gateway behaviour with GMU

    Broadcom Employee
    Posted 15 days ago
    Hello Shawn,
    thanks for including the logs.

    What version of GMU are you running?
    What version of API GW?
    What version of java on the jump box?

  • 7.  RE: Weird Gateway behaviour with GMU

    Posted 14 days ago

    Hi Barry,

    The info is as follows

    GMU Version : GatewayMigrationUtility-1.6.00-729
    API GW Version : 9.4 CR5

    Java on Jumpbox : JDK 1.8.0_271

    We figured that the Jumphost and GMU is working correctly as we are able to import to our other env's GW as well

  • 8.  RE: Weird Gateway behaviour with GMU

    Posted 15 days ago
    Hi David,
    that is quite the issue, as you mentioned, the migratein bundle is 100kb which is quite small!
    And moreover, it works when doing this from inside the Gateway itself.
    This to me smells like a network issue: the error message

    2022-01-12T10:54:19.200+0800 WARNING 822 com.l7tech.server.policy.assertion.ServerRequestSizeLimit: 2: Stream timeout. Exception caught!

    to me is the key: it looks like there is a size  limiter on the size somewhere that stops this.
    I would
    1. look to see if there is any global policy with size limiter
    2. see if the listening port has some limitations as well

    and I would, just to test, disable the size limiter in the restman service just to test.

    I hope this helps.

    If all else fails, although as a long shot, we do provide a tool for migration ourselves, the APIIDA API Gateway Manager that i would not mind testing for that, if you need,

    let me know how it goes,


    Maurizio Garzelli
    APIIDA Chief Technology Advisor APIM