0

Now I am trying to install latest DHF on top of MarkLogic Docker instance (on Docker Mac). I could successfully run MarkLogic 10 docker instance based on the image provided on Docker Hub.

The problem occur when I try to install the latest DHF. I tried both ways, running DHF jar from either "host" or from inside "container instance". However, both ways lead to fails at initiating Data Hub project (48%, Step 24 of 47). I see the following error on the quickstart log window.

java.lang.RuntimeException: com.marklogic.client.MarkLogicIOException: java.io.IOException: unexpected end of stream on Connection{localhost:8010, proxy=DIRECT hostAddress=localhost/0:0:0:0:0:0:0:1:8010 cipherSuite=none protocol=http/1.1} at com.marklogic.hub.deploy.commands.GenerateFunctionMetadataCommand.execute(GenerateFunctionMetadataCommand.java:88)

Any advice will be appreciated.

  • BTW, I've also tried on my Windows 10 laptop, but in this case it actually worked fine... –  Feb 27 '20 at 08:19

2 Answers2

1

The above answer is incorrect. If the deployment tool - either DHF or ml-gradle - is not waiting for a server to reboot, that's a bug. The comment posted by james-kerr for the answer above is most likely the correct solution - to increase the amount of RAM assigned to Docker.

rjrudin
  • 2,108
  • 9
  • 7
  • It used to be true, but you are right that it should be considered a bug if the deployment tool doesn't wait. There could still be reasons why MarkLogic restarts unexpectedly. Insufficient memory can be one of them, and is commonly seen when using docker. There may be other reasons too though. – grtjn Aug 06 '21 at 13:07
0

The unexpected end of stream on Connection type of messages typically occur when MarkLogic has been triggered to reboot to apply config updates (only necessary for particular settings, like port changes, and such), while deployment is not anticipating that. The usual solution is to just repeat the last command, to make it try again, and kind of continue where it left off.

It most often happens during initial deployment when a lot of app servers, and databases are added and configured. It should not occur again once your Data Hub is in place.

HTH!

grtjn
  • 20,254
  • 1
  • 24
  • 35
  • According to your advice I added some additional resource to Docker and proceeded calmly.. Then, it worked perfect! Thank you so much for your advice. –  Feb 28 '20 at 01:46
  • For others that may run into this error, to add to grtjn's answer, it can occur when running ml-gradle or any client against MarkLogic in Docker without enough RAM and can happen sporadically even if MarkLogic doesn't restart. The solution as noted is to increase the RAM available to your MarkLogic container. I run with 4 cpus and 8 GB of RAM for my dev machine. – James Kerr Aug 05 '21 at 08:30