search

tntech-ngin / ndn-hydra

ndn-hydra: A Python-coded NDN distributed repository with five focused attributes: resiliency, scalability, usability, efficiency, and security.
Apache License 2.0
2 stars 1 forks source link

TimeoutError when fetch files from Hydra #3

Closed xai990 closed 3 months ago

xai990 commented 4 months ago

Hello, I came across an issue when I tried to fetch the file from Hydra. My Hydra set up is Node 1 as client node and Node 2, Node 3, Node 4 and Node 5 as the Hydra nodes. I can insert the file to Hydra successfully. However, when I ran the fetch command, I had the timeout error: First Packet Content: /hydra/node/node4/fetch/bio.tgz INFO:root:Timeout /hydra/node/node4/fetch/bio.tgz/seg=0 INFO:root:Timeout /hydra/node/node4/fetch/bio.tgz/seg=0 INFO:root:Timeout /hydra/node/node4/fetch/bio.tgz/seg=0 Client Fetch Command Failed. Client finished Fetch Command! - total time (with disk): 12.0193 secs INFO:root:Manually shutdown

Please see the screenshot for details of the error. Screenshot from 2024-02-20 18-34-21

Thanks,

tntech-ngin commented 4 months ago

Hi Xusheng,

I need some more information.

  1. Did you create a new slice for this? If not, please do.
  2. Run everything from scratch.
  3. Insert a larger file, not just 1 packet.
  4. What are the routes in NFD?
  5. You need at least 4 hydra nodes. It seems you have that – please confirm.
  6. The replication requires some time. Please wait a few minutes before fetching. This is due to the heartbeat going out every 3 seconds or so. I can’t tell whem you are requesting these.

Can you please do these quickly and let me know as soon as possible? If you can let me know today or tomorrow, I can further debug this before our next meeting.

tntech-ngin commented 4 months ago

Manas mentioned publishing under a different name gets rid of this problem. And/or logout and login.

tntech-ngin commented 4 months ago

Please pull the updated code and see if this issue still exists/

sankalpatimilsina12 commented 3 months ago

Added /client route from nodes back to node2. They are then able to pull file from client. This resolved the timeout issues at the client and correspondingly NACK issues at these nodes.