Ozone 1.4.0 got "org.eclipse.jetty.io.EofException: Early EOF" error for uploading large objects #6649
Replies: 1 comment 1 reply
-
@Lucas12138 In my experience S3G Jetty May I know how big is each part? One way to reduce the likelihood is to reduce the size of the parts in the multipart uploads. Since it is a multipart upload, whenever you find this issue, you can continue the upload the parts starting from the part that threw the exception (i.e. part 424 in your example) and complete it afterwards. There is no need to reupload the previous parts. However, you might need to write a custom program to complete the rest of the multipart upload parts. |
Beta Was this translation helpful? Give feedback.
-
Hi,
we recently upgraded from ozone 1.3.0 to 1.4.0 and enabled erasure coding.
After this, we did some testing. For smaller objects (less than 100GB), usually it works fine.
For large objects, e.g. 1TB, we got the following errors after uploading for around 30min (at the speed of 66492914 bytes per second).
It had around 400 records of success logs like:
then went into
and then the upload was stopped.
Can you help see what may be the cause of this? Thank you
Beta Was this translation helpful? Give feedback.
All reactions