Project

General

Profile

Actions

Bug #5301

open

LargeBlock Support lock/semaphore race issue

Added by Patrick Domack over 6 years ago.

Status:
New
Priority:
Normal
Assignee:
-
Category:
zfs - Zettabyte File System
Start date:
2014-11-08
Due date:
% Done:

0%

Estimated time:
Difficulty:
Medium
Tags:
needs-triage
Gerrit CR:

Description

I have been shuffling my data around to play with 1M blocks, and have noticed writes getting hung up in kernel space, not issuing any disk io. I haven't been able to confirm if this affects reads also yet or not.

When doing the process below, it seems to happen pretty quickly.

zfs set recordsize=1M sourcepool/dataset
rsync sourcepool/dataset/ tmppool
rm -rf sourcepool/dataset/*
rsync tmppool/ sourcepool/dataset/

The writing rsync going back to the new 1M dataset, hangs at a seemingly random point.
A reboot will correct this issue, and the copy will make a lot more progress, but will randomly happen again.

It seems to have fixed itself also, if I just wait around 30min, and try again.

No data to display

Actions

Also available in: Atom PDF