Closed renzhengeek closed 5 years ago
Emm, another weird failure as follows.
use lvm can create 1G chunksize thinp;
# lvcreate -L10GM -V2G -n thin1 --thinpool vg/pool0 --chunksize 1G
File descriptor 6 (/dev/pts/1) leaked on lvcreate invocation. Parent PID 4267: -bash
Thin pool volume with chunk size 1.00 GiB can address at most 253.00 PiB of data.
WARNING: Pool zeroing and 1.00 GiB large chunk size slows down thin provisioning.
WARNING: Consider disabling zeroing (-Zn) or using smaller chunk size (<512.00 KiB).
Logical volume "thin1" created.
but largest_data_block_size_succeeds failed to create 1G chunksize using dmsetup
# dmtest run --suite thin-provisioning -t CreationTests
Loaded suite thin-provisioning
CreationTests
activate_thin_while_pool_suspended_fails...PASS
create_lots_of_empty_thins...PASS
create_lots_of_recursive_snaps...PASS
create_lots_of_snaps...PASS
huge_block_size...PASS
largest_data_block_size_succeeds...#<Test::Unit::Error:0x00000000020f9c70
@exception=
#<ProcessControl::ExitError: command failed: dmsetup load test-dev-205043 /tmp/dm-table20190118-15319-kcdnhf>,
@test_name="test_largest_data_block_size_succeeds(CreationTests)">
FAIL
largest_dev_t_succeeds...PASS
non_power_of_2_data_block_size_fails...PASS
too_large_a_dev_t_fails...PASS
too_large_data_block_size_fails...PASS
too_small_a_metadata_dev_fails...PASS
too_small_data_block_size_fails...PASS
kernel error message:
kernel: device-mapper: thin metadata: changing the data block size (from 128 to 2097152) is not supported
kernel: device-mapper: table: 252:0: thin-pool: Error creating metadata object
kernel: device-mapper: ioctl: error adding target to table
`
Ah, the set_mode pool message support is one of your private patch for testing:
commit ef64914d18bba537d92f937860159da7dad8d2ee
Author: Joe Thornber <ejt@redhat.com>
Date: Mon Mar 10 14:39:43 2014 +0000
[dm-thin] set_mode pool message.
This isn't going upstream, just for my testing.
Hi, Many testcases failed in my testing on centos 7.6, but with 4.9 kernel and device-mapper-1.02.149 package:
I haven't look into all failures yet, instead I start with CrashTests::aborting_a_provisioned_thin failure, notice that many cases failed with this error message:
which I can reproduce with dmsetup following the steps in logs.
kernel message:
Does this testsuite need to run on a specific kernel version? I searched around the kernel source(5.0-rc2 and 4.19), but 'set_mode' message is not in pool message list.
Could you please shed some light where I should look at to solve this issue?