[sheepdog] [PATCH] test: Implentment fast random on AES

MORITA Kazutaka morita.kazutaka at lab.ntt.co.jp
Wed May 8 10:19:49 CEST 2013


At Mon, 6 May 2013 20:15:03 -0700,
Joseph Glanville wrote:
> 
> 001  Test master node killing and master transfer
> 002  Test master transfer for crashed cluster

(snip)

> 016  Test vdi clone functionality
> 017  Test vdi tree functionality - output mismatch (see 017.out.bad)
> --- 017.out 2013-05-06 00:56:15.781798666 -0700
> +++ 017.out.bad 2013-05-06 20:01:47.950142025 -0700
> @@ -5,5 +5,5 @@
>  wrote 512/512 bytes at offset 0
>  512 bytes, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
>  test---[DATE]---[DATE]---[DATE]---(you are here)
> -test2---[DATE]-+-[DATE]
> +test2---[DATE]-+-[DATE]---[DATE]
>                             `-[DATE]---(you are here)

This test can be passed with the upcoming QEMU 1.5.

> 018  Test object cache writeback mode
> 019  Test object cache writethrough mode

(snip)

> 047  Test journal file replay
> 048  Test some collie commands - output mismatch (see 048.out.bad)
> --- 048.out 2013-05-06 00:56:15.777798666 -0700
> +++ 048.out.bad 2013-05-06 20:08:51.858152531 -0700
> @@ -20,7 +20,6 @@
>  ---------------------------------------------------
>  127.0.0.1:7001
>  127.0.0.1:7002
> -127.0.0.1:7003
> 
>  obj 807c2b2500000000 locations at epoch 4, copies = 3
>  ---------------------------------------------------
> @@ -31,6 +30,12 @@
>  obj 807c2b2500000000 locations at epoch 5, copies = 3
>  ---------------------------------------------------
>  127.0.0.1:7001
> +127.0.0.1:7002
> +127.0.0.1:7003
> +
> +obj 807c2b2500000000 locations at epoch 6, copies = 3
> +---------------------------------------------------
> +127.0.0.1:7001
>  127.0.0.1:7000
>  127.0.0.1:7002
>  Tracking the object 0x7c2b2500000001 (the inode vid 0x7c2b25 idx 1) with 5
> nodes
> @@ -49,17 +54,22 @@
>  obj 7c2b2500000001 locations at epoch 3, copies = 3
>  ---------------------------------------------------
>  127.0.0.1:7001
> -127.0.0.1:7003
>  127.0.0.1:7002
> 
>  obj 7c2b2500000001 locations at epoch 4, copies = 3
>  ---------------------------------------------------
>  127.0.0.1:7001
> -127.0.0.1:7004
>  127.0.0.1:7003
> +127.0.0.1:7002
> 
>  obj 7c2b2500000001 locations at epoch 5, copies = 3
>  ---------------------------------------------------
> +127.0.0.1:7001
> +127.0.0.1:7004
> +127.0.0.1:7003
> +
> +obj 7c2b2500000001 locations at epoch 6, copies = 3
> +---------------------------------------------------
>  127.0.0.1:7000
>  127.0.0.1:7001
>  127.0.0.1:7004

The local cluster driver has a problem in the leave event handling.
You can pass the test with the corosync driver.

> 049  Test object cache reclaim at startup
> 050  Test IO NIC function
> 051  Test cluster halt
> 052  Test force cluster recovery with new nodes
> 053  Test adding nodes to the gateway only cluster
> 054  Test stale objects purging with different disk size.
> 055  Test md with single failure, multiple failure and all failure
> 056  Test md disk failure with node change events - output mismatch (see
> 056.out.bad)
> --- 056.out 2013-05-06 00:56:15.641798666 -0700
> +++ 056.out.bad 2013-05-06 20:12:37.698158127 -0700
> @@ -8,6 +8,27 @@
>  Epoch Time           Version
>  DATE      2 [127.0.0.1:7000, 127.0.0.1:7001, 127.0.0.1:7002, 127.0.0.1:7003
> ]
>  DATE      1 [127.0.0.1:7000, 127.0.0.1:7001, 127.0.0.1:7002]
> +rm: cannot remove ‘/tmp/sheepdog/056/1/d1’: Directory not empty
> +Failed to write object 7c2b2500000016: No object found
> +Failed to write VDI
> +fixed missing 7c2b250000000d
> +fixed missing 7c2b250000000e
> +fixed missing 7c2b250000000f
> +fixed missing 7c2b2500000016
> +fixed missing 7c2b2500000017
> +fixed missing 7c2b2500000018
> +fixed missing 7c2b250000001a
> +fixed missing 7c2b250000001b
> +fixed missing 7c2b2500000021
> +fixed missing 7c2b2500000025
> +fixed missing 7c2b2500000026
> +fixed missing 7c2b2500000028
> +fixed missing 7c2b250000002a
> +fixed missing 7c2b250000002b
> +fixed missing 7c2b250000002c
> +fixed missing 7c2b250000002d
> +fixed missing 7c2b250000002f
> +fixed missing 7c2b2500000030
>  finish check&repair test
>  Cluster status: running
> 

I also hit this error and it looks like a problem of the test script.
I'll send a fix patch.

> 057  Test md hot-plug and hot-unplug - output mismatch (see 057.out.bad)
> --- 057.out 2013-05-06 00:56:15.685798666 -0700
> +++ 057.out.bad 2013-05-06 20:12:50.010158432 -0700
> @@ -4,12 +4,11 @@
>  0 MASKED GB 32 MB STORE/0/d0
>  1 MASKED GB 32 MB STORE/0/d1
>  2 MASKED GB 40 MB STORE/0/d2
> +Failed to execute request, look for sheep.log for more information
>  Id MASKED Use Path
> -0 MASKED GB 4.0 MB STORE/0/d0
> -1 MASKED GB 20 MB STORE/0/d1
> -2 MASKED GB 28 MB STORE/0/d2
> -3 MASKED GB 20 MB STORE/0/d3
> -4 MASKED GB 24 MB STORE/0/d4
> +0 MASKED GB 28 MB STORE/0/d0
> +1 MASKED GB 32 MB STORE/0/d1
> +2 MASKED GB 36 MB STORE/0/d2
>  finish check&repair test
>  Cluster status: running
> 
> @@ -22,15 +21,11 @@
>  Nodes In Recovery:
>    Id   Host:Port         V-Nodes       Zone
>  Id MASKED Use Path
> -0 MASKED GB 4.0 MB STORE/0/d0
> -1 MASKED GB 20 MB STORE/0/d1
> -2 MASKED GB 28 MB STORE/0/d2
> -3 MASKED GB 20 MB STORE/0/d3
> -4 MASKED GB 24 MB STORE/0/d4
> +0 MASKED GB 28 MB STORE/0/d0
> +1 MASKED GB 32 MB STORE/0/d1
> +2 MASKED GB 36 MB STORE/0/d2
>  Id MASKED Use Path
> -0 MASKED GB 28 MB STORE/0/d2
> -1 MASKED GB 32 MB STORE/0/d3
> -2 MASKED GB 36 MB STORE/0/d4
> +0 MASKED GB 96 MB STORE/0/d2
>  finish check&repair test
>  Cluster status: running
> 
> @@ -43,9 +38,7 @@
>  Nodes In Recovery:
>    Id   Host:Port         V-Nodes       Zone
>  Id MASKED Use Path
> -0 MASKED GB 28 MB STORE/0/d2
> -1 MASKED GB 32 MB STORE/0/d3
> -2 MASKED GB 36 MB STORE/0/d4
> +0 MASKED GB 96 MB STORE/0/d2
>  Cluster status: running
> 
>  Cluster created at DATE

Is this reproducible?  If yes, can you give us your sheep.log after
running "./check 57"?  Seems that md hot-plug failed and some disks
were not added.

> 058  Test discard/trim - output mismatch (see 058.out.bad)
> --- 058.out 2013-05-06 00:56:15.809798666 -0700
> +++ 058.out.bad 2013-05-06 20:12:56.946158604 -0700
> @@ -10,35 +10,135 @@
>  discard 104857600/104857600 bytes at offset 0
>  100 MiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
>  finish check&repair test
> -The inode object 0x7c2b25 idx 0 is not allocated

(snip)

> 059  Test live snapshot - output mismatch (see 059.out.bad)
> --- 059.out 2013-05-06 00:56:15.645798666 -0700
> +++ 059.out.bad 2013-05-06 20:13:12.738158996 -0700
> @@ -1,25 +1,61 @@
>  QA output created by 059
>  using backend farm store
> +Invalid error code

058 and 059 also require the recent QEMU.

Thanks,

Kazutaka



More information about the sheepdog mailing list