[LU-11069] ifort lseek returns wrong position on lustre 2.10.3

• Type: Bug
• Resolution: Fixed
• Priority: Critical
• Fix Version/s: Lustre 2.12.0, Lustre 2.10.5
• Affects Version/s: Lustre 2.12.0, Lustre 2.10.3
• Environment: sles12 lustre 2.10.3
• Severity: 2
• Assignee: John Hammond
• Reporter: Mahmoud Hanafi

Description

On lustre2.10.3 ifort lseek will return the wrong position when using append ">>" operation.

This can be reproduced on

client: sles12 lustre2.103 <–> server: lustre2.7.3

client: sles12 lustre2.103 <–> server: lustre2.10.3

hello.f

      write(6,*) "hello world"
      end 

Working on sles11 and lustre2.7.3

> date > tt
> date >> tt
> cat tt
Wed May 30 13:32:58 PDT 2018
Wed May 30 13:33:04 PDT 2018
> strace ./a.out >> tt  
.....
ioctl(1, SNDCTL_TMR_TIMEBASE or TCGETS, 0x7fffffffbf80) = -1 ENOTTY (Inappropriate ioctl for device)
fstat(1, {st_mode=S_IFREG|0600, st_size=58, ...}) = 0
ioctl(1, SNDCTL_TMR_TIMEBASE or TCGETS, 0x7fffffffbf80) = -1 ENOTTY (Inappropriate ioctl for device)
write(1, " hello world\n", 13)          = 13
lseek(1, 0, SEEK_CUR)                   = 71
ftruncate(1, 71)
                        = 0
> cat tt
Wed May 30 13:32:58 PDT 2018
Wed May 30 13:33:04 PDT 2018
 hello world 
 

• • Note lseek(1,0,SEEK_CUR) set at 71.

Not working on lustre2.10.3.

> rm tt
> date > tt
> date >> tt
> cat tt
Wed May 30 13:41:39 PDT 2018
Wed May 30 13:41:44 PDT 2018

> strace ./a.out >> tt   
....
fstat(1, {st_mode=S_IFREG|0600, st_size=58, ...}) = 0
ioctl(1, TCGETS, 0x7fffffffbf80)        = -1 ENOTTY (Inappropriate ioctl for device)
write(1, " hello world\n", 13)          = 13
lseek(1, 0, SEEK_CUR)                   = 13
ftruncate(1, 13)                        = 0

pfe27:/nobackup/jchang> cat tt
Wed May 30 13pfe27:/nobackup/jchang> 
 

*Note that it's lseek that is getting the **wrong* information about current position after the append. 

Issue Links

• is related to: LU-3044 LSeek SEEK_CUR gives incorrect value after write when file is open with O_APPEND
• is related to: LU-8964 use parallel I/O to improve performance on machines with slow single thread performance

Comments

[Gerrit Updater]

John L. Hammond ([e-mail]) merged in patch https://review.whamcloud.com/32661/
Subject: LU-11069 llite: correct file position after appending writes
Project: fs/lustre-release
Branch: b2_10
Current Patch Set:
Commit: 7bc18a0083b27292dd912dba0ae1aa5010568fb4

[Peter Jones]

Landed for 2.12

[Gerrit Updater]

Oleg Drokin ([e-mail]) merged in patch https://review.whamcloud.com/32641/
Subject: LU-11069 llite: correct file position after appending writes
Project: fs/lustre-release
Branch: master
Current Patch Set:
Commit: f8df907e8aaeaee559f913a68a345011a7aa1f23

[Peter Jones]

Mahmoud

If all goes well it should land to master mid-next week.

Peter

[Mahmoud Hanafi]

The patch fixed the lseek issue in our fortran test. When can we expect for it to land.

$ date > tt
$ date >> tt
$ strace ./a.out >> tt

....
ioctl(1, TCGETS, 0x7fffffff77a0)        = -1 ENOTTY (Inappropriate ioctl for device)
ioctl(1, TCGETS, 0x7fffffffbd00)        = -1 ENOTTY (Inappropriate ioctl for device)
fstat(1, {st_mode=S_IFREG|0600, st_size=58, ...}) = 0
ioctl(1, TCGETS, 0x7fffffffbd00)        = -1 ENOTTY (Inappropriate ioctl for device)
write(1, " hello world\n", 13)          = 13
lseek(1, 0, SEEK_CUR)                   = 71
ftruncate(1, 71)                        = 0
.....


$ cat tt
Fri Jun  8 16:00:58 PDT 2018
Fri Jun  8 16:01:01 PDT 2018
 hello world 

[John Hammond]

Hi Jay,

A port of the change to b2_10 is available at https://review.whamcloud.com/#/c/32661/2.

[Jay Lan (Inactive)]

multiop.c: In function 'main':
multiop.c:677:4: error: implicit declaration of function 'llapi_get_ost_layout_v
ersion' [-Werror=implicit-function-declaration]
   rc = llapi_get_ost_layout_version(fd, &layout_version);
   ^
cc1: all warnings being treated as errors
Makefile:1205: recipe for target 'multiop.o' failed
make[5]: *** [multiop.o] Error 1

[Gerrit Updater]

John L. Hammond ([e-mail]) uploaded a new patch: https://review.whamcloud.com/32661
Subject: LU-11069 llite: correct file position after appending writes
Project: fs/lustre-release
Branch: b2_10
Current Patch Set: 1
Commit: 1f2e3e935946f02950b67e9359a5d4d28216fbfd

[Jay Lan (Inactive)]

More problems in the patch when cherry-picked to b2_10 without being flagged as conflicts. It then caused compilation errors.

Could you please back port to b2_10? Thanks!