therockliu / stressapptest Goto Github PK
View Code? Open in Web Editor NEWAutomatically exported from code.google.com/p/stressapptest
License: Apache License 2.0
Automatically exported from code.google.com/p/stressapptest
License: Apache License 2.0
What steps will reproduce the problem?
1.make CXXFLAGS=-static
What is the expected output? What do you see instead?
g++ -static -pthread -laio -o stressapptest main.o os.o os_factory.o
pattern.o queue.o sat.o sat_factory.o worker.o finelock_queue.o error_diag.o
disk_blocks.o adler32memcpy.o logger.o
worker.o: In function `DiskThread::Work()':
worker.cc:(.text+0x1aac): undefined reference to `io_setup'
worker.cc:(.text+0x1b49): undefined reference to `io_destroy'
worker.o: In function `DiskThread::AsyncDiskIO(DiskThread::IoOp, int, void*,
long long, long long, long long)':
worker.cc:(.text+0x29de): undefined reference to `io_submit'
worker.cc:(.text+0x2b6e): undefined reference to `io_getevents'
worker.cc:(.text+0x2d0c): undefined reference to `io_cancel'
worker.cc:(.text+0x2d1f): undefined reference to `io_destroy'
worker.cc:(.text+0x2d49): undefined reference to `io_setup'
collect2: ld returned 1 exit status
What version of the product are you using? On what operating system?
stressapptest-1.0.2_autoconf, OS is 2.6.18-1.2798.fc6(GNU/Linux)
Please provide any additional information below.
[root@pc-87-12-odi-pod2-br-pc1 stressapptest-1.0.2_autoconf]# ls -l
/usr/lib64/libaio.*
-rw-r--r-- 1 root root 57470 Apr 19 23:40 /usr/lib64/libaio.a
lrwxrwxrwx 1 root root 15 Apr 19 23:40 /usr/lib64/libaio.so ->
libaio.so.1.0.1
lrwxrwxrwx 1 root root 15 Apr 19 23:40 /usr/lib64/libaio.so.1 ->
libaio.so.1.0.1
-rwxr-xr-x 1 root root 19775 Apr 19 23:40 /usr/lib64/libaio.so.1.0.1
[root@pc-87-12-odi-pod2-br-pc1 stressapptest-1.0.2_autoconf]#
[root@pc-87-12-odi-pod2-br-pc1 stressapptest-1.0.2_autoconf]#
[root@pc-87-12-odi-pod2-br-pc1 stressapptest-1.0.2_autoconf]# find / -name
librt.a
/usr/lib/librt.a
/usr/lib64/librt.a
Original issue reported on code.google.com by [email protected]
on 20 Apr 2011 at 6:09
Machines with cpu problems or ecc events often generate machine checks.
Publicly available MCE decode should be printed by stressapptest.
Original issue reported on code.google.com by [email protected]
on 19 Mar 2010 at 4:54
What steps will reproduce the problem?
1.run stressapptest with the command "stressapptest -s 3600"
2.
3.
What is the expected output? What do you see instead?
We expect the result should be 62000 on all machines, but the fact is that some
machines get the result less than 40000, some machines get the result which is
62000.
And it could be reproduced always.
What version of the product are you using? On what operating system?
Stats: SAT revision 1.0.4_autoconf, 64 bit binary
os version is CentOS 4.3, and kernel 2.6.32
Please provide any additional information below.
The machines which run fail had run pass when be changed to REDHAT 6.5.
Original issue reported on code.google.com by [email protected]
on 13 May 2015 at 10:37
the configure.ac handling of target/host is incorrect. you should delete these
lines:
AC_CANONICAL_BUILD
AC_CANONICAL_TARGET
the "host" in autoconf terms is where the code will be run, not the "target".
then change $target_cpu to $host_cpu in the few places it gets used.
further, the code currently runs `uname` to figure out the os. please use
$host_os instead as autoconf will set that up for you.
Original issue reported on code.google.com by [email protected]
on 27 Nov 2012 at 6:57
What steps will reproduce the problem?
1.stressapptest -s 28800
What do you see instead?
The screen will show bellow message:
Show pausing worker threads in preparation for power spike (28200 seconds
remaining)
Resuming worker threads to cause a power spike(28185 seconds remaining)
Why it will show these messages?
It cannot finish and keep running.
What version of the product are you using? On what operating system?
1.0.6 64bit binary
Original issue reported on code.google.com by [email protected]
on 28 Jul 2014 at 3:32
Hi,
the current version of stressapptest fails to build on several architectures
due to an arch check that looks for x86_64, i686, powerpc and armv7a.
We've tracked this issue in Debian's BTS at:
http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=598756
http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=599016
For the build status see:
https://buildd.debian.org/status/package.php?p=stressapptest
Would be nice if the configure script wouldn't have those strict tests and
instead allows compilation on systems other than i386, amd64 and powerpc again.
:)
thanks && regards,
-mika-
Original issue reported on code.google.com by [email protected]
on 16 Feb 2011 at 3:08
the Makefile.am should read:
dist_man_MANS = stressapptest.1
otherwise, the man page doesn't get installed to the right place
then just omit COPYING altogether as automake will import it correctly by
default. you can verify by running `make distcheck`.
Original issue reported on code.google.com by [email protected]
on 27 Nov 2012 at 6:51
the code does this:
if test "$with_static" == "yes"
that should be one = as that is the only one in POSIX
Original issue reported on code.google.com by [email protected]
on 4 Aug 2014 at 7:51
the 3rd arg to shmat() is an integer, not a pointer. the current os.cc file
does:
shmaddr = shmat(shmid, NULL, NULL);
which triggers a warning on recent Linux/glibc systems:
os.cc:416:44: warning: passing NULL to non-pointer argument 3 of ‘void*
shmat(int, const void*, int)’ [-Wconversion-null]
simply change the 3rd arg to 0
Original issue reported on code.google.com by [email protected]
on 27 Nov 2012 at 6:58
What steps will reproduce the problem?
1.../configure --build=x86 --target=armv7a-none-linux-gnueabi
--host=arm-none-linux-gnueabi --program-prefix=arm-none-linux-gnueabi-
--with-static --prefix=/people/pradeepb/test
2.make install
What is the expected output? What do you see instead?
I did not expect to see any warnings about the machine type. However, i see
os.h:160:4: warning: #warning "Unsupported CPU type: Unable to force cache
flushes."
What version of the product are you using? On what operating system?
1_0_7; Linux/Ubuntu
Please provide any additional information below.
Original issue reported on code.google.com by [email protected]
on 9 Sep 2014 at 1:37
Attachments:
What steps will reproduce the problem?
1.
2.
3.
What is the expected output? What do you see instead?
What version of the product are you using? On what operating system?
Please provide any additional information below.
Hi
would like to know which version of linus OS should i install on the machine to
run this stress test will 64 bit version work fine with this latest version of
stress test
thanks
Original issue reported on code.google.com by [email protected]
on 3 Nov 2011 at 4:37
What steps will reproduce the problem?
1. aclocal
2. automake --gnu
3. autoconf
4. ./configure./configure --build=x86 --target=armv7a-none-linux-gnueabi
--host=arm-none-linux-gnueabi
What is the expected output? What do you see instead?
Expected: configuration to be completed successfully.
Instead: checking for pthread_barrier... configure: error: in
`/stressapptest-1.0.6_autoconf':
configure: error: cannot run test program while cross compiling
See `config.log' for more details.
What version of the product are you using? On what operating system?
stressapptest-1.0.6
Please provide any additional information below.
Instead of AC_TRY_RUN perform AC_COMPILE_CHECK to check pthread_barrier.
During cross compilation, perform compilation check instead of run check for
pthread_barrier.
Original issue reported on code.google.com by [email protected]
on 6 Mar 2013 at 12:13
Attachments:
Dear Sir
I using 1.0.3 to running some testing , but i meat a problem, Does this stress-tool support multiple hdds ? i don't knonw how to run multiple diskst testing at the same time, maybe someone can tell me how to do , thanks a lot...
Original issue reported on code.google.com by [email protected]
on 14 Nov 2011 at 2:57
What steps will reproduce the problem?
1.Commandline - stressapptest -f /mnt/1 --filesize 1024000 -s 216000 -l
stress_60hr_sys5_8T
What is the expected output? What do you see instead?
1. stressapptest can not complete and system does not hang up
2. The screen show "Pausing worker threads in preparation for power spike(35400
seconds remaining)"
What version of the product are you using? On what operating system?
1. Stressapptest : SAT revision 1.0.5_autoconf, 64 bit binary
2. OS : CentOS6.4
Please provide any additional information below.
Original issue reported on code.google.com by [email protected]
on 13 Oct 2014 at 10:13
Attachments:
What steps will reproduce the problem?
1. ./configure --host=mips64
2. Compiler used is mips-wrs-linux-gnu-mips64_octeon-glibc_small-g++
3. Finally, we get linking error
What is the expected output? What do you see instead?
Linking should be successful. Instead of that we get below linking error
mips-wrs-linux-gnu-mips64_octeon-glibc_small-c++ -DCHECKOPTS -Wreturn-type
-Wunused -Wuninitialized -Wall -O3 -DNDEBUG -pthread -o stressapptest main.o
os.o os_factory.o pattern.o queue.o sat.o sat_factory.o worker.o
finelock_queue.o error_diag.o disk_blocks.o adler32memcpy.o logger.o
/home/gur16663/gur16663_etpa/common/platform/octeon/kits/windriver/WR-3.0/toolch
ain/x86-linux2/bin/../lib/gcc/mips-wrs-linux-gnu/4.3.2/../../../../mips-wrs-linu
x-gnu/bin/ld: cannot find -lstdc++
What version of the product are you using? On what operating system?
stressapptest-1.0.1_autoconf
Please provide any additional information below.
We want to run stresapptest on cavium network's mips cpu.
Original issue reported on code.google.com by [email protected]
on 20 Oct 2010 at 11:11
What steps will reproduce the problem?
1. Run stressapptest
2. Sometimes crashes or runs very slowly
3. Memory is being swapped to disk, or system has OOMed
What is the expected output? What do you see instead?
stressapptest should avoid swapping or OOM somehow, or report an explicit error.
Original issue reported on code.google.com by [email protected]
on 6 Nov 2011 at 12:57
What steps will reproduce the problem?
1.first do ./configure, sucessfully done
2.then I try to make, got this error
3.
What is the expected output? What do you see instead?
What version of the product are you using? On what operating system?
the version I used is stressapptest-1.0.3-autoconf, the OS is Redhat9.0
Please provide any additional information below.
Original issue reported on code.google.com by [email protected]
on 29 Mar 2011 at 5:33
the current cpuid logic clobbers %ebx. this is OK if the code is not PIC, but
if you're building PIEs, it'll fail because %ebx is the PIC register and gcc
doesn't let you clobber it.
simple fix:
int ax, bx, cx, dx;
__asm__ __volatile__ (
"mov %%ebx, %%edi;"
"cpuid;"
"mov %%ebx, %%esi;"
"mov %%edi, %%ebx;"
:"=a" (ax), "=S" (bx), "=c" (cx), "=d" (dx) : "a"(1) :"edi");
it's a few more insns on x86_64 or i386/non-pic, so if you want, you could do
something like:
#if defined(STRESSAPPTEST_CPU_I686) && defined(__PIC__)
... do pic version here ...
#else
... do non-pic version here ...
#endif
Original issue reported on code.google.com by [email protected]
on 27 Nov 2012 at 7:19
What steps will reproduce the problem?
(1) stressapptest -s 20 -C 4 --cc_test --stop_on_errors -i 4 -W -M 4096 -f
/file.1
or
(2) stressapptest -s 20 -C 4 --cc_test --stop_on_errors -i 4 -W -M 4096 -d
or
(3) stressapptest -s 20 -C 4 --cc_test --stop_on_errors -i 4 -W -M 4096 -d /
What is the expected output? What do you see instead?
(1)
Expect: File Copy: ***value***M at ***value***MB/s
Actual: Process Error: Failed to create file file.1!!
File Copy: 0.00M at 0.00MB/s
(2)
Expect: Disk: ***value***M at ***value***MB/s
Actual: Disk: 0.00M at 0.00MB/s
(3)
Expect: Disk: ***value***M at ***value***MB/s
Actual: Process Error: Fatal issue encountered. See above logs for details.
Disk: 0.00M at 0.00MB/s
What version of the product are you using? On what operating system?
stressapptest revision 1.0.6_autoconf, 64 bit binary
Please provide any additional information below.
I am trying to get the speed of the primary hard drive, in other words the
speed of the drive the os is running on. I do not care if this is by using -d
or by using -f but I would prefer to use stressapptest as I am already using it
for some other tests. I am able to get -f info for USB drives I have in the
system, but not the primary SSD drive. Attached you will find the output of the
terminal window where I have attempted to get the SSD information.
Original issue reported on code.google.com by [email protected]
on 10 Apr 2014 at 6:03
Attachments:
Download, and
1. ./configure - ok
2. make - error:
Making install in src
g++ -DHAVE_CONFIG_H -I. -g -O2 -DCHECKOPTS -Wreturn-type -Wunused -
Wuninitialized -Wall -O3 -funroll-all-loops -funroll-loops -DNDEBUG -MT
main.o -MD -MP -MF .deps/main.Tpo -c -o main.o main.cc
In file included from main.cc:17:
sattypes.h: In function `std::string ErrorString(int)':
sattypes.h:144: error: invalid conversion from `int' to `const char*'
sattypes.h:144: error: initializing argument 1 of
`std::basic_string<_CharT, _Traits, _Alloc>::basic_string(const _CharT*,
const _Alloc&) [with _CharT = char, _Traits = std::char_traits<char>,
_Alloc = std::allocator<char>]'
In file included from sat.h:31,
from main.cc:18:
worker.h:29:27: linux/aio_abi.h: No such file or directory
In file included from sat.h:31,
from main.cc:18:
worker.h: At global scope:
worker.h:730: error: `aio_context_t' does not name a type
*** Error code 1
Stop in ~/my/stressapptest-1.0.0_autoconf/src.
*** Error code 1
Stop in ~/my/stressapptest-1.0.0_autoconf.
Original issue reported on code.google.com by [email protected]
on 20 Oct 2009 at 6:52
What steps will reproduce the problem?
1. Run stressapptest
2. Sometimes crashes or runs very slowly
3. Memory is being swapped to disk, or system has OOMed
What is the expected output? What do you see instead?
stressapptest should avoid swapping or OOM somehow, or report an explicit error.
Original issue reported on code.google.com by [email protected]
on 6 Nov 2011 at 12:57
What steps will reproduce the problem?
Run stressapptest on CentOS5.2 text mode.
What is the expected output? What do you see instead?
Assertion failed at sattypes.h:106
Assertion failed at logger.cc:74
Assertion failed at sattypes.h:106
What version of the product are you using? On what operating system?
stressapptest-1.0.1_autoconf.tar.gz
CentOS5.2
Please provide any additional information below.
CentOS5.2 can only install stressapptest-1.0.1, it will show make error when we
using stressapptest-1.0.6 version.
Original issue reported on code.google.com by [email protected]
on 10 Apr 2014 at 1:39
Attachments:
From DmitriyM:
Process Error: while initializing empty_ list
Some digging revealed that FineLockPEQueue::PutEmpty returns false
as it can't unlock FineLockPEQueue::pagelocks_ mutexes ->
Sat::InitializePages fails. Mutexes are never locked as
FineLockPEQueue::GetRandomWithPredicateTag (which locks them) is
actually never called.
Original issue reported on code.google.com by [email protected]
on 23 Oct 2009 at 4:09
What steps will reproduce the problem?
1. stressapptest -f stresstest -f stresstest
What is the expected output? What do you see instead?
It should start only one thread per file/device and possibly complain
about concurrent usage of the same file. The File/device should be opened
exclusively.
What version of the product are you using? On what operating system?
Stats: SAT revision 1.0.1_autoconf, 64 bit binary
Debian Lenny
Original issue reported on code.google.com by [email protected]
on 3 Dec 2009 at 3:27
What steps will reproduce the problem?
1. I have tried various options for running the Disk tests like below but the
stats is not printing the disk and data check performance numbers after the
test runs.
./stressapptest -d /dev/sdb -s 10 -v 20 --destructive
./stressapptest -d /dev/sdb -s 10 -v 20 --destructive --findfiles
./stressapptest -d /dev/sda1 --write-block-size 4096 -v 20 -s 10 --destructive
--cache-size 54mb --write-threshold 200000
./stressapptest -d /dev/sda1 --write-block-size 4096 -v 20 -s 10 --destructive
--cache-size 64mb --write-threshold 200000 --findfiles
./stressapptest -d /dev/sda1 --write-block-size 4096 -s 10 --destructive
--cache-size 64mb --write-threshold 200000
./stressapptest -f /mnt/sda1/file2 -f /mnt/sda2/file1
./stressapptest -f /mnt/sda1/file2 -f /mnt/sda2/file1 --filesize 20m
--read-block-size 1024
./stressapptest -d /dev/sda1 -d /dev/sda2
./stressapptest -d /dev/sda
./stressapptest -d /mnt/sda1 --read-block-size 1024
./stressapptest -d /dev/sda1 --read-block-size 1024
./stressapptest -d /dev/sda1 --read-block-size 1024 --verbose 20
./stressapptest -d /dev/sda1 --read-block-size 1024 -v 20
./stressapptest -d /dev/sda1 -f /mnt/sda1/file3 --read-block-size 1024 -v 20
./stressapptest -d /dev/sda1 -f /mnt/sda1/file3 --write-block-size 4096 -v 20
-s 10
./stressapptest -d /dev/sda1 -f /mnt/sda1/file3 --write-block-size 4096 -v 20
-s 10 --destructive --cache-size 54mb
./stressapptest -d /dev/sda1 -f /mnt/sda1/file3 --write-block-size 4096 -v 20
-s 10 --destructive --cache-size 54mb --write-threshold 200000 use
./stressapptest -d /dev/sda1 -f /mnt/sda1/file3 --write-block-size 4096 -v 20
-s 10 --destructive --cache-size 54mb --write-threshold 200000 usec
./stressapptest -d /dev/sda1 -f /mnt/sda1/file3 --write-block-size 4096 -v 20
-s 10 --destructive --cache-size 54mb --write-threshold 200000
What is the expected output? What do you see instead?
The disk stats & data check should print some number
Stats: Found 0 hardware incidents
Stats: Completed: 72356.00M in 10.00s 7235.11MB/s, with 0 hardware incidents, 0
errors
Stats: Memory Copy: 72356.00M at 7235.34MB/s
Stats: File Copy: 0.00M at 0.00MB/s
Stats: Net Copy: 0.00M at 0.00MB/s
Stats: Data Check: 0.00M at 0.00MB/s
Stats: Invert Data: 0.00M at 0.00MB/s
Stats: Disk: 0.00M at 0.00MB/s
What version of the product are you using? On what operating system?
stressapptest-1.0.6_autoconf on ubuntu
Linux host2 2.6.32-24-generic #39-Ubuntu SMP Wed Jul 28 05:14:15 UTC 2010
x86_64 GNU/Linux
Ubuntu 10.04.1 LTS
Please provide any additional information below.
Please let me know how can we get the disk and data check numbers. I have tried
a lot of options in the 1.0.6 build.
Original issue reported on code.google.com by [email protected]
on 1 Oct 2013 at 7:00
What steps will reproduce the problem?
1. Download latest version of cppcheck from sourceforge (1.57 was used)
2. run cppcheck --enable=all .
What is the expected output? What do you see instead?
What version of the product are you using? On what operating system?
stressapptest-1.0.4_autoconf / Scientific Linux 6.x
Please provide any additional information below.
Multiple issues/warnings are listed but only the error is being entered for
tracking
Original issue reported on code.google.com by [email protected]
on 16 Oct 2012 at 3:06
Attachments:
What steps will reproduce the problem?
1. ./configure
make[2]: Entering directory `/tmp/stressapptest-1.0.0_autoconf/src'
g++ -DHAVE_CONFIG_H -I. -g -O2 -DCHECKOPTS -Wreturn-type -Wunused
-Wuninitialized -Wall -O3 -funroll-all-loops -funroll-loops
-O2 or -O3 ? :)
Original issue reported on code.google.com by dixlor
on 19 Oct 2009 at 10:26
What steps will reproduce the problem?
1. On system that is lacking static version of librt and/or libaio
2. ./configure
What is the expected output? What do you see instead?
Error message is "checking if pthreads is supported... configure: error:
Cannot find a proper pthread library". But pthread library is installed
and there are no problems there.
config.log shows the real problem:
configure:6039: checking if pthreads is supported
configure:6066: gcc -o conftest -g -O2 conftest.c -static -lrt
-pthread -laio >&5
/usr/bin/ld: cannot find -lrt
collect2: ld returned 1 exit status
i.e. librt.a is missing.
After installing static librt, I get same problem with libaio. Gives
misleading pthread message again. Problem is that default libaio is shared
library.
What version of the product are you using? On what operating system?
stressapptest-1.0.3_autoconf
Fedora 12.
Please provide any additional information below.
On Fedora 12, to get a static version of librt.a, you need to install the
"glibc-static" package. For static version of libaio, you need to install
the "libaio-devel" package.
Question: does conftest.c really need to be compiled with the -static flag?
Original issue reported on code.google.com by [email protected]
on 1 Jun 2010 at 6:12
What steps will reproduce the problem?
1.Run with '-W' warm copy on a machine with SSE2
2.No SSE2 workload is run
3.Generic C implementation is run instead
What is the expected output? What do you see instead?
SSE2 workload should be run, if CPU supports it
os layer should detect cpu type and feature set.
Original issue reported on code.google.com by [email protected]
on 12 Mar 2010 at 7:27
What steps will reproduce the problem?
1. typing "stressapptest" as root under commandline
The output is:
Log: Commandline - stressapptest
Stats: SAT revision 1.0.0_autoconf, 32 bit binary
Log: root @ [...] on Mon Oct 19 20:03:44 CEST 2009 from open source release
Log: 1 nodes, 1 cpus.
Log: Defaulting to 1 copy threads
Log: Total 1010 MB. Free 373 MB. Hugepages 0 MB. Targeting 859 MB (84%)
Log: Flooring memory allocation to multiple of 4: 856MB
Process Error: Unsupported system, no error reporting available
Log: Command line option '-A' bypasses this error.
Process Error: Sat::Initialize() failed
Status: FAIL - test encountered procedural errors
Process Error: Fatal issue encountered. See above logs for details.
What version of the product are you using? On what operating system?
Debian Lenny
Original issue reported on code.google.com by [email protected]
on 19 Oct 2009 at 6:13
What steps will reproduce the problem?
1.
2.
3.
What is the expected output? What do you see instead?
What version of the product are you using? On what operating system?
Please provide any additional information below.
Original issue reported on code.google.com by [email protected]
on 7 Mar 2014 at 6:53
What steps will reproduce the problem?
Execute following command
1. aclocal
2. automake --gnu
3. autoconf
4. ./configure./configure --build=x86 --target=armv7a-none-linux-gnueabi
--host=arm-none-linux-gnueabi
What is the expected output? What do you see instead?
Expected: configuration to be completed successfully.
Instead: checking for pthread_barrier... configure: error: in
`/stressapptest-1.0.6_autoconf':
configure: error: cannot run test program while cross compiling
See `config.log' for more details.
What version of the product are you using? On what operating system?
stressapptest-1.0.6
Please provide any additional information below.
Instead of AC_TRY_RUN perform AC_COMPILE_CHECK to check pthread_barrier.
During cross compilation, perform compilation check instead of run check for
pthread_barrier.
Original issue reported on code.google.com by [email protected]
on 6 Mar 2013 at 12:07
Attachments:
please do not hardcode optimization flags by default. configure.ac currently
does:
CXXFLAGS="$CXXFLAGS -O3 -funroll-all-loops -funroll-loops ..."
could that at least be put behind a configure flag ? like
--disable-default-optimizations ?
AC_ARG_ENABLE([default-optimizations],
AS_HELP_STRING([--disable-default-optimizations], [Disable default optimization flag overrides])])
AS_IF([test x"$enable_default_optimizations" != xno], [
CXXFLAGS="$CXXFLAGS -O3 -funroll-all-loops -funroll-loops"
])
Original issue reported on code.google.com by [email protected]
on 27 Nov 2012 at 7:02
the current os.cc code does:
shmaddr = mmap64(NULL, length, PROT_READ | PROT_WRITE,
MAP_SHARED | MAP_NORESERVE | MAP_LOCKED | MAP_POPULATE,
shm_object, NULL);
the last arg is an off_t, not a pointer, so that should be 0, not NULL
otherwise, building on recent glibc/linux yields an ugly warning:
os.cc:473:44: warning: passing NULL to non-pointer argument 6 of ‘void*
mmap64(void*, size_t, int, int, int, __off64_t)’ [-Wconversion-null]
Original issue reported on code.google.com by [email protected]
on 27 Nov 2012 at 7:00
We noticed a failure while running SAT under Linux back to back. test_loop.sh
shows how it can reproduce. The failure mode is that one of the
SAT applications will fail because of an allocation failure. We looked into
this and found that there are a couple of scenarios that will cause the memory
not to be freed to the OS at process exit time frame.
1. libc refuses to free the memory back to the OS on free(). memalign() isn't
guaranteed to work with free(). However, glibc claims it does. posix_memalign()
is suppose to work with free(), but in our tests strace showed that glibc
refused to give back the virtual memory space to the OS.
Once the above behavior occurs the following can happen which cause the memory
not to be released to the OS.
a. Someone reading certain /proc files of SAT's can take a reference on the
OS's mm structure. This can lead to a delayed free of the mm's memory.
b. There can be a thread deep in a run queue that is signalled to die but
hasn't yet. The memory won't be given back to the OS until this last reference
to the mm is dropped.
My proposal is to perform the explicit mmap()/munmap() in SAT so that external
reference counts on the mm within Linux won't cause the memory to be tied up. A
patch is attached to do just that.
Original issue reported on code.google.com by [email protected]
on 16 Sep 2010 at 5:39
Attachments:
What steps will reproduce the problem?
1. Provision a new ubuntu server
2. apt-get update
3. install stressaptest
4. run stressapptest
What is the expected output? What do you see instead?
I expected performance benchmarks. Instead I saw the following:
alan@dev1:~$ lsb_release -a
No LSB modules are available.
Distributor ID: Ubuntu
Description: Ubuntu 9.04
Release: 9.04
Codename: jaunty
alan@dev1:~$ stressapptest
Log: Commandline - stressapptest
Stats: SAT revision 1.0.0_autoconf, 64 bit binary
Log: root @ dev1 on Thu Oct 15 17:52:53 PDT 2009 from open source release
Log: 1 nodes, 8 cpus.
Log: Defaulting to 8 copy threads
Log: Total 7894 MB. Free 7441 MB. Hugepages 0 MB. Targeting 7308 MB (92%)
Process Error: Unsupported system, no error reporting available
Log: Command line option '-A' bypasses this error.
Process Error: Sat::Initialize() failed
Status: FAIL - test encountered procedural errors
Process Error: Fatal issue encountered. See above logs for details.
alan@dev1:~$ uname -a
Linux dev1 2.6.28-15-generic #52-Ubuntu SMP Wed Sep 9 10:48:52 UTC 2009
x86_64 GNU/Linux
alan@dev1:~$
What version of the product are you using? On what operating system?
I am using version " revision 1.0.0_autoconf, 64 bit binary" on an Ubuntu
LInux OS.
Please provide any additional information below.
Pretty straightforward. I installed g++, did ./configure, make, make
install, then run and got the FAIL error.
Original issue reported on code.google.com by [email protected]
on 16 Oct 2009 at 12:58
What steps will reproduce the problem?
1. Install googlestress app 1.0.6 autoconf rpm
2. Run script to start CPU and DIMM stress for specific time
3.
What is the expected output? What do you see instead?
There shouldn't has error message on DIMM
What version of the product are you using? On what operating system?
Intel Avoton C2750 8C CPU
8G Micron 1600MHZ SODIMM *2
Centos6.3
Please provide any additional information below.
Please see the test result on attach.
Original issue reported on code.google.com by [email protected]
on 24 Apr 2014 at 4:33
Attachments:
What steps will reproduce the problem?
1.stressapptest -A -d /dev/sdb
2.Process Error: Unable to seek to sector 4840934 in
What is the expected output? What do you see instead?
Should not error. off64_t should be used for 32bit binary compatibility.
Original issue reported on code.google.com by [email protected]
on 31 Dec 2009 at 1:12
A declarative, efficient, and flexible JavaScript library for building user interfaces.
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
An Open Source Machine Learning Framework for Everyone
The Web framework for perfectionists with deadlines.
A PHP framework for web artisans
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
Some thing interesting about web. New door for the world.
A server is a program made to process requests and deliver data to clients.
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Some thing interesting about visualization, use data art
Some thing interesting about game, make everyone happy.
We are working to build community through open source technology. NB: members must have two-factor auth.
Open source projects and samples from Microsoft.
Google ❤️ Open Source for everyone.
Alibaba Open Source for everyone
Data-Driven Documents codes.
China tencent open source team.