I would prefer not
having a configuration option as that is an extra thing to document and
test and some users who need it might not even find out about the option.
I think it would
be better to aim for a maximum of 3500 to 4000 open files, either by increasing
the segment size or by capping the number of segments.
The code seems
to take the (number of identifiers + 1) / 500000 then find the power of
two just bigger than this, so for 1,168,946,725 identifiers this could
create 4096 segments.
I don't know why
the number of segments is chosen to be a power of two.
The code including
giant segments is explained here.
https://bugs.eclipse.org/bugs/show_bug.cgi?id=372548
It was updated
for multithreaded with bug 277422 and a minor update made inbound references
to a class object go in ascending order.
So we need to
understand if the number of segments needs to be a power of two and what
a good upper limit would be and how many other open files there normally
are.
Andrew Johnson
From:
"Kevin
Grigorenko" <kevin.grigorenko@xxxxxxxxxx>
To:
"Memory
Analyzer Dev list" <mat-dev@xxxxxxxxxxx>
Date:
14/07/2021
16:13
Subject:
Re:
[mat-dev] java.io.FileNotFoundException: /mat-work/nikhil/MAT1.11/mat/../../sacbay-m-big77GB.inbound.index3096.log
(Too many open files)
Sent
by: "mat-dev"
<mat-dev-bounces@xxxxxxxxxxx>
ZjQcmQRYFpfptBannerEnd
Andrew, what do you think about a configurable setting for the 500,000
identifier split that Nikhil could increase?
--
Kevin Grigorenko
IBM App Platform SWAT
From: "Nikhil
Dongre" <nikhil.vilas@xxxxxxxxx>
To: "Memory
Analyzer Dev list" <mat-dev@xxxxxxxxxxx>
Date: 07/14/2021
06:10 AM
Subject: [EXTERNAL]
Re: [mat-dev] java.io.FileNotFoundException: /mat-work/nikhil/MAT1.11/mat/../../sacbay-m-big77GB.inbound.index3096.log
(Too many open files)
Sent by: "mat-dev"
<mat-dev-bounces@xxxxxxxxxxx>
HI Kevin, Andrew, That means with the latest version we cannot process
Large Heap dumps? Regards Nikhil Dongre On Tue, Jul 13, 2021 at 11:03 PM
Kevin Grigorenko <kevin.grigorenko@xxxxxxxxxx> wrote: Hi Nikhil,
ZjQcmQRYFpfptBannerStart
This Message Is From an External Sender
This
message came from outside your organization.
ZjQcmQRYFpfptBannerEnd
HI Kevin, Andrew,
That means with the latest version we cannot process Large Heap dumps?
Regards
Nikhil Dongre
On Tue, Jul 13, 2021 at 11:03 PM Kevin Grigorenko <kevin.grigorenko@xxxxxxxxxx>
wrote:
Hi Nikhil,
You might be able to set ulimits if you run a privileged container (which
generally isn't recommended) but otherwise I don't think you can. Example
reproducing your error:
$ docker run --rm fedora sh -c "whoami; ulimit -n; sleep 1; ulimit
-n 10000000"
root
1048576
sh: line 1: ulimit: open files: cannot modify limit: Operation not permitted
As Andrew mentioned, you would need to specify the ulimit when launching
the container; for example, for Docker: --ulimit nofile=90000
See https://docs.docker.com/engine/reference/commandline/run/#set-ulimits-in-container---ulimit
--
Kevin Grigorenko
IBM App Platform SWAT
From: "Nikhil
Dongre" <nikhil.vilas@xxxxxxxxx>
To: "Memory
Analyzer Dev list" <mat-dev@xxxxxxxxxxx>
Date: 07/13/2021
10:00 AM
Subject: [EXTERNAL]
Re: [mat-dev] java.io.FileNotFoundException: /mat-work/nikhil/MAT1.11/mat/../../sacbay-m-big77GB.inbound.index3096.log
(Too many open files)
Sent by: "mat-dev"
<mat-dev-bounces@xxxxxxxxxxx>
Hi Andrew,Kevin, I am not able to modify the ulimit beyond 4096. I tried
the below command and got an exception. [opc@ci-iad-mat-lab-01 ~]$ ulimit
-n 90000 -bash: ulimit: open files: cannot modify limit: Operation not
permitted ZjQcmQRYFpfptBannerStart
This Message Is From an External Sender
This message came from outside your organization.
ZjQcmQRYFpfptBannerEnd
Hi Andrew,Kevin,
I am not able to modify the ulimit beyond 4096.
I tried the below command and got an exception.
[opc@ci-iad-mat-lab-01
~]$ ulimit -n 90000
-bash:
ulimit: open files: cannot modify limit: Operation not permitted
Here
is ulimit command output
[opc@ci-iad-mat-lab-01
mat]$ ulimit -a
core
file size (blocks, -c) 0
data
seg size (kbytes, -d) unlimited
scheduling
priority (-e) 0
file
size (blocks, -f) unlimited
pending
signals (-i) 239754
max
locked memory (kbytes, -l) 64
max
memory size (kbytes, -m) unlimited
open
files
(-n) 4096
pipe
size (512 bytes, -p) 8
POSIX
message queues (bytes, -q) 819200
real-time
priority (-r) 0
stack
size (kbytes, -s) 8192
cpu
time (seconds, -t) unlimited
max
user processes (-u) 4096
virtual
memory (kbytes, -v) unlimited
file
locks
(-x) unlimited
Regards
Nikhil
Dongre
On Mon, Jul 12, 2021 at 11:43 PM Nikhil Dongre <nikhil.vilas@xxxxxxxxx>
wrote:
Hi Andrew,
I think this is a very genuine scenario with 70GB + hprof files with MAT
1.11 as by default Linux os max support 4096 open files.
And in our case 4089 (sacbay-m-big77GB.inbound.index4089.log)
files I can see on my disk.
Once we started using MAT 1.11 it started falling for large hprof files
and in our scenario we are processing upto 100GB files with old MAT 1.8
it was very very slow but rarely failed
And if we think of the above solution, is there any way to calculate the
on the basis of hprof file size? (Reason: upto 50GB hprof files we didn't
face any issue and we may need to process max 100GB file)
"It
might cause performance or memory problems for other people if we reduced
the number and so increased the size of the open temporary index files."
I agree with the above comment.
But is there any way to make this working for Large Heap Dumps.
Regards
Nikhil Dongre
On Fri, Jul 9, 2021 at 1:32 PM Andrew Johnson <andrew_johnson@xxxxxxxxxx>
wrote:
Nikhil,
The code for this is here:
https://git.eclipse.org/r/plugins/gitiles/mat/org.eclipse.mat/+/refs/tags/R_1.12.0/plugins/org.eclipse.mat.parser/src/org/eclipse/mat/parser/index/IndexWriter.java
I think the current code is a reasonable compromise - I think it has one
inbound.indexNNNN.log file for every 500,000 identifiers, so for the maximum
possible 2^31 identifiers it could allocate 4295 files.
There are suggestions on the Internet that you can change the number of
open files when starting a container:
https://docs.docker.com/engine/reference/commandline/run/#set-ulimits-in-container---ulimit
https://mtyurt.net/post/docker-how-to-increase-number-of-open-files-limit.html
E.g. $ docker run --ulimit nofile=1024:1024
--rm debian sh -c "ulimit -n"
It might cause performance or memory problems for other people if we reduced
the number and so increased the size of the open temporary index files.
Regards,
Andrew Johnson
From: "Nikhil
Dongre" <nikhil.vilas@xxxxxxxxx>
To: "Memory
Analyzer Dev list" <mat-dev@xxxxxxxxxxx>
Date: 07/07/2021
17:56
Subject: [EXTERNAL]
Re: [mat-dev] java.io.FileNotFoundException: /mat-work/nikhil/MAT1.11/mat/../../sacbay-m-big77GB.inbound.index3096.log
(Too many open files)
Sent by: "mat-dev"
<mat-dev-bounces@xxxxxxxxxxx>
Hi Andrew,
Thanks for quick work around.
But we have scenarios where we may not be able to change the limit as its
container based solution. Once Hprof files are processed and HTML reports
are generated the container will vanish and throughout we will not have
access to the container.
And another scenario is hprof file size it may vary from 1gb to 100gb so
fixing at OS level may not be the right solution.
Please suggest.
Regards
Nikhil Dongre
On Wed, Jul 7, 2021 at 5:09 PM Andrew Johnson <andrew_johnson@xxxxxxxxxx>
wrote:
Memory Analyzer does create a lot of files at the re-indexing stage as
it helps to build the inbound index. These are deleted once that stage
is over, so you don't normally notice them.
You need to increase the limit on the number of open files for your process.
Something like
ulimit -S -n 4096
would increase the soft limit for the process - if 4096 is too small for
MAT with this dump you will still get the error. There is a hard limit
of the number of open files:
ulimit -H -n
and an ordinary user can't increase the soft limit beyond the hard limit.
If you need more then you need root access or a system administrator to
increase the limits for your userid
The administrator might need to change something like the following.
/etc/security/limits.conf
* - nofile 16384
Once that is done you might need to log in again and run
ulimit -S -n 16384
Andrew Johnson
From: "Nikhil
Dongre" <nikhil.vilas@xxxxxxxxx>
To: "Memory
Analyzer Dev list" <mat-dev@xxxxxxxxxxx>
Date: 07/07/2021
06:51
Subject: Re:
[mat-dev] java.io.FileNotFoundException: /mat-work/nikhil/MAT1.11/mat/../../sacbay-m-big77GB.inbound.index3096.log
(Too many open files)
Sent by: "mat-dev"
<mat-dev-bounces@xxxxxxxxxxx>
Here are some more details
Please find the command ulimit -a
output below :
[opc@ci-iad-mat-lab-01
mat]$ ulimit -a
core
file size (blocks, -c) 0
data
seg size (kbytes, -d) unlimited
scheduling
priority (-e) 0
file
size (blocks, -f) unlimited
pending
signals (-i) 239754
max
locked memory (kbytes, -l) 64
max
memory size (kbytes, -m) unlimited
open
files
(-n) 1024
pipe
size (512 bytes, -p) 8
POSIX
message queues (bytes, -q) 819200
real-time
priority (-r) 0
stack
size (kbytes, -s) 8192
cpu
time (seconds, -t) unlimited
max
user processes (-u) 4096
virtual
memory (kbytes, -v) unlimited
file
locks
(-x) unlimited
[opc@ci-iad-mat-lab-01
mat]$
Hprof
file size is : 77GB
MAT
version : 1.11
And
its failing after this step "Re-indexing outbound index"
[Task:
Parsing /mat-work/nikhil/MAT1.11_Normal/mat/../../sacbay-m-big77GB.hprof
[
Subtask:
Scanning /mat-work/nikhil/MAT1.11_Normal/mat/../../sacbay-m-big77GB.hprof
[....................
[INFO]
Wrote threads call stacks to /mat-work/nikhil/MAT1.11_Normal/mat/../../sacbay-m-big77GB.threads
[....................
[INFO]
Heap /mat-work/nikhil/MAT1.11_Normal/mat/../../sacbay-m-big77GB.hprof contains
1,168,946,725 objects
[....................
Subtask:
Extracting objects from /mat-work/nikhil/MAT1.11_Normal/mat/../../sacbay-m-big77GB.hprof
[................................................................................Task:
Removing unreachable objects
[
Subtask:
Searching for unreachable objects
[
Subtask:
Marking reachable objects
[.
Subtask:
Re-indexing objects
[.
[INFO]
Removed 72,901,319 unreachable objects using 6,692,864,976 bytes
[..
Subtask:
Re-indexing classes
[...
Subtask:
Writing /mat-work/nikhil/MAT1.11_Normal/mat/../../sacbay-m-big77GB.idx.index
[....
Subtask:
Writing /mat-work/nikhil/MAT1.11_Normal/mat/../../sacbay-m-big77GB.o2c.index
[.....
Subtask:
Writing /mat-work/nikhil/MAT1.11_Normal/mat/../../sacbay-m-big77GB.a2s.index
[......
Subtask:
Re-indexing outbound index
[......]
eclipse.buildId=unknown
java.version=1.8.0_292
java.vendor=Red
Hat, Inc.
BootLoader
constants: OS=linux, ARCH=x86_64, WS=gtk, NL=en_US
Framework
arguments: -application org.eclipse.mat.api.parse -DhprofStrictnessWarning=true
../../sacbay-m-big77GB.hprof org.eclipse.mat.api:suspects org.eclipse.mat.api:top_components
org.eclipse.mat.api:overview
Command-line
arguments: -os linux -ws gtk -arch x86_64 -consolelog -application
org.eclipse.mat.api.parse -DhprofStrictnessWarning=true ../../sacbay-m-big77GB.hprof
org.eclipse.mat.api:suspects org.eclipse.mat.api:top_components org.eclipse.mat.api:overview
!ENTRY
org.eclipse.osgi 4 0 2021-07-07 05:32:46.312
!MESSAGE
Application error
!STACK
1
org.eclipse.mat.SnapshotException:
Error opening heap dump 'sacbay-m-big77GB.hprof'. Check the error log for
further details.
at
org.eclipse.mat.parser.internal.SnapshotFactoryImpl.parse(SnapshotFactoryImpl.java:331)
at
org.eclipse.mat.parser.internal.SnapshotFactoryImpl.openSnapshot(SnapshotFactoryImpl.java:167)
at
org.eclipse.mat.snapshot.SnapshotFactory.openSnapshot(SnapshotFactory.java:147)
at
org.eclipse.mat.internal.apps.ParseSnapshotApp.parse(ParseSnapshotApp.java:138)
at
org.eclipse.mat.internal.apps.ParseSnapshotApp.start(ParseSnapshotApp.java:110)
at
org.eclipse.equinox.internal.app.EclipseAppHandle.run(EclipseAppHandle.java:203)
at
org.eclipse.core.runtime.internal.adaptor.EclipseAppLauncher.runApplication(EclipseAppLauncher.java:137)
at
org.eclipse.core.runtime.internal.adaptor.EclipseAppLauncher.start(EclipseAppLauncher.java:107)
at
org.eclipse.core.runtime.adaptor.EclipseStarter.run(EclipseStarter.java:401)
at
org.eclipse.core.runtime.adaptor.EclipseStarter.run(EclipseStarter.java:255)
at
sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
at
sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:62)
at
sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43)
at
java.lang.reflect.Method.invoke(Method.java:498)
at
org.eclipse.equinox.launcher.Main.invokeFramework(Main.java:657)
at
org.eclipse.equinox.launcher.Main.basicRun(Main.java:594)
at
org.eclipse.equinox.launcher.Main.run(Main.java:1447)
at
org.eclipse.equinox.launcher.Main.main(Main.java:1420)
Caused
by: java.io.FileNotFoundException: /mat-work/nikhil/MAT1.11_Normal/mat/../../sacbay-m-big77GB.inbound.index3096.log
(Too many open files)
at
java.io.FileOutputStream.open0(Native Method)
at
java.io.FileOutputStream.open(FileOutputStream.java:270)
at
java.io.FileOutputStream.<init>(FileOutputStream.java:213)
at
java.io.FileOutputStream.<init>(FileOutputStream.java:162)
at
org.eclipse.mat.parser.index.IndexWriter$InboundWriter.log(IndexWriter.java:1237)
at
org.eclipse.mat.parser.internal.GarbageCleaner.clean(GarbageCleaner.java:360)
at
org.eclipse.mat.parser.internal.SnapshotFactoryImpl.parse(SnapshotFactoryImpl.java:282)
...
17 more
On Wed, Jul 7, 2021 at 9:03 AM Nikhil Dongre <nikhil.vilas@xxxxxxxxx>
wrote:
Hi Kevin,
Thanks for the quick response.
Please find the command output below :
[opc@ci-iad-mat-lab-01
mat]$ ulimit -a
core
file size (blocks, -c) 0
data
seg size (kbytes, -d) unlimited
scheduling
priority (-e) 0
file
size (blocks, -f) unlimited
pending
signals (-i) 239754
max
locked memory (kbytes, -l) 64
max
memory size (kbytes, -m) unlimited
open
files
(-n) 1024
pipe
size (512 bytes, -p) 8
POSIX
message queues (bytes, -q) 819200
real-time
priority (-r) 0
stack
size (kbytes, -s) 8192
cpu
time (seconds, -t) unlimited
max
user processes (-u) 4096
virtual
memory (kbytes, -v) unlimited
file
locks
(-x) unlimited
[opc@ci-iad-mat-lab-01
mat]$
Regards
Nikhil
Dongre
On Wed, Jul 7, 2021 at 8:52 AM Kevin Grigorenko <kevin.grigorenko@xxxxxxxxxx>
wrote:
The error "Too many open files" may occur because your file ulimit
is too small, system-wide limits have been exceeded, or there is a bug
in MAT creating too many file descriptors (in decreasing order of probability).
Please start a terminal and send the output of the following command:
$ ulimit -a
--
Kevin Grigorenko
IBM App Platform SWAT
From: "Nikhil
Dongre" <nikhil.vilas@xxxxxxxxx>
To: mat-dev@xxxxxxxxxxx
Date: 07/06/2021
06:01 PM
Subject: [EXTERNAL]
[mat-dev] java.io.FileNotFoundException: /mat-work/nikhil/MAT1.11/mat/../../sacbay-m-big77GB.inbound.index3096.log
(Too many open files)
Sent by: "mat-dev"
<mat-dev-bounces@xxxxxxxxxxx>
Hi Team, When i tried to process process gz hprof file i got this exception
org.eclipse.mat.SnapshotException: Error opening heap dump 'sacbay-m-big77GB.hprof.gz'.
Check the error log for further details. at org.eclipse.mat.parser.internal.SnapshotFactoryImpl.parse(SnapshotFactoryImpl.java:331)
Hi Team,
When i tried to process process gz hprof file i got this exception
org.eclipse.mat.SnapshotException:
Error opening heap dump 'sacbay-m-big77GB.hprof.gz'. Check the error log
for further details.
at
org.eclipse.mat.parser.internal.SnapshotFactoryImpl.parse(SnapshotFactoryImpl.java:331)
at
org.eclipse.mat.parser.internal.SnapshotFactoryImpl.openSnapshot(SnapshotFactoryImpl.java:167)
at
org.eclipse.mat.snapshot.SnapshotFactory.openSnapshot(SnapshotFactory.java:147)
at
org.eclipse.mat.internal.apps.ParseSnapshotApp.parse(ParseSnapshotApp.java:138)
at
org.eclipse.mat.internal.apps.ParseSnapshotApp.start(ParseSnapshotApp.java:110)
at
org.eclipse.equinox.internal.app.EclipseAppHandle.run(EclipseAppHandle.java:203)
at
org.eclipse.core.runtime.internal.adaptor.EclipseAppLauncher.runApplication(EclipseAppLauncher.java:137)
at
org.eclipse.core.runtime.internal.adaptor.EclipseAppLauncher.start(EclipseAppLauncher.java:107)
at
org.eclipse.core.runtime.adaptor.EclipseStarter.run(EclipseStarter.java:401)
at
org.eclipse.core.runtime.adaptor.EclipseStarter.run(EclipseStarter.java:255)
at
sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
at
sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:62)
at
sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43)
at
java.lang.reflect.Method.invoke(Method.java:498)
at
org.eclipse.equinox.launcher.Main.invokeFramework(Main.java:657)
at
org.eclipse.equinox.launcher.Main.basicRun(Main.java:594)
at
org.eclipse.equinox.launcher.Main.run(Main.java:1447)
at
org.eclipse.equinox.launcher.Main.main(Main.java:1420)
Caused
by: java.io.FileNotFoundException: /mat-work/nikhil/MAT1.11/mat/../../sacbay-m-big77GB.inbound.index3096.log
(Too many open files)
at
java.io.FileOutputStream.open0(Native Method)
at
java.io.FileOutputStream.open(FileOutputStream.java:270)
at
java.io.FileOutputStream.<init>(FileOutputStream.java:213)
at
java.io.FileOutputStream.<init>(FileOutputStream.java:162)
at
org.eclipse.mat.parser.index.IndexWriter$InboundWriter.log(IndexWriter.java:1237)
at
org.eclipse.mat.parser.internal.GarbageCleaner.clean(GarbageCleaner.java:360)
at
org.eclipse.mat.parser.internal.SnapshotFactoryImpl.parse(SnapshotFactoryImpl.java:282)
...
17 more
MemoryAnalyzer:
An
error has occurred. See the log file_______________________________________________
mat-dev mailing list
mat-dev@xxxxxxxxxxx
To unsubscribe from this list, visit https://www.eclipse.org/mailman/listinfo/mat-dev
_______________________________________________
mat-dev mailing list
mat-dev@xxxxxxxxxxx
To unsubscribe from this list, visit https://www.eclipse.org/mailman/listinfo/mat-dev_______________________________________________
mat-dev mailing list
mat-dev@xxxxxxxxxxx
To unsubscribe from this list, visit https://www.eclipse.org/mailman/listinfo/mat-dev
Unless stated otherwise above:
IBM United Kingdom Limited - Registered in England and Wales with number
741598.
Registered office: PO Box 41, North Harbour, Portsmouth, Hampshire PO6
3AU
_______________________________________________
mat-dev mailing list
mat-dev@xxxxxxxxxxx
To unsubscribe from this list, visit https://www.eclipse.org/mailman/listinfo/mat-dev_______________________________________________
mat-dev mailing list
mat-dev@xxxxxxxxxxx
To unsubscribe from this list, visit https://www.eclipse.org/mailman/listinfo/mat-dev
Unless stated otherwise above:
IBM United Kingdom Limited - Registered in England and Wales with number
741598.
Registered office: PO Box 41, North Harbour, Portsmouth, Hampshire PO6
3AU
_______________________________________________
mat-dev mailing list
mat-dev@xxxxxxxxxxx
To unsubscribe from this list, visit https://www.eclipse.org/mailman/listinfo/mat-dev_______________________________________________
mat-dev mailing list
mat-dev@xxxxxxxxxxx
To unsubscribe from this list, visit https://www.eclipse.org/mailman/listinfo/mat-dev
_______________________________________________
mat-dev mailing list
mat-dev@xxxxxxxxxxx
To unsubscribe from this list, visit https://www.eclipse.org/mailman/listinfo/mat-dev_______________________________________________
mat-dev mailing list
mat-dev@xxxxxxxxxxx
To unsubscribe from this list, visit https://www.eclipse.org/mailman/listinfo/mat-dev
_______________________________________________
mat-dev mailing list
mat-dev@xxxxxxxxxxx
To unsubscribe from this list, visit https://www.eclipse.org/mailman/listinfo/mat-dev
Unless stated otherwise above:
IBM United Kingdom Limited - Registered in England and Wales with number
741598.
Registered office: PO Box 41, North Harbour, Portsmouth, Hampshire PO6
3AU
_______________________________________________
mat-dev mailing list
mat-dev@xxxxxxxxxxx
To unsubscribe from this list, visit https://www.eclipse.org/mailman/listinfo/mat-dev