SEO

....has anyone used this tool? wondering how accurate it is on HPUX
11.11i.
I ran it but it lists alot of IPv4 ports listed as *.* and IDLE.

Ht a écrit :
> ...has anyone used this tool? wondering how accurate it is on HPUX
> 11.11i.
> I ran it but it lists alot of IPv4 ports listed as *.* and IDLE.
We use it a lot at work.
If you see wrong result, try recompiling it since lsof go into private
internal kernel structures.

On Nov 25, 6:46 am, Alain <ota1...@hotmail.com> wrote:
> Ht a écrit :> ...has anyone used this tool? wondering how accurate it is on HPUX
> > 11.11i.
> > I ran it but it lists alot of IPv4 ports listed as *.* and IDLE.
>
> We use it a lot at work.
> If you see wrong result, try recompiling it since lsof go into private
> internal kernel structures.

This is bizarre.
So i rebuilt it on the machine i'm running it. When i run it, it says
the application is holding multiple (in the order of 10) file
discriptors to IDLE ports as well as a particular configuration file
and curiously directories as well !!

Does this describe multiple open files or the same one? They point to
the same device and some node name...

COMMAND PID USER FD TYPE DEVICE SIZE/OFF NODE NAME
java 2264 root 208u REG 64,0x5 3079 38214 /opt
(/dev/vg00/lvol5)
java 2264 root 209u REG 64,0x5 3079 38214 /opt
(/dev/vg00/lvol5)
java 2264 root 210u REG 64,0x5 3079 38214 /opt
(/dev/vg00/lvol5)
java 2264 root 211u REG 64,0x5 3079 38214 /opt
(/dev/vg00/lvol5)
java 2264 root 212u REG 64,0x5 3079 38214 /opt
(/dev/vg00/lvol5)
java 2264 root 213u REG 64,0x5 3079 38214 /opt
(/dev/vg00/lvol5)
java 2264 root 214u REG 64,0x5 3079 38214 /opt
(/dev/vg00/lvol5)
java 2264 root 215u REG 64,0x5 3079 38214 /opt
(/dev/vg00/lvol5)

On Nov 25, 8:43 am, Ht <hung...@gmail.com> wrote:
> On Nov 25, 6:46 am, Alain <ota1...@hotmail.com> wrote:
>
> > Ht a écrit :> ...has anyone used this tool? wondering how accurate itis on HPUX
> > > 11.11i.
> > > I ran it but it lists alot of IPv4 ports listed as *.* and IDLE.
>
> > We use it a lot at work.
> > If you see wrong result, try recompiling it since lsof go into private
> > internal kernel structures.
>
> This is bizarre.
> So i rebuilt it on the machine i'm running it. When i run it, it says
> the application is holding multiple (in the order of 10) file
> discriptors to IDLE ports as well as a particular configuration file
> and curiously directories as well !!

Ht a écrit :
> Does this describe multiple open files or the same one? They point to
> the same device and some node name...
>
> COMMAND PID USER FD TYPE DEVICE SIZE/OFF NODE NAME
> java 2264 root 208u REG 64,0x5 3079 38214 /opt
> (/dev/vg00/lvol5)
> java 2264 root 209u REG 64,0x5 3079 38214 /opt
> (/dev/vg00/lvol5)

Same inode: 38214 which is a directory (/opt). In my opinion, nothing to
worry about.
Try "ls -lid /opt" to get /opt'inode.
lsof is useful to find which process is listening at a port; we use it
also to find removed huge file but still hold by a process,...

I see what you mean Alain. They point to a file that has been
deleted...but i thought i'd see one of them not the dozens i'm
seeing...which I'm trying to explain. I actually see the number
increasing over time. I'm almost wondering if I've got a leak in some
file reader.

On Nov 25, 10:15 am, Alain <ota1...@hotmail.com> wrote:
> Ht a écrit :
>
> > Does this describe multiple open files or the same one? They point to
> > the same device and some node name...
>
> > COMMAND PID USER FD TYPE DEVICE SIZE/OFF NODE NAME
> > java 2264 root 208u REG 64,0x5 3079 38214 /opt
> > (/dev/vg00/lvol5)
> > java 2264 root 209u REG 64,0x5 3079 38214 /opt
> > (/dev/vg00/lvol5)
>
> Same inode: 38214 which is a directory (/opt). In my opinion, nothing to
> worry about.
> Try "ls -lid /opt" to get /opt'inode.
> lsof is useful to find which process is listening at a port; we use it
> also to find removed huge file but still hold by a process,...

Any idea how long these idle ports stick around? ...I would have
thought they would not be in the idle state but wait_close or
something terminal.
I'm trying to track down issues where the system runs out of file
descriptors.

java_q3p 29846 root 88u IPv4 0x626ac640 0t0 TCP *:*
(IDLE)
java_q3p 29846 root 89u IPv4 0x6176c0c0 0t0 TCP *:*
(IDLE)
java_q3p 29846 root 90u IPv4 0x5dbfe4c0 0t0 TCP *:*
(IDLE)
java_q3p 29846 root 91u IPv4 0x61294280 0t0 TCP *:*
(IDLE)
java_q3p 29846 root 92u IPv4 0x59c2a200 0t0 TCP *:*
(IDLE)
java_q3p 29846 root 93u IPv4 0x5def3940 0t0 TCP *:*
(IDLE)
java_q3p 29846 root 94u IPv4 0x5dd31e40 0t0 TCP *:*
(IDLE)
java_q3p 29846 root 95u IPv4 0x5b538580 0t0 TCP *:*
(IDLE)
java_q3p 29846 root 96u IPv4 0x61737b00 0t0 TCP *:*
(IDLE)
java_q3p 29846 root 97u IPv4 0x61666e80 0t0 TCP *:*
(IDLE)
java_q3p 29846 root 98u IPv4 0x5cce5540 0t0 TCP *:*
(IDLE)
java_q3p 29846 root 99u IPv4 0x61419940 0t0 TCP *:*
(IDLE)
java_q3p 29846 root 100u IPv4 0x607536c0 0t0 TCP *:*
(IDLE)



On Nov 25, 10:45 am, Ht <hung...@gmail.com> wrote:
> I see what you mean Alain. They point to a file that has been
> deleted...but i thought i'd see one of them not the dozens i'm
> seeing...which I'm trying to explain. I actually see the number
> increasing over time. I'm almost wondering if I've got a leak in some
> file reader.
>
> On Nov 25, 10:15 am, Alain <ota1...@hotmail.com> wrote:
>
> > Ht a écrit :
>
> > > Does this describe multiple open files or the same one? They point to
> > > the same device and some node name...
>
> > > COMMAND PID USER FD TYPE DEVICE SIZE/OFF NODE NAME
> > > java 2264 root 208u REG 64,0x5 3079 38214 /opt
> > > (/dev/vg00/lvol5)
> > > java 2264 root 209u REG 64,0x5 3079 38214 /opt
> > > (/dev/vg00/lvol5)
>
> > Same inode: 38214 which is a directory (/opt). In my opinion, nothing to
> > worry about.
> > Try "ls -lid /opt" to get /opt'inode.
> > lsof is useful to find which process is listening at a port; we use it
> > also to find removed huge file but still hold by a process,...

Ht <hungtru@gmail.com> writes:

> Any idea how long these idle ports stick around? ...I would have
> thought they would not be in the idle state but wait_close or
> something terminal.
> I'm trying to track down issues where the system runs out of file
> descriptors.
>
> java_q3p 29846 root 88u IPv4 0x626ac640 0t0 TCP *:*
> (IDLE)

Java is always a good source for useless syscalls (and a waste of
performance). See "tusc" for details...

Ulrich

Ulrich Windl <Ulrich.Windl@rz.uni-regensburg.de> wrote:
> Ht <hungtru@gmail.com> writes:
> > Any idea how long these idle ports stick around? ...I would have
> > thought they would not be in the idle state but wait_close or
> > something terminal.

If one just calls socket(AF_INET,SOCK_STREAM,...), then a socket could
look like that. If one goes all the way through ESTABLISHED and then
does shutdown() I'm not sure if it ends-up showing unbound as well,
but I would think it would go to IDLE.

> > I'm trying to track down issues where the system runs out of file
> > descriptors.
> >
> > java_q3p 29846 root 88u IPv4 0x626ac640 0t0 TCP *:*
> > (IDLE)

> Java is always a good source for useless syscalls (and a waste of
> performance). See "tusc" for details...

Indeed.

ftp://ftp.cup.hp.com/dist/networking/tools/ among other places.

rick jones
--
denial, anger, bargaining, depression, acceptance, rebirth...
where do you want to be today?
these opinions are mine, all mine; HP might not want them anyway...
feel free to post, OR email to rick.jones2 in hp.com but NOT BOTH...