Say I have two systems with two hot-swappable drives and have created mirrors
for root, var, and swap across those two drives on each system. If I take a
drive from one system and insert it into the other system, it appears that the
mirror providers on that drive automatically insert themselves
We use iostat to collect statistics of hard drive activity. We've been seeing
some values for the transaction wait column that look suspicious. This is easy
to reproduce by just running iostat repeatedly over a short period of time, as
I show below. Notice the third from last column. From what
This looks like a bug in iostat. 4294967295 == 2 * 32 - -1
It seems that some call returns (unsigned long)-1, e.g. to indicate a failing
system/library call but iostat still prints the result:
Which _precise_ version of FreeBSD are you using?
Well, that's a good question. The particular
You might be able to reduce the iso size some by making a tarball of /var
(using tar -y or tar -z) instead of keeping /var2 as a tree.
Granted you would then need to have tar(1) in the iso, which may cancel out
much of the savings if you would not otherwise have needed it.
Actually, /var is
Can you write a few shell scripts? You'ld need to create a tarball of the
/var contents you need on the box, and explode it onto
/var at boot time -- if you're using auto-var on MFS all the time, you'll
need to set that up to happen on every reboot.
Obviously I can do that. What I was really
I'm probably missing something here, but I'm not sure that's correct. If the
OP wants his own /var, then diskless(8) describes how
/var can be automagically populated (see also /etc/rc.initdiskless). The
nanobsd.sh script (designed with flash drives in mind) uses
this method. I looked into
Not that I know of, unless you use the advantages of mfs then. Full circle,
bud. Now you're asking for necessities of the mfs or mfsroot systems.
I don't want to go there, and don't need to. I came up with a simple way to
populate /var from the original contents so I'm happy. The CD boots,
If FreeBSD cannot write to /tmp or /var on boot, it automatically
creates a MFS filesystems for those mountpoints and mounts them during boot.
You don't need to do anything.
It works as the same readonly compactflash environments out there.
What incidentally does /var get populated with? Our
What incidentally does /var get populated with? Our image has a custom
directory under /var but this did not show up in the MFS versions of this
directory. I can get around this but I wonder what else might not be included?
I found something else that's missing--/var/db/pkg is empty. It looks
You can try to solve the problem by:
-
# echo kern.cam.boot_delay=1 /boot/loader.conf
-
We've put a pretty sizeable delay already directly in the kernel. There is now
a noticeable pause before the mount root step is about to be performed. We
don't see the problem often now, but
One of the distribution sets that comes on a standard release DVD is base.
This includes the core set of binaries as well as the files under /etc and a
few other text files. Running make installworld doesn't collect everything
that's needed. Is there a make option to gather all of the files? I
We have a USB boot stick based cloning process that we're considering porting
to a DVD based media. I'm not sure though that it's possible due to the
restrictions I've seen in the mfsroot environment we'd have to use. For
example, in our USB disk procedure, we create partitions using gpart and
I use make distributionin /usr/src
to create the rest of the /etc files.
That seems to be exactly what I need. Thanks!
___
freebsd-questions@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send
It sounds like http://mfsbsd.vx.sk/ would be helpful to you.
(I havent used it yet due to lack of time but it looks good.)
Hmmm, that just might do the trick. I'll check it out, thanks.
___
freebsd-questions@freebsd.org mailing list
But ... why are you constricting yourself to use mfs_root? I have many times
ran FreeBSD completely from CDrom, which
will give you all 700 (or a DVD, 4.3G) usable space.
I'd be happy to help, if you have questions. but please direct the questions
to the mailing list.
The reason I was
If FreeBSD cannot write to /tmp or /var on boot, it automatically creates a
MFS filesystems for those mountpoints
and mounts them during boot. You don't need to do anything.
It works as the same readonly compactflash environments out there.
D'oh! Man, wish I had known that. I just tried it and
We clone systems from specially prepared USB flash sticks and this all works
well, except that occasionally the flash stick fails to boot. It fails at the
mount root step, saying that it cannot mount the specified root partition. We
use a labeled partition on the disk to make it device
What's the best what to test the status of an Ethernet interface
programmatically? We've been using this code similar to this:
struct ifmediareq ifmr;
memset(ifmr, 0, sizeof(ifmr));
strcpy(ifmr.ifm_name, nfe0);
ioctl(sockfd, SIOCGIFMEDIA, (caddr_t)ifmr)
and then checking the value of
I was going to suggest that you look at the ifconfig(8) source code, but then
I did so myself - it looks like you're doing it pretty much exactly how they
are. I've never noticed ifconfig(8) returning an incorrect value, not to say
it's not possible.
Are you sure that nothing is causing
Is there some trick to know when the power supply sensor readings returned by
ipmitool actually reflects that there is a power supply issue? Our difficulty
is that no one seems to use the same sensor values when it comes to power
supply reporting, and even if there are two power supplies the
Theoretically, doing a straight dd copy of one disk to another and then
swapping in that disk should work. I've done it, with no other tweaking needed.
I've never done it with mixed OS instances on the same disk, or for that matter
with a solid state drive. You'll lose the trailing 12GB of your
] On Behalf Of Peter Steele
Sent: Tuesday, March 30, 2010 8:42 AM
To: freebsd-questions@freebsd.org
Subject: Making sense out of impitool power supply readings
Is there some trick to know when the power supply sensor readings returned by
ipmitool actually reflects that there is a power supply issue? Our
I've moved this to freebsd-hackers...
-Original Message-
From: owner-freebsd-questi...@freebsd.org
[mailto:owner-freebsd-questi...@freebsd.org] On Behalf Of Peter Steele
Sent: Thursday, March 25, 2010 9:54 PM
To: freebsd-questions@freebsd.org
Subject: Very suspicious stack trace
We had
Does FreeBSD 8 support the Dell H700 RAID controller? We've been using a 3Ware
controller but may need to switch to this controller. What we'd like to have is
a command line interface similar to the tw_cli command so we can create RAID
sets on a booted system instead of doing it in the BIOS.
I see that these PERC controllers are all SAS instead of SATA. What kind of
cost differential is there between SAS and SATA disks?
___
freebsd-questions@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To
We had an app crash and the resulting core dump produced a very
suspicious/confusing stack trace:
#0 0x0008011d438c in thr_kill () from /lib/libc.so.7
#1 0x0008012722bb in abort () from /lib/libc.so.7
#2 0x0008011fb70c in malloc_usable_size () from /lib/libc.so.7
#3
Thinking about this some more, a good trick would be to generate a hostname
from the MAC address of the host, since that is guaranteed to be unique.
In fact, this is what we are currently using. Unfortunately I guess I wasn't
entirely clear. I was looking for a facility that actually *assigns*
I've ended up writing a service that runs after netif is complete and sets the
hostname based on the MAC address and also updates /etc/hosts. It does what I
need...
Thanks for all the replies on this...
___
freebsd-questions@freebsd.org mailing list
Is there any facility in FreeBSD for generating a random hostname? We have a
template with a fixed hostname that has to be changed after the template is
closed. It would be useful to have a hostname generated randomly.
___
We have a system controlled through a Java GUI and one of the commands provided
in the GUI is to change the date/time, including the time zone. When the time
zone is changed the FreeBSD system immediately recognizes the change (that is,
the date command from the command line shows the correct
Are there any good instructions for creating a customized bootable .iso image?
I've done the work for creating a bootable USB image, but a .iso is a different
beast in that the boot media is read-only and a virtual disk has to be created
as part of the boot process. Any pointers would be
You can do this with the native make release process for freebsd.
http://www.freebsd.org/doc/en_US.ISO8859-1/articles/releng/release-build.html
Is this what you are looking for?
Hmmm. This might very well do what we need. I'll check it out. Thanks.
What's the proper way to calculate kernel/user/idle time? I know the raw values
come from sysctl kern.cp_time, but these values need to be massaged based on
the number of CPUs and so on. Can someone explain briefly what the algorithm is
calculating the final percentages representing these
They shouldn't need to be massaged. Just sample the values at two intervals,
and your percentages can be calculated by dividing
each delta by the sum of the deltas (since the sum equals the total CPU usage
over the interval, by definition). If you want to
calculate per-cpu usage, use the
We've seen this error sporadically when using the gpart command:
gpart: Cannot get GEOM tree: Cannot allocate memory
What would cause this? It does not happen often but I wouldn't think we should
never see it, not with a simple gpart show command.
We use gpart to create GPT style partitions. For example:
# gpart show ad4
= 34 490234685 ad4 GPT (234G)
34 161 freebsd-boot (8.0K)
50 671088642 freebsd-swap (32G)
67108914 671088643 freebsd-swap (32G)
134217778 104857604
We boot off USB disks all the time without issues. As long as the disk is
listed first in the BIOS and it's a proper FBSD image, it works fine...
-Original Message-
From: owner-freebsd-questi...@freebsd.org
[mailto:owner-freebsd-questi...@freebsd.org] On Behalf Of Fbsd1
Sent: Wednesday,
We have an application where the user can change the date/time via a GUI. One
of the options the user has is to specify that the time is to be synced using
ntp. Our coding worked fine under BSD 7 but since we've moved onto BSD 8 we've
encountered a problem where the command that we initiate
ntpq -pc rv localhost
cat /etc/ntp.conf
My ntp.conf looks like this:
# General Configuration
server 0.us.pool.ntp.org
server 1.us.pool.ntp.org
server 2.us.pool.ntp.org
server 3.us.pool.ntp.org
# Drift file
driftfile /var/db/ntpd.drift
The output from ntpq for the BSD 7 system is this:
Resending this message. For some reason my post never showed up...
-Original Message-
From: Peter Steele
Sent: Wednesday, February 17, 2010 11:51 AM
To: freebsd-questions@freebsd.org
Subject: RE: What would make ntpd hang in BSD 8?
ntpq -pc rv localhost
cat /etc/ntp.conf
My
I suspect I know the problem. The tool I'm building links with a bunch of
other libraries we've developed, which I didn't write. I only modified the
makefile of my
own code. I'm going to have to tweak the makefiles of a dozen different
library modules.
Unfortunately the problem isn't quite as
I've set up a system with gpart and have the swap partition first followed by
root, var, and so on. This works fine but I've seen documents that always have
root first, then swap. Is there any reason that root should be the first
partition or can it follow swap space?
I've used the syntax
1:ad(1,a)/boot/loader
in boot.config to specify the boot device. This doesn't work with GPT
partitions. What's the correct syntax in boot.config for GPT partitions?
I looked at the source code to boot.c and there doesn't seem to be anything
specifically related to GPT
The root partition should always be the 'a' partition, but it doesn't have to
be the first in physical order on the disk (ie. starting at cylinder 0). So
long as partitions don't overlap (with the historical exception of the 'c'
partition, which should cover the whole drive) you can put them
The easiest way would probably be the following.
# SOMEDIR=/path/to/fbsd8buildenv
# mkdir -p ${SOMEDIR}
# cd /path/to/FreeBSD-8.0/src
# make buildworld
# make installworld DESTDIR=${SOMEDIR}
Then adding --sysroot=${SOMEDIR} to all invocations of gcc/ld and/or liberal
use of -I and -L gcc
You could check that the tool is actually linked to the correct libraries with
ldd(1). If all else fails, you could try building a full FreeBSD 8 jail or
chroot.
However running FBSD 8 userland on a 7 kernel is unsupported so I have no idea
if that will actually work well enough to build
I've used the syntax
1:ad(1,a)/boot/loader
in boot.config to specify the boot device. This doesn't work with GPT
partitions. What's the correct syntax in boot.config for GPT partitions?
___
freebsd-questions@freebsd.org mailing list
I have a BSD 7 system with the full BSD 8 sources loaded on it, and we use this
box to build our custom BSD 8 kernel and tools. We do not install the custom
code on the BSD 7 box but simply collect the artifacts as a basis for our
custom BSD 8 image. I have a standalone tool that has previously
Okay, that looks doable. I'll see how this works out. Thanks very much for the
info!
-Original Message-
From: Pieter de Goeje [mailto:pie...@degoeje.nl]
Sent: Saturday, February 06, 2010 5:28 PM
To: freebsd-questions@freebsd.org
Cc: Peter Steele
Subject: Re: What is easiest way to build
Because we have large drives (2TB) we're switching to gpart to partition our
disks. I had previously been using fdisk/bsdlabel and setting up specially
configured partition tables that would work with gmirror. This involved faking
the size of the c partition to make sure there was space for
Assuming I have the name of an interface, what's the easiest way
programmatically to get the status of the interface?
___
freebsd-questions@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any
I use this function to retrieve a list of network interfaces on a system. This
works well enough with one exception--here seems to be multiple entries for
each of the interfaces, and the data in the duplicate entries isn't the same.
What is the proper way to filter the list of structures
Add my name to the list--we get tons of these messages since upgrading to
8.0
-Original Message-
From: owner-freebsd-questi...@freebsd.org
[mailto:owner-freebsd-questi...@freebsd.org] On Behalf Of Bernt Hansson
Sent: Wednesday, December 09, 2009 10:24 AM
To: J.D. Bronson
Cc:
I started out doing this a couple of months ago, and the project has been a
wealth of experience and knowledge, to say the least.
Once you are able to do manage building a release, you can actually build your
own update server and distribute binary updates on your custom kernel.
Hmm,
We have 3U systems with 3Ware raid controllers configured to give us large 11TB
logical drives. The diskinfo command shows this:
# diskinfo -v da1
da1
512 # sectorsize
1133104128 # mediasize in bytes (11T)
23437369344 # mediasize in sectors
In the subject line, you wrote large partition, so I assume you won't want
to boot from from the device, but use it as a big storage area instead.
Correct me if I'm wrong.
For simplicity I didn't include all the details. In fact we need three slices,
one for the OS, one for swap, and the
You cannot use fdisk for this, because fdisk creates MBR partition tables and
these partitions are limited to 2 TB. You have three
options:
1. Use GPT instead of MBR. This is handled by gpt (FreeBSD 7) and gpart
(FreeBSD 8) commands.
We're running 8.0. I'll have to check out gpart.
2. Use a
Excuse me, you're mixing up terminology here. Let me explain:
A SLICE is what Windows calls a DOS primary partition, often just named a
partition.
Yes, I know what a BSD slice is compared to a BSD partition. Considering that
fdisk uses partition interchangeably in cases with slice, I often do
b) go with gpt / gpart, which is okay if FreeBSD will
be the only OS that accesses the disk(s) in question,
as I may assume by your statements.
That's correct; these will be strictly BSD accessible drives.
___
freebsd-questions@freebsd.org mailing
We had been building systems from the binary release plus our custom kernel. In
doing that we just used the base collection from the 8.0 binary release DVD and
copied our kernel files to /boot/kernel. We are now doing both installworld and
installkernel. The issue I'm having is figuring out
By the way, I beleive a good part of /etc is installed/reinstalled by
a make installworld and/or mergemaster.
I should have clarified my scenario better. I am building a BSD image from
scratch, as part of an automated process. I am not updating an existing system.
I had been doing this by
Ah, you _probably_ want to do 'make release' -- I have no experience with this
however to be much more help if that _is_ what you need.
Didn't know about that one. I'll have to check it out--thanks.
___
freebsd-questions@freebsd.org mailing list
I've prepared a USB boot key and it comes up okay until the point it's about to
mount root. It sits hung at the Trying to mount root from: /dev/da0s1a, and a
few lines earlier I see the message root mount waiting for: umass_cam_rescan.
Anyone know what this is about?
I suspect I know the answer to this question but I'll ask it anyway. We're
dealing with some very large disks (11TB raid array) and a newfs operation
takes a significant time. Is there any way to get a volume formatted faster
than the typical newfs does?
Okay, thanks for the reply. I know where I need to go from here...
-Original Message-
From: Maxim Khitrov [mailto:mkhit...@gmail.com]
Sent: Friday, December 04, 2009 12:23 PM
To: Peter Steele
Cc: freebsd-questions@freebsd.org
Subject: Re: Is there the equivalent of a Windows fast format
I have a USB disk that I partitioned with fdisk and bsdlabel. I used the -w
option of bsdlabel to write a standard label. The label itself looks fine:
# bsdlabel da0s1
# /dev/da0s1:
8 partitions:
#size offsetfstype [fsize bsize bps/cpg]
a: 7823576 164.2BSD 2048
Okay, so I've seen this error many times and its cause has always been clear.
In this case I'm stumped. I've got a 3U SuperMicro server with 16 drives hooked
up to two 3Ware controllers. The drives are configured into two logical drives
da0 and da1. I've installed a FreeBSD 8.0 OS on da0 but
/cyl)
Do you want to change our idea of what BIOS thinks ? [n]
Should I answer yes to this query?
-Original Message-
From: Matt Szubrycht [mailto:ma...@bmihosting.com]
Sent: Wednesday, December 02, 2009 5:43 PM
To: Peter Steele
Cc: freebsd-questions@freebsd.org
Subject: Re: missing
I ended up using
ls /dev/ad*|sort -g -k1.8
Not quite as generic as I wanted but it works...
From: Oliver Mahmoudi [mailto:olivermahmo...@gmail.com]
Sent: Sunday, November 29, 2009 10:36 AM
To: Peter Steele
Cc: freebsd-questions@freebsd.org
Subject: Re: Sorting a device list
you can try
Can anyone recommend a quick and dirty way to sort a device list? For example,
if I do this:
ls /dev/ad* | sort
I get something like this:
/dev/ad10
/dev/ad4
/dev/ad6
/dev/ad8
I can add -g, but it doesn't help:
ls /dev/ad* | sort -g
/dev/ad10
/dev/ad4
/dev/ad6
/dev/ad8
I need to skip the
I had tried that. It doesn't work:
# ls -d1 /dev/ad* | sort -n
/dev/ad10
/dev/ad4
/dev/ad6
/dev/ad8
I want the ad10 to appear last...
-Original Message-
From: Giorgos Keramidas [mailto:keram...@ceid.upatras.gr]
Sent: Saturday, November 28, 2009 4:31 PM
To: Peter Steele
Cc: freebsd
I have a script that configures a mirrored file system and have been using this
successfully in 7.0 without any issues. The commands I use to set up the mirror
are:
gmirror label -v -n -b round-robin gm0 ad4s1
gmirror configure -a gm0
This never gave me any trouble in 7.0, but in 8.0 the
What's the best place to pick up the latest 7.0 and 8.0 packages for postgresql
client and server?
___
freebsd-questions@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to
I've installed a BSD 7.0 system and I am getting the error login: Audit record
was not committed when I try to login. What would cause this? I am using a
custom process to do the BSD install onto USB boot disks, but we've been using
it for months and I've never seen this error when we've booted
been built on an 8.0 system with a custom kernel that was built on a
7.0 system?
-Original Message-
From: Norbert Papke [mailto:npa...@acm.org]
Sent: Tuesday, November 17, 2009 10:47 PM
To: freebsd-questions@freebsd.org
Cc: Peter Steele
Subject: Re: ELF library not found error
On November
Message-
From: owner-freebsd-questi...@freebsd.org
[mailto:owner-freebsd-questi...@freebsd.org] On Behalf Of Peter Steele
Sent: Wednesday, November 18, 2009 5:14 AM
To: freebsd-questions@freebsd.org
Subject: RE: ELF library not found error
This was not an upgrade but rather a fresh install. I
Presumably (and I am speculating), the 8.0 packages are not yet finalized and
therefore inconsistent. Perhaps you will have better luck after the official
8.0 Release?
I was thinking the same thing--too much version mismatching going on. I'm going
to take your suggestion though and compile
I did a search for this error and got numerous hits, none which really seemed
to explain my situation. I've installed an 8.0 RC3 system and included Python
2.5, 2.6, and 3.1. The 2.6 version appears to run fine. However, for both 2.5
and 3.1 I get the error:
ELF interpreter
I should have mentioned that this shared library mentioned in the error is in
fact present. For some reason though these apps seem to think it's missing...
-Original Message-
From: owner-freebsd-questi...@freebsd.org
[mailto:owner-freebsd-questi...@freebsd.org] On Behalf Of Peter Steele
When a system has a USB drive present, the system typically names it /dev/da0.
However, if the system has SATA drives hooked to an LSI controller, or if the
system has SCSI drives, the same drive prefix is used as is for the USB drive.
For our purposes, we'd like to separate the USB drive from
If you could elaborate more on the goal you are seeking, other useful
information could come from this list. At the moment, it's just my wild
guessing. :-)
I have a bootable FreeBSD image on a USB stick that clones itself on the target
system's hard drive, creating partitions and other
In this case, labelling the USB stick would be a good chioce. The /etc/fstab
entries then refer to those labels instead of device names (that could change).
You can use generic labels as well as UFS labels here; even a reference to the
UFSID would be possible, as well as independant from da
Labels are an excellent solution in this case.
I've done some quick research on this and it will indeed solve the issue
regarding a generic fstab. I have a related question though. I want to take
this a step further and convert the bootable USB stick into a bootable CD-ROM
image. This is a
I have a FreeBSD image that I install on USB sticks to build new systems. When
the stick boots it automatically clones itself on the system's hard drive,
creating partitions and other configuration parameters that are programmed into
the stick's cloning logic. I want to create a similar
Anyone know if there is a port for OpenWBEM for FreeBSD? I did some searching
but couldn't find anything, although I did find references saying that there is
a FreeBSD port. I couldn't find anything in the usual places though...
___
I want to call the mount() function to perform the same action as running the
following mount command from the command line:
mount -t ufs -o noatime /dev/adXXX /mnt
The man page lists the signature of mount() as
int mount(const char *type, const char *dir, int
See /usr/src/sbin/mount/mount_ufs.c from RELENG_6. Essentially, this
argument should be a ufs_args struct as defined in
/usr/include/ufs/ufs/ufsmount.h.
Okay, got it. Thanks very much!
___
freebsd-questions@freebsd.org mailing list
I have an application running a number of threads. I've had recent instances
where the code below is causing a core dump to occur:
char fstatCmd[200];
char *fstatOut = /tmp/fstat.out;
sprintf(fstatCmd, fstat | grep -v USER | wc -l %s, fstatOut);
rc = system(fstatCmd);
The call is simply
In UNIX it is not safe to perform arbitrary actions after forking a
multi-threaded process. You're basically expected to call exec soon after the
fork, although
you can do certain other work if you are very careful.
The reason for this is that after the fork, only one thread will be running in
We just purchased an HP Proliant DL320 G6, a 1U server with two Broadcom NICs.
When configured as standalone interfaces, the two NICs work fine. However, when
configured as a failover lagg pair, we cannot assign an IP to the lagg0
interface. We are using the following entry in our rc.conf file:
I posted this on the -net list but didn't get any responses. I'm hoping a wider
audience might help.
We're using the lag driver to provide automatic failover in case of a network
outage. The default configuration looks like this:
lagg0: flags=8843UP,BROADCAST,RUNNING,SIMPLEX,MULTICAST
Not really, unless you manually change master. However I believe this also
causes a slight or even bigger network outage. Any reason you're not using
loadbalance algorithm, since it seems to suit you better?
Our resident network guru is quite opposed to using the loadbalancing option
since it
mdconfig for swap space
On Tue, Sep 08, 2009 at 07:52:59PM -0400, Jerry McAllister wrote:
On Tue, Sep 08, 2009 at 04:51:20PM -0500, Peter Steele wrote:
Are there any advantages to using mdconfig and creating a virtual disk for
swap space as opposed to having a designated swap partition
Nowadays having swap twice as RAM is not necessary. If your system wasn't
swapping much in the past you can safely stay with 4G in my opinion...
extending it to 16G would be waste of space :)
I won't bore you with the details but in fact our application *does* require
this much swap space, but
It's easy to *try* the swap files. Then measure the performance.
If the behaviour is really as specific to your custom application as you
indicate, then general advice may not apply either.
In fact, after discussing this with the team, we are going to do exactly that.
We'll allocate an extra
Are there any advantages to using mdconfig and creating a virtual disk for swap
space as opposed to having a designated swap partition? For example, I could do
something like this:
mdconfig -a -t swap -f /var/swap0 -s 4g
swapon -a /dev/md0
to add 4G to the system swap space backed by the file
We have a suite of applications with a Java GUI controlling everything.
One of the actions the user can perform is to set the time zone. We do
this through our Java application and update the /etc/localtime as
required. We also make an API call to tell the JVM that the time zone as
changed, and
Did you try unsetting TZ and then calling tzset()? The man page
implies that doing so will force a reread of /etc/localtime
(http://www.freebsd.org/cgi/man.cgi?query=tzsetsourceid=opensearch):
The tzset() function initializes time conversion information used by
the library routine localtime(3).
I wonder if you'd get more insight by asking the question in -hackers.
Perhaps there are some libc experts listening there.
Well, I still haven't found the magic so I'll try my luck there...
Thanks for the feedback.
___
freebsd-questions@freebsd.org
We've developed a server platform around FreeBSD 7.0 consisting of
several applications written in C and one primary application written in
Java (JDK 1.6). We're seeing cases in some of ouor stress tests where
some threads in our JAVA application appear to get no CPU time for
extended periods of
1 - 100 of 176 matches
Mail list logo