Samba 3.4.3-3.2.1-2306-SUSE-SL11.2 load average: 787.00, 787.03, 788.63

English Network/Internet
1

The conf file is basically equiv to the one that came with a typical install of 11.2 with the exception I added

[erin]

inherit acls = Yes

    write list = erin
    path = /home/erin
    read only = No
    force user = erin
    guest ok = No
    create mask = 0644
    directory mask = 0775

This appears to coincide with the automated Win7 backup erin setup on his share. From what I can tell Win7 creates a zip file in increments of 2GB … thus quite a few thousand. I intend to configure samba properly for a production environment, and to use Bacula for backups, but I’ve been obsessing over getting a USR5610C to work with Hylafax 6.x in class 2.0 mode (with CID enabled).

That being said, how likely is it that I can shut this box down remotely? Last time I tried nothing happened, kill -9, shutdown -r now, init 6, init S, **** did I really just do that? Yes, sigh, drive to the office, where I found the box still responsive, but ultimately had to hard reset. The only thing I didn’t try was shutdown -n (although I don’t think it would have made a difference).

If anyone has any advice as to why this is occurring and how to shut down at this point I’d appreciate it. Below are details that may be pertinent in discerning why this is happening:

I’ve setup /home as ext4 data=writeback,noacl,user_xattr,commit=60,nobarrier,nobh,stripe=384 1 2

There are two kernel paging request bugs in dmesg, only posting the first one:
[198640.313857] BUG: unable to handle kernel paging request at ffffffff0000000a
[198640.313872] IP: <ffffffff8110bd85>] put_compound_page+0x25/0x60
[198640.313884] PGD 1003067 PUD 0
[198640.313890] Oops: 0002 #1] PREEMPT SMP
[198640.313896] last sysfs file: /sys/devices/system/cpu/cpu3/cache/index2/shared_cpu_map
[198640.313904] CPU 0
[198640.313908] Modules linked in: ip6t_LOG xt_tcpudp xt_pkttype ipt_LOG xt_limit snd_pcm_oss snd_mixer_oss snd_seq snd_seq_device edd af_packet ip6t_REJECT nf_conntrack_ipv6 ip6table_raw xt_NOTRACK ipt_REJECT xt_state iptable_raw iptable_filter ip6table_mangle nf_conntrack_netbios_ns nf_conntrack_ipv4 cpufreq_conservative cpufreq_userspace cpufreq_powersave nf_conntrack nf_defrag_ipv4 ip_tables ip6table_filter ip6_tables x_tables acpi_cpufreq fuse loop dm_mod snd_hda_codec_realtek ppdev sr_mod ohci1394 cdrom joydev snd_hda_intel snd_hda_codec snd_hwdep snd_pcm ieee1394 iTCO_wdt serio_raw iTCO_vendor_support parport i2c_i801 r8169 sg snd_timer pcspkr snd snd_page_alloc button intel_agp ext4 jbd2 crc16 fan processor ide_pci_generic ide_core ata_generic aacraid pata_jmicron thermal thermal_sys [last unloaded: preloadtrace]
[198640.314005] Pid: 46, comm: pdflush Not tainted 2.6.31.12-0.1-desktop #1 P35-DS3P
[198640.314012] RIP: 0010:<ffffffff8110bd85>] <ffffffff8110bd85>] put_compound_page+0x25/0x60
[198640.314021] RSP: 0018:ffff88021d3418b0 EFLAGS: 00010206
[198640.314026] RAX: ffffffff00000002 RBX: 000000000000000e RCX: 0000000000000000
[198640.314032] RDX: ffffea0006fedc20 RSI: 000000000000000e RDI: ffffea0007180380
[198640.314038] RBP: ffff88021d3418c0 R08: ffffea00009ad618 R09: 0000000000000002
[198640.314044] R10: ffffea000746bde8 R11: 000000000000000e R12: ffffea0007180380
[198640.314050] R13: 0000000000000000 R14: ffff88021d341a40 R15: ffff88021d341a30
[198640.314057] FS: 0000000000000000(0000) GS:ffff88000904a000(0000) knlGS:0000000000000000
[198640.314063] CS: 0010 DS: 0018 ES: 0018 CR0: 000000008005003b
[198640.314069] CR2: ffffffff0000000a CR3: 0000000213912000 CR4: 00000000000006f0
[198640.314075] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
[198640.314081] DR3: 0000000000000000 DR6: 00000000ffff4ff0 DR7: 0000000000000400
[198640.314087] Process pdflush (pid: 46, threadinfo ffff88021d340000, task ffff88021d33e340)
[198640.314093] Stack:
[198640.314096] ffff88021d3419a0 00000000f7c3286e ffff88021d3419b0 ffffffff8110c51e
[198640.314103] <0> ffff880000012600 0000000000000005 0000000000000000 0000000000000030
[198640.314111] <0> 0000000000000000 0000000000000000 00000000000060ae ffffea0000ac4278
[198640.314122] Call Trace:
[198640.314134] <ffffffff8110c51e>] release_pages+0x26e/0x280
[198640.314143] <ffffffff8110cd10>] __pagevec_release+0x30/0x60
[198640.314164] <ffffffffa00a280c>] ext4_num_dirty_pages+0x23c/0x2c0 [ext4]
[198640.314186] <ffffffffa00a72da>] ext4_da_writepages+0x57a/0x660 [ext4]
[198640.314201] <ffffffff81109cf5>] do_writepages+0x35/0x70
[198640.314211] <ffffffff81172d29>] writeback_single_inode+0x1a9/0x450
[198640.314221] <ffffffff81173430>] generic_sync_sb_inodes+0x1a0/0x560
[198640.314229] <ffffffff81173835>] sync_sb_inodes+0x45/0x60
[198640.314238] <ffffffff81173975>] writeback_inodes+0x65/0x120
[198640.314246] <ffffffff8110917e>] background_writeout+0xbe/0x110
[198640.314255] <ffffffff8110ada9>] __pdflush+0x139/0x270
[198640.314263] <ffffffff8110af68>] pdflush+0x88/0xb0
[198640.314273] <ffffffff8108c1d6>] kthread+0xb6/0xc0
[198640.314281] <ffffffff8100d7ba>] child_rip+0xa/0x20
[198640.314289] Code: e8 d1 f3 f5 ff 90 55 48 89 f8 48 89 e5 48 83 ec 10 65 48 8b 14 25 28 00 00 00 48 89 55 f8 31 d2 48 8b 17 f7 c2 00 00 02 00 75 22 <f0> ff 48 08 0f 94 c2 84 d2 74 06 48 89 c7 ff 50 60 48 8b 45 f8
[198640.314335] RIP <ffffffff8110bd85>] put_compound_page+0x25/0x60
[198640.314343] RSP <ffff88021d3418b0>
[198640.314347] CR2: ffffffff0000000a
[198640.314352] — end trace 554a801ea9e04041 ]—

2

On Sun March 28 2010 01:46 am, Shane R wrote:

>
> The conf file is basically equiv to the one that came with a typical
> install of 11.2 with the exception I added
>
> [erin]
> # inherit acls = Yes
> write list = erin
> path = /home/erin
> read only = No
> force user = erin
> guest ok = No
> create mask = 0644
> directory mask = 0775
>
> This appears to coincide with the automated Win7 backup erin setup on
> his share. From what I can tell Win7 creates a zip file in increments
> of 2GB … thus quite a few thousand. I intend to configure samba
> properly for a production environment, and to use Bacula for backups,
> but I’ve been obsessing over getting a USR5610C to work with Hylafax 6.x
> in class 2.0 mode (with CID enabled).
>
> That being said, how likely is it that I can shut this box down
> remotely? Last time I tried nothing happened, kill -9, shutdown -r now,
> init 6, init S, **** did I really just do that? Yes, sigh, drive to the
> office, where I found the box still responsive, but ultimately had to
> hard reset. The only thing I didn’t try was shutdown -n (although I
> don’t think it would have made a difference).
>
> If anyone has any advice as to why this is occurring and how to shut
> down at this point I’d appreciate it. Below are details that may be
> pertinent in discerning why this is happening:
<snip>
>
ShaneR;

[quote=ShaneR]
I’ve been running into a similar issue, and posted about it a while
back. I’ve changed the logging level per your suggestion and hope that
will shed some insight into what is going on. While I’m not positive it
only applies to windows 7, I speculate that during the Windows 7 backup
process it opens a new smbd to serve it.

'The link to my old thread.
’ (http://tinyurl.com/yasaulj)

If it is of any help. I can provide any additional information if
anyone thinks they can help me troubleshoot this. During setting up
windows 7 backup it will not let me use a drive letter already bound
(ie: Z to a samba share, it wants \samba-server\samba-share with
credentials). I’d assume that would work well as it is requiring
credentials, but who knows. At any rate, the backups have occurred
similar to last week and now I have 1010.0 load average.

17:36:59 up 5 days, 13:00, 5 users, load average: 1010.01, 1010.01,
1010.00

IE: all smbd agents, can’t kill them they are all blocked:

shane 11846 0.0 0.0 81076 2964 ? D Apr06 0:00
/usr/sbin/smbd
-D -s /etc/samba/smb.conf

The only way to get rid of the samba processes (that I have found) is
to reboot the machine without properly shutting it down. I can still
talk to it, and even put it in single user mode, but kill -9, killall
-9, pkill -9, none of those work.

[quote]

Try updating your Samba to version 3.5.1, following the directions in the
sticky at the head of this forum.

As a note: Samba3.5.2 was just released today. I’m not sure it has made it to
the mirrors yet, but if you discover the current update version is 3.5.2
don’t be surprised. It should be OK to use 3.5.2 as well.

–
P. V.
“We’re all in this together, I’m pulling for you.” Red Green

3

On Wed April 7 2010 09:20 pm, PV wrote:

> On Sun March 28 2010 01:46 am, Shane R wrote:
>
<snip>
>>
>> That being said, how likely is it that I can shut this box down
>> remotely? Last time I tried nothing happened, kill -9, shutdown -r now,
>> init 6, init S, **** did I really just do that? Yes, sigh, drive to the
>> office, where I found the box still responsive, but ultimately had to
>> hard reset. The only thing I didn’t try was shutdown -n (although I
>> don’t think it would have made a difference).
>>
>> If anyone has any advice as to why this is occurring and how to shut
>> down at this point I’d appreciate it. Below are details that may be
>> pertinent in discerning why this is happening:
> <snip>
<snip>
>
> Try updating your Samba to version 3.5.1, following the directions in the
> sticky at the head of this forum.
>
> As a note: Samba3.5.2 was just released today. I’m not sure it has made it
to
> the mirrors yet, but if you discover the current update version is 3.5.2
> don’t be surprised. It should be OK to use 3.5.2 as well.
>
PS: Welcome to the forums. Did you execute “shutdown” with root authority?


su
shutdown -r now
(or maybe)
su
shutdown -h now

If you only need to kill smbd you could just use:


su
rcsmb stop
(or maybe)
su
rcsmb restart

–
P. V.
“We’re all in this together, I’m pulling for you.” Red Green

4

shutdown -r does nothing, same thing as changing init to 6:

planner:~ # shutdown -r now
Broadcast message from root (pts/2) (Thu Apr 8 01:42:25 2010):
The system is going down for reboot NOW!

from another console

planner:~ # w
01:45:06 up 5 days, 21:08, 7 users, load average: 1010.93, 1010.42, 1010.15
USER TTY LOGIN@ IDLE JCPU PCPU WHAT
shane :0 Fri04 ?xdm? 2:25m 0.11s /bin/sh /usr/bin/startkde
shane pts/1 Fri04 5days 0.01s 0.01s /bin/bash
shane pts/3 Fri04 5days 0.01s 0.01s /bin/bash
shane pts/4 Fri04 5days 0.00s 5.12s kdeinit4: kded4 [kdeinit]
root pts/2 01:42 2:41 0.14s 0.09s shutdown -r 0 w
root pts/0 17:20 7:30m 0.10s 0.10s -bash
root pts/5 01:42 0.00s 0.05s 0.00s w
planner:~ #

planner:~ # rcsmb stop
Shutting down Samba SMB daemon Warning: daemon not running. done

Rather than listing out all of the processes:
planner:~ # ps aux | grep smbd | tee | wc -l
1005

5

I’m new to openSUSE, but I’ve been running FreeBSD, RHEL, and CentOS boxes for a decade. They all have their quirks, but what bothers me is when I give the signal SIGKILL I expect the kernel to kill that process regardless of whether it is blocked or not.

Currently the smdb processes are blocked: (one of 1000 shown below from a ps):
24034 ? D 0:00 /usr/sbin/smbd -D -s /etc/samba/smb.conf
(Where D is indicating the process is blocked right?)

Doing a top:

top - 01:59:23 up 5 days, 21:22, 8 users, load average: 1011.04, 1010.97, 101
Tasks: 1219 total, 1 running, 1218 sleeping, 0 stopped, 0 zombie
Cpu(s): 0.4%us, 0.3%sy, 0.0%ni, 99.2%id, 0.0%wa, 0.1%hi, 0.0%si, 0.0%st
Mem: 8191768k total, 5967040k used, 2224728k free, 551200k buffers

shutdown -r does nothing, same thing as changing init to 6, init S will drop you into single user mode, not a whole lot of help there.

planner:~ # shutdown -r now
Broadcast message from root (pts/2) (Thu Apr 8 01:42:25 2010):
The system is going down for reboot NOW!

from another console

planner:~ # w
01:45:06 up 5 days, 21:08, 7 users, load average: 1010.93, 1010.42, 1010.15
USER TTY LOGIN@ IDLE JCPU PCPU WHAT
shane :0 Fri04 ?xdm? 2:25m 0.11s /bin/sh /usr/bin/startkde
shane pts/1 Fri04 5days 0.01s 0.01s /bin/bash
shane pts/3 Fri04 5days 0.01s 0.01s /bin/bash
shane pts/4 Fri04 5days 0.00s 5.12s kdeinit4: kded4 [kdeinit]
root pts/2 01:42 2:41 0.14s 0.09s shutdown -r 0 w
root pts/0 17:20 7:30m 0.10s 0.10s -bash
root pts/5 01:42 0.00s 0.05s 0.00s w
planner:~ #

Initiating another session after shutdown being issued:

Apr 8 01:42:25 planner shutdown[1622]: shutting down for system reboot
Apr 8 01:46:33 planner sshd[1671]: Accepted keyboard-interactive/pam for root from 192.168.2.99 port 3805 ssh2

planner:~ # rcsmb stop
Shutting down Samba SMB daemon Warning: daemon not running. done

Rather than listing out all of the processes:
planner:~ # ps aux | grep smbd | tee | wc -l
1005

I will try updating samba tonight. This appears to only pertain to windows 7 boxes, but it sure is annoying to have to reboot the box once a week or so.

6

On Thu April 8 2010 01:56 am, Shane R wrote:

>
> shutdown -r does nothing, same thing as changing init to 6:
>
>> planner:~ # shutdown -r now
>> Broadcast message from root (pts/2) (Thu Apr 8 01:42:25 2010):
>> The system is going down for reboot NOW!
>
> from another console
>
>> planner:~ # w
>> 01:45:06 up 5 days, 21:08, 7 users, load average: 1010.93, 1010.42,
>> 1010.15
>> USER TTY LOGIN@ IDLE JCPU PCPU WHAT
>> shane :0 Fri04 ?xdm? 2:25m 0.11s /bin/sh
>> /usr/bin/startkde
>> shane pts/1 Fri04 5days 0.01s 0.01s /bin/bash
>> shane pts/3 Fri04 5days 0.01s 0.01s /bin/bash
>> shane pts/4 Fri04 5days 0.00s 5.12s kdeinit4: kded4
>> [kdeinit]
>> root pts/2 01:42 2:41 0.14s 0.09s shutdown -r 0 w
>> root pts/0 17:20 7:30m 0.10s 0.10s -bash
>> root pts/5 01:42 0.00s 0.05s 0.00s w
>> planner:~ #
>>
>
> planner:~ # rcsmb stop
> Shutting down Samba SMB daemon Warning: daemon not running.
> done
>
> Rather than listing out all of the processes:
> planner:~ # ps aux | grep smbd | tee | wc -l
> 1005
>
>
ShaneR;

  1. Are you running apparmor? If so are there any messages
    in /var/log/apparmor/ or /var/log/messages relating to the shutdown?

  2. Did you upgrade Samba to version 3.5.2?
    –
    P. V.
    “We’re all in this together, I’m pulling for you.” Red Green

7

I was not running apparmor, just checked and found it had re-enabled itself.

I did not update to the latest samba, the repository shows the x86 version as 64bit and the x86_64 as 32 bit.

This is my first suse box…I’m starting to think it might be in my best interest to start over, and since the two of us have gotten so acquainted over the last two months I’d probably stick with openSUSE. My only other experience is with Red Hat and FreeBSD a long time ago.

Gentoo has always interested me, but I digress. This is intended to be a production file server; it easily saturates the uplink bandwidth, but I didn’t necessarily want a workstation install and it seems to have complicated things (as in there is a samba, and a samba-yast2).

Should I start over? It’s already filled the root partition up a quarter of the way with ****.

Furthermore, if I’m running RAID 5EE ext4 do I want stripe=384 or stripe=256?

planner:/ # df
Filesystem 1K-blocks Used Available Use% Mounted on
/dev/sda2 20641788 4428168 15164980 23% /
udev 4095884 232 4095652 1% /dev
/dev/sda3 7656263764 2728604508 4538743368 38% /home

Starting over isn’t appealing, but it seems like I’ve made a mess and I lean more toward the OCD side. (Like I ran synthetic I/O benchmarks on R0, R10, R50, R5, R5EE, R6)

Thanks for the help and the input

8

When attempting to configure software repositories it requests registration, yet the step after the privacy manager states:
Manual Interaction Required - followed by a URL and another statement about helping me manage systems. I can’t seem to add a system and I did not click that I had a registration code, so I’m assuming this isn’t free? Regardless, it does not allow me past “Manual Interaction Required”. This appears to be something like RHEL / Red Hat provisioning system?

To the point, I haven’t made any changes to the system; I’m still working on getting samba updated – I’ve added the community repository, but have not found an updated version specified. I found it under software.opensuse.org/search (and am installing it), but two questions:

(1) how do I get it to automatically look for updates via zypper or yast?
(2) why are some things like apache2-prefork locked? Upon reviewing the structure of apache it would appear updating wouldn’t affect its configuration (per all the includes).

Thanks,
Shane

9

On Fri April 9 2010 10:16 am, Shane R wrote:

>
> When attempting to configure software repositories it requests
> registration, yet the step after the privacy manager states:
> Manual Interaction Required - followed by a URL and another statement
> about helping me manage systems. I can’t seem to add a system and I did
> not click that I had a registration code, so I’m assuming this isn’t
> free? Regardless, it does not allow me past “Manual Interaction
> Required”. This appears to be something like RHEL / Red Hat
> provisioning system?
>
> To the point, I haven’t made any changes to the system; I’m still
> working on getting samba updated – I’ve added the community repository,
> but have not found an updated version specified. I found it under
> software.opensuse.org/search (and am installing it), but two questions:
>
> (1) how do I get it to automatically look for updates via zypper or
> yast?
> (2) why are some things like apache2-prefork locked? Upon reviewing
> the structure of apache it would appear updating wouldn’t affect its
> configuration (per all the includes).
>
> Thanks,
> Shane
>
>
Shane;
No registration code is needed for openSuse. It is, however, needed for
SLED/SLES. You can check the version of SuSE you are using with:


cat /etc/SuSE-release

As for adding repositories:
Via YaST follow the information given in the sticky at the top of this forum
by Swerdna.

Via zypper:
Follow this howto:
http://forums.opensuse.org/information-new-users/advanced-how-faq-read-only/429157-opensuse-software-installation-hints.html

You can also just download the needed rpm’s from the repository and install
them with the rpm command. If you say save the rpms in a directory
called “Samba3.5.2” they can be installed by:


su
rpm -Uvh <PathToDirectory>/Samba3.5.2/*.rpm

Since * matches all names, make sure the directory only contains the rpm’s you
wish to install. Actually I usually just cd to the desired directory and run
the rpm command. Downloads are easiest from an ftp mirror.

As for apparmor, this probably kept you from shutting down. There could also
be problems if you changed the permissions setting from the default(easy) to
secure or paranoid.

–
P. V.
“We’re all in this together, I’m pulling for you.” Red Green

10

I ended up using the 1-click install, which worked flawlessly. I’m really starting to like this distro.

I’m currently playing with Windows7 and samba 3.5.2-1.1-2345-SUSE-SL11.3; hopefully the issues will be resolved.

I’ve disabled AppArmor again. Is there a particular reason to run it? This is an internal server behind a hardware firewall, VPN used for anything remote.

Thank you for everything, the community support here is great.

11

On Sun April 11 2010 02:56 pm, Shane R wrote:

>
> I ended up using the 1-click install, which worked flawlessly. I’m
> really starting to like this distro.
>
> I’m currently playing with Windows7 and samba
> 3.5.2-1.1-2345-SUSE-SL11.3; hopefully the issues will be resolved.
>
> I’ve disabled AppArmor again. Is there a particular reason to run it?
> This is an internal server behind a hardware firewall, VPN used for
> anything remote.
>
> Thank you for everything, the community support here is great.
>
>
ShaneR;
I’ve never felt it necessary to use apparmor, it always seems to get in the
way and I never spent the time trying to get it configured. Others will most
likely disagree and properly configured it adds a certain level of security.

Hope 3.5.2 works out for you. If you need additional help feel free to post
again. There are many users willing to help.

P. V.
“We’re all in this together, I’m pulling for you.” Red Green