gmane.linux.tagcose.general

UPS problem in low voltage, power fluctuation,PC getting restart..

Eric Shubert — 2013-05-23T17:10:45

This thread is pretty informative. I found it interesting.
http://www.thinkdigit.com/forum/hardware-q/141106-ups-problem-low-voltage-power-fluctuation-pc-getting-restart.html

Linux Kernel /etc/sysctl.conf Security Hardening

Eric Shubert — 2013-05-20T14:21:25

We should be sure that such measures are taken where appropriate.
http://www.cyberciti.biz/faq/linux-kernel-etcsysctl-conf-security-hardening/

Re: READM FPDMA QUEUED errors

2013-05-16T21:35:59

_______________________________________________
list mailing list
list< at >tagcose.com
https://lists.tagcose.com/mailman/listinfo/list

Re: READM FPDMA QUEUED errors

Eric Shubert — 2013-05-16T21:34:45

I wrote too soon. :(
While this may have indeed fixed those errors from being thrown, the 
server's networking on the wan side has seized up again.

At this point I'm suspecting the virtio network driver, mostly because 
I'm not seeing this problem with any of the other servers (which don't 
use that).

I'll post more here as I find the solution.

READM FPDMA QUEUED errors

Eric Shubert — 2013-05-16T19:34:33

One of my servers was throwing messages to the console (dmesg) like:
ata 3.00: exception mask ... frozen
failed command: READ FPDMA QUEUED
status: { DRDY }                (timeout)

These errors showed for both ata 3.00 and 1.00, which correspond to the 
mirrored host drives (software, not udat).

I've seen this error only on this one (of 6) using the Asus E35M1-M Pro 
motherboard. What's a little different with this host are the hard 
drives. They're both WD2500JS-75 models, circa 2007 IIRC. I suspect 
these older drives do not (fully) support some functionality that the 
software is attempting to use.

The apparent solution was found here:
https://bugzilla.redhat.com/show_bug.cgi?id=611350
Adding the "pcie_aspm=off" kernel parameter has indeed eliminated the 
messages.

I haven't the time to investigate further, but wanted to leave this 
breadcrumb here for starters.
HTH.

Re: Parted User's Manual

dwayne — 2013-05-07T00:32:29

thanks. When is a machine valid for production, and when is the machine in
the developmental stage. Maybe the answer is too long for an email and/or
has particular variables that have to be set before the question can be
answered. I think at least there is some general guidelines. Phrased a
different way, how is one determined from the other (production vs.
developmental machine)?


On Mon, May 6, 2013 at 5:22 PM, Eric Shubert <shubes< at >tagcose.com> wrote:

_______________________________________________
list mailing list
list< at >tagcose.com
https://lists.tagcose.com/mailman/listinfo/list

Parted User's Manual

Eric Shubert — 2013-05-07T00:22:45

http://www.gnu.org/software/parted/manual/parted.html
This manual is indispensable when working with gpt partitions.
(note to self)

Fwd: VPN Troubleshooting

Eric Shubert — 2013-05-06T15:35:58

Here are some good VPN/network troubleshooting tips from Kerry on the 
IPCop users list:



Also, what are your results for the following:

pings (packet sizes of 64bytes, 1500bytes, 4096bytes, 1400 bytes, etc. 
with No Fragment options set) from host in source network to host in 
destination network and vice versa.
[you will have to experiment with this, but make sure the Do Not 
Fragment option is set and attempt packets larger than 1500bytes not 
only to destination network, but another host/site out on the Internet 
as well (as a control).  You should find a size that does not give 
response as packet size is too large, then work between packet sizes 
that are successful and those that are not until you find maximum 
successful size.]

traceroute to/from source/destination network as well as Internet 
control host.
(traceroute also has mtu discovery options etc, but sometimes there are 
issues and version differences - read more in manpages)

pings to intermediate routers/devices from above traceroute re

Thanks I'll check it out Re: Router/Bridge Linux Firewall

Dwayne — 2013-05-01T15:01:02


Eric Shubert <shubes< at >tagcose.com> wrote:

Router/Bridge Linux Firewall

Eric Shubert — 2013-05-01T14:53:33

http://www.zeroshell.org/

Zeroshell might be a suitable replacement for IPCop in situations where 
WAN network bonding and QoS are needed. It also has strong 
authentication features. I think it's worth a close look.

Re: Linux fatware? These distros need to slim down | Data Center - InfoWorld

Eric Shubert — 2013-04-10T01:15:19

Very good piece. So true.
Thanks Dwayne.

Linux fatware? These distros need to slim down | DataCenter - InfoWorld

dwayne — 2013-04-10T00:55:28

http://www.infoworld.com/d/data-center/linux-fatware-these-distros-need-slim-down-215757?page=0,0
_______________________________________________
list mailing list
list< at >tagcose.com
https://lists.tagcose.com/mailman/listinfo/list

Re: Nice pice on virtio

2013-04-09T00:44:01

Thanks. We are working on Logstash today and tomorrow.
------Original Message------
From: Eric Shubert
Sender: list-bounces+dwaynehaught=gmail.com< at >tagcose.com
To: Tagcose
ReplyTo: Tagcose
Subject: [tagcose] Nice pice on virtio
Sent: Apr 8, 2013 5:41 PM

http://www.ibm.com/developerworks/linux/library/l-virtio/index.html

Nice pice on virtio

Eric Shubert — 2013-04-09T00:41:50

http://www.ibm.com/developerworks/linux/library/l-virtio/index.html

» Linuxaria – Everything about GNU/Linux and Open source Monitor your bandwidth from the Linux shell

dwayne — 2013-03-31T18:22:36

http://linuxaria.com/article/monitor-your-bandwidth-from-the-linux-shell?lang=en
_______________________________________________
list mailing list
list< at >tagcose.com
https://lists.tagcose.com/mailman/listinfo/list

Backup email with a single command

dwayne — 2013-03-31T17:43:04

http://www.linux-magazine.com/Online/Blogs/Productivity-Sauce/Back-up-Email-with-a-Single-Command
_______________________________________________
list mailing list
list< at >tagcose.com
https://lists.tagcose.com/mailman/listinfo/list

Fwd: Re: Hardware suggestions for flash-based IPCop

Eric Shubert — 2013-03-26T17:49:22

This describes pretty well what the flash install of IPCop does.
We need to make similar considerations for USB installs of PVE.

-------- Original Message --------
Subject: Re: Hardware suggestions for flash-based IPCop
Date: Thu, 31 Jan 2013 18:37:47 -0800
From: David W Studeman <dave-smqDtNBZND1ekjLnd1UXUg< at >public.gmane.org>
To: ipcop-user-5NWGOfrQmneRv+LV9MX5uipxlwaOVQ5f< at >public.gmane.org
Newsgroups: gmane.comp.security.ipcop.user
References: <51017B77.705< at >xxiii.com>

On 1/24/2013 10:20 AM, tar23-nQ5yDzRoN1QAvxtiuMwx3w< at >public.gmane.org wrote:

CF cards are alive and well thanks to Industrial Grade CF cards. They
can be had in sizes as little as 128MB. You can also use DOM's which
plug directly into drive headers or even SD down to microSD with the
proper adapter.

The ideal size for IPCop 2.x is 512MB. When you choose Flash Install
during installation, it makes fixed sized boot and var/log_compressed
partitions. The remainder of the disk is used as root and any larger
size with do nothing as root never gr

Linux Security – Server Hardening | Pinehead.TV

dwayne — 2013-02-15T17:11:47

http://pinehead.tv/linux/linux-security-server-hardening/
_______________________________________________
list mailing list
list< at >tagcose.com
https://lists.tagcose.com/mailman/listinfo/list

Intel brings flash cache acceleration to Linux servers |ZDNet

dwayne — 2013-02-15T17:02:45

http://www.zdnet.com/intel-brings-flash-cache-acceleration-to-linux-servers-7000011206/
_______________________________________________
list mailing list
list< at >tagcose.com
https://lists.tagcose.com/mailman/listinfo/list

Re: list Digest, Vol 57, Issue 1

Eric Shubert — 2013-01-08T18:29:13

Good. To be honest, I'm a bit confused as to why there would be more 
than one gateway. Load balancing or failover scenarios I'm guessing, but 
I haven't done that stuff.



The later. The VM controls the raid array of the data store. That way it 
can easily be split off for scalability (separate backplane).


We need to write this up. I just came across a pitfall yesterday. I'll 
tell you about it on Saturday. The gist of it is be sure to use 
different md# device numbers for each of the arrays on the host. Grub2 
will get them confused otherwise. I use /dev/md7 for the data store.

Also, I *highly* recommend against using raid5. Raid5 will put a 
significant load on the cpu. With the cost of HDDs as low as it is, 
raid5 doesn't make cents any more. If you already have 4 1TB drives that 
you want to use, I recommend using raid-10, which will give you 2TB of 
mirrored storage.


I typically create a store just for ISOs, but that's not really 
necessary. Mostly a hangover from VMware Server. A PVE store can

Re: list Digest, Vol 57, Issue 1

James Dugger — 2013-01-04T06:07:45

Eric,

Thanks for the updated information.  I think I got confused with the dual
gateway stanzas shown in the list of the installation wiki.  Proxmox
wouldn't set both up and when it initially set it up I was left without
access to the hosts webserver.  That is back and I have access to the web
UI again.

A couple of new questions:  I have set up the host on the flash drives as
described. However I have 4 - 1TB SATA HDD's that are currently in a
software RAID 5 that I have in the host system that I need to use to set up
the 1st data store.  They are currently data-less and not assembled in the
current system.

Should these drives be assembled by mdadm at the host level or should they
be re-configured and assembled in Proxmox trough a vm?  If vm inside
Proxmox is the way to go than I have yet to find a proceedure to do this
and need some direction here.

The 32GB flash based RAID is setup and mounted at /opt/vms  and has been
added to Proxmox as a data store as opt.  Do I need to add ISO's for
installation on