January 7, 2010
mitch @ 1:12 pm on January 7, 2010
According to an article I found today… this is the case.. I would have to concur to a degree… I would however say that my “low” was more last year than this… As this one has been more of an upswing… slowly but up none the less. I still don’t have the time any more to collect my thoughts and thoroughly research things as much as I would like. Also the reason I haven’t rewritten my blog or changed the way it looks in years. However… this will change… soon. It may not be prettier… but I will have written it all… and… I will be modifying it and adding new as I see my own needs for it change.
Work — No Comments
November 18, 2009
mitch @ 9:07 am on November 18, 2009
So, I installed my first 10GigE card the other week… for a high throughput NAS head for a digital printing company I contract for. For the last couple months I’ve been planning and preparing to swap out their entire network core with new Juniper equipment… I’ve installed a couple stacks of the Juniper EX4200 switches, and we even purchased a 10GigE module for a 300 meter fiber link on both ends of the building. As part of this… we thought we could improve their through put to the NAS server (CentOS 5.4 x86_64 with Samba and SAN disk, and RedHat Cluster Server) by trunking ports. As soon as I had two 1Gbit copper links trunked to these… the speed seemed to be cut in half with samba… SSH seemed to be very fast… some googling and I’m seeing reports that in a bonded nic senario, samba drops to half speed… but NFS and other protocols seem to be able to take advantage of it… Now… I’m sure an upgrade to samba would fix this… but I don’t want to introduce new variables going into their busy season… So, we purchased a 10GbE card for the server so we wouldn’t need to trunk it.
Moved over to this server with the NetXen 10GbE card… pluged it into the Juniper Switch… the link works fine… we move the digital presses over to it and it works fine for a while… after a couple hours running, I start getting errors in the log:
Nov 17 20:34:54 gondor kernel: NETDEV WATCHDOG: eth4: transmit timed out
Nov 17 20:34:54 gondor kernel: netxen_nic eth4: transmit timeout, resetting.
Nov 17 20:34:58 gondor kernel: NETDEV WATCHDOG: eth4: transmit timed out
Nov 17 20:34:58 gondor kernel: netxen_nic eth4: transmit timeout, resetting.
Did some googling, and I’ve found an article to turn off the tcp segmentation offload
so I run:
ethtool -K eth4 tso off
and you can check it with:
ethtool -k eth4
I will update this post if it makes any difference
—- UPDATE —- Nov 30, 2009 3:55 PM MST —-
After looking around… HP wants to use a nx_nic driver instead of the netxen_nic driver that comes with CentOS/RedHat Enterprise by default. So, I’ve compiled the SRPM with my build box, and Installed it on the new server. Version 4.0.516. I then upgraded the firmware on the NIC to the same. I will do some more testing to see if this is the fix I require….
Linux· Open Source· RedHat· Work — No Comments
August 23, 2009
mitch @ 12:15 pm on August 23, 2009
So, I ran into an issue that didn’t seem to be really documented anywhere that I could find anyway.. and it has to deal with a LAG (Link Aggregation Group) between a J6350 (any J-Series would be similar) and a stack of EX4200T switches in the DMZ for a company I’m doing some consulting for. The J6350 has an additional uPIM of 8×1gige ports on it, and I had taken two of them and connected them to two interfaces on the EX switch. I had created three zones on the J6350, one for the Internet drop or Untrust, a second for the Core network or Trust zone, and the third being the DMZ zone. I setup policies permiting the Trust and DMZ zones to access the Untrust zone, and the trust zone is allowed to access the DMZ zone. The DMZ zone is routed on the J6350 for ease of managing access between the different VLANS. This is where the point of confusion came. Zones need the connecting interface added to their zone, and I had been thinking that adding the LAG interface (ae0.0) to the zone would work… It wasn’t. Once I added the virtual vlan interfaces (vlan.100, vlan.101, vlan.103 and vlan.190) everything worked as it should.
Technology· Work — No Comments
February 19, 2009
mitch @ 3:14 pm on February 19, 2009
I ran into a slight issue today with OS X (Leopard), where it would ignore the file permissions forced by my samba file server… In order to maintain users and groups with the correct permissions, I have shares setup like:
[share]
comment = share Volume
path = /srv/SHARE
browseable = yes
writeable = yes
Valid Users =@DOMAIN+"Domain Users"
create mask = 0660
force create mode = 0660
directory mask = 0770
force directory mode = 0770
force group =@DOMAIN+Domain Users
hide files = /__TMP__/.DS_Store/lost+found/
so, I should get files that look like:
-rw-rw---- 1 root root 0 Feb 19 15:07 test1
But the default umask on OS X is 022, so I would get files showing up like:
-rw-r--r-- 1 root root 0 Feb 19 15:07 test2
On the above share… This would prevent other users from changing the files….
Turns out, Apple started pushing for more POSIX compliance and as a result their files would turn up with the permissions assigned to the file on the local system or with their default umask as its created. In a shared world… not always ideal, but I don’t want to have to change the umask on all of the OSX clients…. found an option in samba that I added under the [global] section as follows:
[global]
.....
# Make OS X Leopard play nice with others
unix extensions = off
This fixed my problem.
Apple· Linux· RedHat· Work — No Comments
August 15, 2008
mitch @ 8:46 am on August 15, 2008
My blog will probably become unlisted from Google for this, but it needs to be said.
Polyserve from HP or as its now called “HP Clustered Gateway” is the biggest pile of crap software ever.
I normally don’t like to post about work stuff on my personal blog. However, I’ve noticed that HP seems to be wanting to cover up and hide the fact that this software is the most worthless pile of crap that they have ever purchased and released.
Backing it up, worthless, you’ll get file locking issues that can’t be rectified short of taking snapshots of the volumes. When you have 12TB volumes… those snapshots aren’t THAT small.
Anyway, I really just wanted to start creating some links around to start getting the word out that Polyserve is Junk so it become googleable
Heres a link to a blog with the same info: http://polyserveisjunk.blogspot.com/
Work — No Comments
July 18, 2008
mitch @ 1:40 pm on July 18, 2008
I’m not sure what it is over the last couple months… but lately I don’t really feel into my job. Not sure if its a result of them constantly ignoring what I say or what… All I know is that I normally would have been able to digest something I really wanted to learn, but I find myself unable to concentrate on django, and well anything else I want to learn.
I’m not really sure if its because of work or school… and to be honest… either could be doing it. As I feel that both are ridiculous and neither really feel like they would reward creativity.
Why is it that school (especially IT related) is worthless? Yet everyone requires it? If theres one thing I’m going to learn about me attending school… is that I will not make it a requirement for the people I hire to attend. As just showing an aptitude for learning on your own is self sufficient in my mind.
Life· Work — No Comments
April 13, 2008
mitch @ 8:03 am on April 13, 2008
Since getting the MacBook Pro from my boss, I’ve had some time to play with it. As it turns out, I still only like to use it as much as I like to use my wife’s laptop, mostly just for checking email and web browsing. It just seems slower than my Lenovo, even with its faster procs and same amount of ram… to test I got it dual booting Ubuntu, but even that seems slow. All that I could deal with, however, I really don’t like the keyboard. Or rather the lack of an insert, and a delete and back space keys. Page Up and Page Down also seem to be missing. These keys I use alot, and they’re all very accessible on my Lenovo. So, I’ve decided, I won’t be fully switching from my Lenovo as much as my boss wants me to. The biggest reason so far that I could have used the mac was more for the 1Gb nic it has in it. At work I’ve had to do some testing of the network and I have some scripts and other applications that can fully maximize the throughput of the network, and my current hardware peaks at 100Mbit so, I didn’t see some problems that others in the building were experiencing because of it.
I had the choice to buy a laptop with a 1Gb nic in it when I bought this, it would have cost me more, but to be honest, it was a laptop, when would I EVER actually be able to utilize a gigabit nic. Well… saving to or reading from the hard drive its impossible… but reading and writing straight from ram, you can actually get around to filling a gigabit pipe.
So, I have a reason to switch, I just don’t think I will be. If anything it will give me a reason to buy a new Thinkpad T61 
Apple· Technology· Thoughts· Work — No Comments
April 3, 2008
mitch @ 8:33 am on April 3, 2008
I don’t talk about work much on my blog, mostly because I don’t feel like it is appropriate. However, there are things that I learn while at work that would be appropriate to write about. The purpose of this post is just such a situation. I currently work as a network system consultant and I’ve been doing work onsite for a company for the last 8 months or so. They move large files around, (they’re a printing company) and have also put in a huge investment into gigabit ethernet equipment everywhere. However, most of their buildings are fairly old, and the last time they were cabled was around 1998. Most things are run with your standard Category 5 cable (not even Cat5e). Now the sharp person will know immediately that Cat5 is not supposed to run gigabit. Someone forgot to tell everyone else here, because they put the swithes in place and put gigabit ethernet on the desktops as well… and they negotiated at 1 gigabit…. so it should work right? Wrong! Gigabit ethernet is *very* sensitive to the type of cable being run over it, so much so, that technically you can run it over Cat5e, but if your really shooting for gigabit to the desktop, I wouldn’t go less than Cat6 and making sure you do not run the cable closer than 3 feet from florescent light blasts, and keep it the same distance from all power runs as well. We had a situation in one of the buildings, that they’ve been fighting speed issues for months/years, and in no way did they think it was possible to get faster speeds out of Cat5 cable if I were to drop their speeds to 100Mbit full duplex. Instead of the 1Gbit it was currently negotiating at. Finally we got the ok to force everyone in the building down to 100 Full (actually the switches are HP Procurves, and they have an option to set ports to auto-100 which allows the port to negotiate between full and half duplex and speeds of 10 or 100Mbit, essentially leaving out the option for gigabit). And sure enough speed times increased 3 fold. Imagine that.
Technology· Work — No Comments
November 19, 2007
mitch @ 5:58 pm on November 19, 2007
At my current place of employement, gigabit is deployed everywhere. I’d rather not get into the details as to why it is, but it goes without saying that its not really needed *everywhere*. None the less, I had cacti up and graphing for a couple months, before I thought I’d really like to see cacti graph a fully utilized gigabit port. Well, it turns out that the default configuration of the ucd/net snmp network interface ports on cacti will not graph a gigabit port. It will appear to work, but seems to wrap around at about 120Mbit. The reason is, cacti by default uses 32bit counters for this setting. It has a drop down under the network interfaces when creating graphs on a device, to use 64-bit counters. This is required for gigabit ports (if you want to get a true representation for the graph). I will see if I can grab some screen shots of this when I get back to work.
But overall, I love cacti and even after looking at similar tools, I much prefer cacti over anything else.
Open Source· Technology· Work — No Comments
April 18, 2006
mitch @ 2:35 pm on April 18, 2006
For the last couple years I’ve been wanting to move my career towards network and systems architecture, which I think of it as “Infrastructure Architecture”, and I’m just starting to figure out what type of a background do I need to beable to carry such a title. I get fairly bored when studing routers and switches to be honest, which is why I really haven’t wanted to learn or study them very closely before. However, now that I have a much better understanding of my likes and dislikes, and even some (alteast perceived) understanding of the direction of the computer industry in general. Instead of learning more of the unix’s I’m focusing more on network related technologies and sticking with Linux as the network OS. So my next move will be Cisco certifications. First up will be the CCNA which I’ve studied alot already and having some experience with cisco routers and switches so far its alot easier to remember than when I was studying originally… not sure if its because back then I really wanted my RHCE more than anything else or if I was more into becomming a Sys Admin than a Network Admin/Engineer. Either way, its making this time around alot easier 
Next up after that will be what I’m really interested in which would be the CCDA, or the Cisco Certified Design Associate. Since, like I already said, actually administering routers and switches is somewhat of a bore to me, but networks in general I find very interesting. Designing them is definately intriguing. Hopefully that gets me off on the track I really want, and maybe I’ll even go for the CCDP after that. And somewhere in the middle of all of this will be some VoIP. Anyway, thats the current plan thought I’d lay it all out for myself just because I can.
Life· Technology· Thoughts· Work — No Comments
