Archive for the ‘Network Operating Systems’ category

« Older Entries

Upgrades

July 29th, 2009

This is going to be a quick one.  We’ve upgraded to eDirectory 8.8, NetWare 6.5 SP8 on all the netware boxes, and introduced the first OES Linux server into the directory.  We’ve also upgraded to GW8 just today. Initial thoughts:

The Good:  The webaccess client is as big of a leap foward as it was from 6.5 to 7.0, maybe even more.  Seamless scrolling past 20 messages and enhanced javascript functions really makes the web client a viable option for people who are in between the power user and I just check my email crowds.  The Windows client seems to be a nice improvement.  I’m really going to like the threaded topic view.  ConsoleOne on Linux with the latest snappins doesn’t actually require you to hold the mouse button down to select attribute screens.

The Bad:  The installer, both NetWare/Windows and especially on Linux still leave a lot to be desired.  The Linux installer is still incapable of properly resizing windows without cutting off buttons even after its initial release with 6.5.  ConsoleOne on Linux is still a slug with the java swing(or is it awt) UI.  The linux agents need the ability to display the live agents screen(the gray screen) without having to unload the daemon and loading it as an application.  I’d prefer to see everything live instead of using the http screen or gwmonitor.

The Ugly:  What’s up with that client icon?

No comments »

Posted in Directory Services, Email, Network Operating Systems, wtf

Why I shouldn’t make any big changes on a Friday afternoon

July 14th, 2008

Work was empty on Friday so I had planned on moving the GWIA from an old 4u Dell Poweredge running NetWare 5.1 to a nice and tidy 1u PE 1950 running SLES 10.  Actually, I ended up not transfering the domain and gateway, rather I created both new.  The domain install and config was quick and painless.  The GWIA install and config was not quick and painless.  There were two “gotchas” which I ran into.  The first gotcha was that the GroupWise system needs to be told which GWIA to send outbound mail to.  Makes perfect sense.  You can find that setting in C1:

Tools/GW System Operations>Internet Addressing>Internet Agent for outbound SMTP/MIME messages:

That got outbound message flowing.  The next step was inbound.  I couldn’t figure out why inbound message processing just wasn’t working.  I was banging my head against the wall until I saw a vague novell forum posting mentioning postfix fighting with the GWIA.  A quick trip into yast’s runlevel module showed that postfix was indeed running.  Shutting that off and restarting the gwia agent allowed inbound messages to flow in.

1pm:  Feeling my oats I decided to move the domain for some large post offices to the cluster.  The domain resided on the other remaining Netware 5.1 server.  This server has some mental issues.  It has been around for longer than I have, logically anyway.  What I mean by that is the server began life as a Netware 3.12 server on different hardware.  It then got a software upgrade to Netware 4.11.  Then a hardware upgrade to a Dell PE server.  Then another software upgrade to NetWare 5.1.  And finally another hardware upgrade to a Dell PE 2650.  This happened over roughly a ten to fifteen year period.  I’ve had plenty of problems with this server.  Disk I/O is completely screwed up.  Writing to or deleting from the volumes takes multiple times over the amount of time it should.  GW domain maintenance operations on the box generally called for a database rebuild to get anything moving.  I had GWAVA 3.6 which vomited tens of thousands of .log files for every piece of mail recieved into the files system, already straining the tired traditional file system based volumes.  In the event of a server abend, an automatic vrepair to the sys volume(I didn’t mention everything was loaded on the sys volume for some unknown reason) would render the server useless for up to 7 hours.  To sum it all up, a painful experience.

At any rate, I shut down the MTA and GWAVA.  I copied the domain to the cluster resource(NetWare).  I installed GWAVA4 on the resource as well.  I got the MTA up and running, but the post offices wouldn’t see a differnet outbound MTP IP address.  I tried rebuilding the domain, restarting the POAs, nothing seemed to work.  I decided to pull the plug due to the time.  I moved the domain back and all of the sudden the POAs were seeing the new domain IP.  Crap!  I moved the domain back to the new location and messages finally started to flow.  I created a new GWAVA4 scanner and set up basic notification digests and called it a day.

Fast forward to today.  I came in and started to do some cleanup work, specifically fixing the cluster resource load scripts to load the MTA up and shut it down.  I noticed that one of the cluster nodes, the node that was running GWAVA and the MTA wiped out early this morning at some point.  So, GWAVA loaded up on the next node, but no MTA.  I fixed that piece and got the busted node back online.  I migrated the resource and the server immediately abended.  It migrated the resource over to the next node, which promptly abended.  Thankfully the resource went comatose before it could completely wipe out the entire cluster.  To get to the point, GWMTAVS.NLM has some serious bug which will drag NetWare to its knees.  I quickly decided to ditch GWAVA on NetWare and put in on the SLES box with the GWIA and just do a GWIA scanner.

Things to remember:

  • Disable postfix on install of SLES
  • Tell the system which GWIA to use
  • C1 sucks, groupwise administration sucks, and file based administration should have been eliminated in GroupWise 7.  Everything should be TCP/IP.
  • GroupWise is very flexible, sometimes to a fault
  • Aside from file sharing duties, Novell’s own GroupWise modules, and Zenworks, don’t use NetWare for anything else important.  It will eat itself to death given the opportunity.  It’s like “Pizza the Hut” from Spaceballs.  Yes, I went there.
  • Don’t make any critical changes on a friday afternoon if you don’t have the time to deal with it over the weekend.  It doesn’t help with stress.

3 comments »

Posted in Email, Network Operating Systems, Rants

OES2 going gold…already?

September 28th, 2007

The public beta for OES2 was out for less than a month and already Novell is moving to the gold release shortly. I haven’t even gotten a spare few minutes to install the beta!

http://www.novell.com/coolblogs/?p=1023

No comments »

Posted in Network Operating Systems

OES2 public beta released

September 10th, 2007

Check it out:

http://www.novell.com/coolblogs/?p=1008

No comments »

Posted in Network Operating Systems

Backups == RAGE

August 9th, 2007

I finally took a leap of faith and rebuilt part of the backup system. When I set it the first time around, I carved the MSA20 drive arrary into two logical RAID devices and in turn two logical drives on the Windows 2003 box it is directly attached to.

We use Backupexec 10 and I set up 36 or so B2D devices for each backup job, per day, per server to be backed up, all full backup jobs. I did this for ease of restore over efficiency. Due to the more limited space restrictions, I had to move B2D devices all over the place to get proper backups. It wasn’t working for me.

I blew the entire drive array away and made one big RAID5 logical drive just to build a bit of fault tolerance into the setup. I also created three B2D policies per server. Every other Friday, a full backup will run. The other Fridays, a separate full backup will run. Monday through Thursday, incremental backups will span the gaps between fulls. This way, I’m hoping to have two full weeks of hot backups. I’m going to add automatic duplicate jobs to copy the full backups to tape so I can have off site backups. We’re not a 24×7 shop, so backups during the weekend are mostly redundant. If space ends up not being an issue, I’ll add incremental backups for Saturday and Sunday as well just to cover everything. If I have a flaw in this design, please feel free to leave a comment tell me what I’m off on and what you might do differently.

Finally, Backupexec for Windows is not NCS(Novell Cluster Services) aware. So because of this, I have to make sure that I know where my services are located. If I have a resource attached to a specific node, I need to keep it there when it comes time for the backup to run otherwise it will fail. This throws a small wrench into everything. I really wanted to keep the nodes and resources completely non dependent on each other, but that’s not going to work out.

No comments »

Posted in Network Operating Systems, Tools

Cluster completely online

July 26th, 2007

I have all three nodes online now. iPrint is running. I’m beginning to start the GroupWise migration. I’ve got a test PO that will migrate properly from node to node. I plan to start moving live POs to the cluster next week. 1 domain will come over after that to a different resource.

During my install process and reading docs, I came across TID 3839149 and saw this in the environment section about iScsi; Poor man’s SAN. A little humor in a technical doc is always appreciated.

No comments »

Posted in Email, Network Operating Systems

Cluster, ZenWorks imaging, air conditioning, OpenSUSE

July 19th, 2007

After lots of trial, error, and reading, I’ve finally gotten our Novell cluster online.

Just to clarify, no, the cluster isn’t called totally_sweet_cluster, although that would be totally sweet. Right now it’s just two nodes(Dell PowerEdge 1950) running iScsi initiator and one server(Dell PowerEdge 2950) running as the iScsi Target. I have one extra 1950 that I still need to install NetWare on and configure. I’m working on getting the proper cluster licenses so I can install the extra node as OES only gives you licenses for two nodes out of the box.  I will play with the service migration tomorrow assuming I have the time.

We’re moving forward with our imaging.  Multicast imaging with ZenWorks is a huge benefit for us.  It even saves the computer name and ip address so we don’t have to manually configure that on each machine when we image a lab.  Once the summer is over, I need to sit down and figure out how to build a Zen image disk, specifically how to upgrade the kernel and initrd in the ISO.  Until then, http://www.zenimaging.info is a great site to grab up to date zen ISOs and PXE related software.

The air conditioner in our server room has been faltering.  It’s a rather large dedicated unit so it should have no problem handling the heat our servers generate.   However, the temperature in the room would swing by a matter of 10 degrees F from 70 degrees to 80 degrees in half an hour or less.  Lately it has been acting far worse.  Twice yesterday the room hit 90+ degrees.  Obviously, that’s not good.  Thanks to our facilities department, the problem was discovered and fixed.  It was sitting around 70 degrees when I left.  I’m hoping the issue is solved.  A hot data center isn’t a good environment, especially when you’re looking to add more servers.

Finally, here’s an obligatory Linux screen shot.  I’ve been trying to use my OpenSUSE install as much as possible lately. ;)

No comments »

Posted in Desktop Management, Network Operating Systems, Tools

Sophos, cluster

July 9th, 2007

We got our Sophos Virus scan licensing info today so I started getting it installed. I’m installing the Enterprise Console on a Server 2003 box. It’s supposed to download updates and drop them on one of the NetWare servers if I’m following the documentation correctly. However, it’s unable to authenticate to my NetWare server. So I have a server call into the company. Hopefully I can get it sorted tomorrow so I can roll this out and check it off my list of things to do.

We received the first couple pieces for the iSCSI cluster. We got the device which will act as the iSCSI Target(SAN). Sitting in another office is also a boxed up Cisco Catalyst 2560G switch which will act as the interconnect device. I’m still waiting on the three iSCSI initiator nodes. However, I probably won’t get to this project until next week, given the work load I need to deal with first.

No comments »

Posted in Network Operating Systems, Tools

Gearing up for summer hell…

June 29th, 2007

I had to curb my Linux fun for a bit, or so I thought, to work on some summer tasks. The first thing I needed to do was get a framework for our imaging in place. Historically, we have used Ghost from back when Binary Labs owned a produced it. We never really got into multi casting for no particularly good reason. We continued to lean on the old Dos ODI based boot disks which ran on IPX. Some might remember the old lsl, 3c90x, ipxodi, vlm commands like they were yesterday. I do too, mostly because they were yesterday for us, literally.

We have had ZenWorks in place for a while. I even installed the Zen 7 imaging bits last summer but never got around to playing with it until recently. In the past week, I have finally been able to create boot CDs and USB thumb drives which boot to a menu driven shell script for all of our normal everyday tasks. I even tested multi casting and it seems to be working well. I think we’ve officially caught up to the year 2000 with our imaging.

I have all the server equipment ordered I need for the summer. It shaping up that our iSCSI cluster is a go. If anyone has any pro tips with Novell Cluster Services, feel free to drop me a line.  You know, anything that I might want to avoid because a year down the road I might be convinced to try and build a time machine out of a Delorean so I can produce the necessary 1.21 gigawatts I need for time travel to come back and slap myself in the back of the head for the dumb decision I’m about to make…in the past…or something.

On an unrelated note, I was accused of being a Novell employee or a plant on a news site/message board this week. I’ve been called plenty of things in my life, that’s the first time I’ve been tagged as a corporate shill. Should I be honored? ;)

No comments »

Posted in Desktop Management, Network Operating Systems, Tools

Cluster

March 29th, 2007

I’ve submitted my hardware requirements, pricing, and licensing. All that’s left is approval(and me crossing my fingers). I’m going with OES NetWare for both the target and initiators at this point. I’m not comfortable putting OES Linux into production due to the issues I’ve read, and OES2 will not be ready.

One node will handle file duties, one will handle GroupWise duties, and the third will handle iPrint. Each service will be able to fail over to a “living” node in the event of failure. If I get the nod for this, I hope to have the hardware in place and everything installed by the start of summer. I’ll migrate services over during the summer then and be able to roll out the afore mentioned iPrint.

I wasn’t able to get to Brainshare, but from what I’ve been reading, OES2 is shaping up to be what finally moves us away from bare metal NetWare for sure, and NetWare in general(bare metal or para). Unfortunately it’s not ready yet for this project, but I’ll be able to migrate OES2 nodes into the cluster I assume down the road, so I’m not stuck(ah the beauty of not having Windows, Exchange, and Active Directory).

No comments »

Posted in Network Operating Systems