June 2006 Archives

Going on vacation

I'm going to be taking three weeks off starting Monday. Three whole weeks. Not here. All in a single block. Yep.

Never taken that much time off in one solid hunk before. But then I get five weeks a year of vacation at this job, so this may happen again. *ahem*. At OldJob I had enough time accumulated that I could take that kind of time off, but I never did since I was a) a lot more indispensible there, and b) couldn't afford to do Neat Things.

As it is, this time I'll be visiting family for the whole three weeks so I'll not be incuring a lot of cost in the way of hotel fees. Which is good for the budget.

Flip side of that is that I'll not be posting here during that time. And when I get back on the 24th I'll be spending a good hunk of the day mucking out my e-mail and getting caught up on projects, not posting to the blog.

See you in a few weeks.

DS updates

We're rolling out eDir right now. This allows some of us to go to SP5 if we wish, and also gives others of us (er, me) more time to engineer an SP5 process that'll fit our environment. Since I manage our main file-serving cluster, I'm using a lot more NetWare products than anyone else. This means I get to apply a lot more post-SP5 updates to make things work.

Plus, I'm going on vacation for three weeks starting Monday. I don't want to hold anyone else back. I'll probably roll SP5 out during the inter-session between the end of Summer Session and Fall Quarter.

Tags: ,

HSM on NetWare part 2

Over at Novell's Cool Blogs Richard Jones has posted a pair of entries related to HSM. I covered the first one last week. This appears to be a series on HSM in general. Richard Jones was the presenter for IO 202: Data Protection Concepts at BrainShare this year, which I attended.

It was in IO 202 that concept of 'Shadow Volumes' was discussed. Shadow Volumes, as they were explained, combines the best features of Distributed File Systems and HSM. He didn't go into very much detail, but in essence I believe Shadow Volumes work by migrating old files to different file systems and use DFS features to allow instant access to that data without all the fun migrate/demigrate problems. Secondly, Shadow Volumes work on last-modified-date rather than last-accessed date; last-accessed date being an increasingly unreliable data-point in this era of Google Desktop, Spotlight, Beagle, and Vista Search.

To give you an idea as to how scanty my information is, these are my notes from the session, as transcribed directly from my note-book:
- Shadow Volume (not shipping) combines DFS & HSM to a degree. (Would be good)
Not a lot. Going to the sessions slides now that I have them, I find more data! And this key phrase: "Like DFS, but at a file-level instead of a Directory level." Ahah! So you take two different file systems, and some DFS-like technology unifies the directory listing of both to present a complete file-system. Then the DFS-like technology in the client is smart enough to be told which actual file-system the file lives on and go retrieve it from there silently. That way you get around the migrate/demigrate lag that traditional HSM involves, and the associated disk-space hit. These days hard-drives are cheaper than tapes on a per-gig basis for certain technologies, as the backup software vendors will tell you at great length, so it makes sense to migrate from your Fibre-Channel SAN to a SATA JBOD for instance.

What's also important about this file-level DFS is that it permits different backup policies on the two physical file-systems. The old/slow file-system can be backed up once a month, where the new/fast file-system gets some attention every day. "What's the big deal," you ask, "that's what the 'Archive' bit is for. Software has been doing that for years. Decades." By splitting your data this way, you can backup a whole volume in complete, rather than a whole volume backup that requires the software to scan the meta data of every single file and directory to check to see if that data needs backing up. That will speed up backups.

Shadow Volumes will probably require a change on the Client to make work. So we're still probably a few years away from a release involving the technology. Though, it could also work if the server itself handles the redirection... hmm. We'll have to see how it pans out.

Tags: ,

Novell gets a new CEO

Novell announced this morning that Messman has been dumped, and Hovsepian put in his place. Plus the CFO is changing.

Ooooh boy.

I didn't like Messman, so I'm not sad to see him go. However, the timing doth suck. The fact it is coming right now is a bad sign all around. This looks to be a move aimed directly at improving the share price in the short term. That can be done by 'cutting costs' or selling off bits of the company. The layoff axe bit pretty hard in November as it was, and I fear what another round would look like.

Unfortunately, the stock analysts don't read me or they'd know that Novell would have been in for a few bad quarters. The stock took a drubbing after the Q3 earnings report. An entirely predictable drubbing. As I've stated before the unambiguous statement Novell made about the future of NetWare is going to completely trash NetWare license renewals over the 12-24 months following BrainShare 2006. This started happening, and the stock tanked. This decline in revenue will happen again in Q4, and again in Q1-07. And to a lesser extent in Q2-07. I predicted this would happen as a result of Novell's statement, so I expected it to happen. And it did.

The Board of Directors moved about 6-9 months too soon. How they're going to 'restore share-holder value' is beyond me at this point. The fundamentals of the product say that they are just plain going to have a few bad quarters, and then revenues will pick up slowly afterwards as the Linux revenue continues to slowly increase. Changing CEO's will restore some confidence in the product, but the post-announcement pickup in stock price still hasn't recovered the value it lost after the Q3 announcement.

But making share-holders happy will have to involve maintaining revenues and cutting costs. Dropping or spinning off unprofitable products is probably going to happen. Yet more lay-offs will happen, very probably in the already diminished NetWare development groups. Novell has a lot of cash, so stock buy-backs to prop up value will almost definitely occur; a move that will diminish available resources for research and development (and mergers).

For the product that Novell sells, this move holds no good news.


Several updates

As the pesimists expected, the Blackboard upgrade isn't going as well as it could have. At least we're on the new hardware. The software update though... there are service calls in the works.

On the 'new grandboss' front, rumor has it that a candidate has been selected. But the mill has not supplied which candidate it is yet. Also, the existing guy will be working in a part time capacity in 'Space Administration. While lofty sounding, this is the department that deals with office shuffles and the domino-effects therein. God-speed, and good hunting. Bring your asbestos long-johns.

One of our three eDir servers will almost definitely be migrated to OES-Linux before fall quarter. This is an older box, so the Linux kernel should provide a few more percentage points of performance. I'd prefer to upgrade the hardware, but the budget isn't there yet.

We're still working on getting a disaster recovery data-center into the basement of Bond Hall. Yep, still plugging.

Novell created the support-forums for Suse Linux Enterprise Desktop yesterday. Perhaps release is soon? Who knows.

HSM on NetWare

Richard Jones over at CoolBlogs recently posted a piece about Hiearchical Storage Management and the OES product line. NetWare has had HSM support since the 1980's, so there is nothing new there. But Linux is another story, and that's still in the pipeline. The article is a good read.

Go read it.

Tags: , ,

That said NetWare

SANs drop the shackles
IP SANs to grow 211% over the next three years


It said NetWare! Specifically:
However, NetWare and Linux environments are also becoming popular for iSCSI, and solutions for small RISC servers running Unix are just starting to emerge.



Removing IPX

IPX routing has been turned off in the router core, and only small grumblings in the corners of the network have resulted. I'll be unbinding IPX from my eDir servers pretty soon. It's already commented out in the autoexec.ncf files so the next reboots will do it anyway.

Passing of an era, but it is time.

Tags: ,
An interesting thing has shown up on the Novell Wiki. Someone has been posting procedures to install Suse Linux Enterprise Desktop 10.

Take a look for yourself.

This is a document in progress, so take it for what you will. But still, interesting!

Tags: ,

Macs and our cluster

One of the projects I've had on my burner for some time is to re-jigger our AFP setup on the cluster to reflect the cluster names. Since we haven't touched the default names, volumes present as the default [node-name].[volume-name] rather than [virtual-server-name].[volume-name] I'd like. This has the effect of making AFP mounts not cluster aware, which is bad.

The presented name is changed in the SYS:\ETC\AFPVOL.CFG file, by the way. It is a well documented file.

The problem I'm facing is that we've managed to work around this problem before Novell supplied a way to fix it. And that workaround is what our two mac-labs use. If I change it suddenly, then those mac labs stop working. Therefore, we need to test. I've tried to get this changed pushed through in the past, but I haven't pushed hard enough.

So, I'm going to push. I'm setting a line in the sand, beyond which the current way will no longer work. Such lines in the sand are, of course, mutable, but my hope is that we really won't come up against any real show-stoppers. Heck, this is how it is SUPPOSED to work. I figure the biggest problems will be retraining, rather than technical. We'll see how it goes.

Tags: ,

SUSE 10.1

I got a new Dell E1505 two months back, and only just recently got it repartitioned enough that I could install a linux on it. I've done this a few times with my elderly circa 11/00 Dell, to mixed results. Slackware and NLD were both hard to shim onto that elderly hardware, and wireless was a problem both times as I ended up having to borrow an Orinoco card each time.

This time around things went smoother. This isn't a full install review like others have been, this is just my impressions of how my own experience went.

Quick Look Good Things
  • Xgl is, of course, gorgeous. Even with the integrated Intel 945GM card. It isn't beefy enough to handle transparency with live updates and wibbly windows at the same time, but it is plenty beefy enough to handle screen translations.
  • The install went rather quick, and the add-on disk was handy.
  • Suspend-to-RAM exists! Yay!
  • The wireless card I have does NOT require ndiswrapper! ipw3945!
  • Basic unsecured wireless connectivity works great.
  • The Power Manager now recognizes the lid-switch and can handle events for it. This was not the case with my old laptop as it was too old to present said switch in a way any Linux power-manager could recognize.
Quick Look Bad Things
  • Compiz is a fragile little thing. Any change at all to screen res, themes, or anything that touches look-n-feel seems to cause it to crash on this laptop.
  • Suspend-to-RAM is so new, it is buggy.
  • Resume-from-Suspend/Hibernation is a chancy proposition at best. Disabling screen-saver locking may help with this.
  • The i945GM graphics does stutter when first copying a screen pic to VRAM for translations, noticable. We'll see if it turns into the 'dripping faucet' kind of noticable.
  • When flipping from Windows to SUSE, it is best to COLD BOOT the system. Something goes weird in the video channel and resolution on the SUSE boot is... funky. This kind of problem is one I've had with Linux and XWindows since, oh heck, that PII 450 system I had a ways back. Getting better, though. Still not there yet.
  • Overall system responsiveness seems slower than when booted to WinXP. That might be improved by going pure-Gnome with no compiz or Xgl.
All in all, this represents a big step forward in Linux usability. Unfortunately, it still isn't 'there' yet. Part of it is that OpenOffice's WordPerfect support isn't all that good, and that's what most of my word-processing documents are in. Part of it is the general fragility of the pieces. And part of it is, I'm sure, a poor choice in mobile hardware to try it out on. But it is what I have, and it works 'pretty good'.

I could EASILLY do BrainShare with this laptop. No worries there at all.

Tags: ,

Novell Open Audio & SAMBA

Novell OpenAudio released a podcast last Friday that is quite good. The meat of the program is an interview with one of the big-wigs over at the Samba project. It actually had some nice technical details for those of us who like such things. It gave clues at where Samba v4 is aimed, as well as some of the technical challenges they face.

I highly recommend this pod-cast. About 43 minutes long.

Tags: ,

More printing stats, weekends

This morning I noticed on our bandwidth tracker, that our internet connection had a busier weekend than any previous this quarter (when there wasn't a Bit Torrent running). Since this week is finals week, I hypothesized that the computer lab printers would be busier than normal.

Date Pages Sat Sun Ratio
6/3-6/4 26501 9468 16994 0.557138
5/20-5/21 15002 4384 10678 0.410564
5/13-5/14 11454 3973 7481 0.531079
5/6-5/7 16963 3797 13150 0.288745
4/29-4/30 14237 3963 10274 0.385731
4/22-4/23 12202 3463 8739 0.39627
4/15-4/16 9670 3115 6555 0.47521
4/8-4/9 13418 4257 9161 0.464687
4/1-4/2 13208 4302 8732 0.492671

So, yeah. It was. 26K pages this weekend, which is 9538 pages more than the next busiest weekend. The other thing that lept out is that the work was more even. The Saturday/Sunday ratio was MUCH closer to even than the past few weekends. And the Saturday numbers are also very interesting, since it shows an output this one Saturday versus any TWO previous Saturdays.

What is also interesting are the weekends of 4/15-16 and 5/13-14. Both of these weekends had markedly reduced page outputs. I wonder why that is? I strongly suspect that the chief driver of page output on weekends is having a paper due on Monday. Perhaps the profs of this University just don't like grading papers in the middle of the month? Hmmmm.

Tags: ,

Printer statistics

It has been a while since I've done page-count stats, so I figured I'd give some. The period the below stats represent is the week 5/21-5/27. A regular week, and the week before the strange tradition of this university known as 'dead week'.

Breakdown by Hour (chart):

Breakdown by Day (chart):

Take a look at the Day chart. I like that smooth curve from Monday through Friday. Monday, Tuesday, and Wednesday have slowly decreasing page-rates, and Thursday and Friday show greatly decreasing page-rates. Miniscule printing is done on Saturday.

As for the Hourly chart, once again the trend of massive printing between 11am and Noon is shown. I'm not sure why that particular hour is popular, but it clearly is, and has clearly been so for some time.

Fun stuff!

Tags: ,

Grandboss candidates

One of the questions I've been trying to ask is their views on two areas:

1: How do you balance in-house applications versus off-the-shelf applications?
2: How do see the third category, open-source?

Candidate 1: He spent too much time talking about managment methodologies. Plus, other people were leading the questioning so I didn't get a word in edgewise. Plus, he seemed to be more interested in the not-yet-open great-grand-boss position. Erm.

Candidate 2: He had a dim view of anything that looked to internal resources for support. Anything developed in-house had to have very solid reasons for being developed in-house, and a firm committment for long-term maintenance of the project. Open-Source was in the same boat, it MUST have very clear ways for long term support. This candidate would not look on pure open-source projects like OpenCMS with favor. Where I suspect something like RedHat and Novell, which provide an OSS solution but solid support behind it, could be more palatable. But sole-developed solutions, such as Microsoft, would still be preferrable over the community-developed solutions on principle.

Candidate 3: He had an open mind. To paraphrase, "I like to go best-of-breed wherever possible, but the budget doesn't always allow that." From this NetWare administrator's point of view, those are welcome words. Of the three, I think we could talk this one into going the Open Enterprise Server route for the next round of file-server upgrades. The other two I suspect would be more in the, "dump it, we're going to Microsoft," camp.