It appears I spoke too soon…

Another day of down time and this time on the new server. Thought things were settling down, but I guess not. According to the email I got from the folks at TekTonic the server was choking on a file creation it was attempting to do which only took them some 5 or 6 hours to figure out and resolve which occurred around 3PM yesterday afternoon. Then it seems the server went through a process of repeated crashes throughout the afternoon so at times it was available and at other times it wasn’t. I didn’t get a chance to check again until 11PM last night and at that point the sites were offline until Elwed managed to restart the VPS. Then I received several reports of corrupted mySQL tables which magically appear to have corrected themselves.

So I just finished downloading a backup of all the sites I host here and now I’m sitting here contemplating my next move. Originally this was going to be my “I’m calling it quits” entry because I was so angry with the sites and myself that I just didn’t want to deal with it anymore, but between Deadscot, ***Dave, and Elwed I’ve been convinced I should try and get some sleep before making decisions I might regret. I’m amazed and pleased by how successful SEB has become, but that same success is causing a lot of problems I don’t know how to fix and/or can’t afford to. The last thing I want to do is stop blogging, but I’m not doing as much of that as I am sitting around waiting for the damned server to behave. It’s amazingly frustrating.

So now I’m sitting here and being whiny about my own incompetence and lack of Bill Gate’s sized fortune to be able to just buy the resources and dedicated tech support team so I can just blog and not worry about the crap behind the scenes. If I’m going to do this I want to do it properly, but that goal is beginning to look more and more remote with each day we’re down for hours at a time.

*Sigh* It’s 12:46AM and I’m due to get up at 6AM so I should probably go and try and get some sleep. I’m tired from worrying about the sites and the stress isn’t going to help me sleep, but I need to try. More later when my head is clearer and my emotions aren’t so high.

11 thoughts on “It appears I spoke too soon…

  1. Yes.  Sleep.  Don’t rush into a decision based on frustration at technical pianos falling on your head.

    I understand that frustration.  Unfortunately, you may need to choose between having all the widgets and customizations you’ve put together with EE or being able to just blog.  Which royally sucks, but—lacking that Gatesean fortune or MacGuyveresque l337 SKILLZ to create a high-powered and reliable server out of a toaster and a pad of Post-It(tm) notes, may be the decision.

    Frankly, given the choice between reading SEB in a stripped down format (hell, in Blogger) and not reading SEB at all, I’ll take the former.  Though, of course, I know you’re not doing it all just for me.  (Though I have to wonder why the hell not.)

    Anyway, hopefully this comment finds you rested and refreshed and all that jazz.  Don’t let the server gremlins grind you down.

  2. Don’t get stressed.  We understand the problems you are having.  This is just a blog after all, though it means more than that to the community you have built here.  Don’t stop doing something that you love.  Take this as a learning experience.  You are a computer support guy after all.  Why not learn a little about Linux? 

    In fact, I have several machines here running Linux.  The slowest is a P3 300 MHZ machine with 64 MB of RAM.  When I get a free moment I am going to install a Linux system from scratch.  I will try to compile and install everything from source.  I love the learning experience, and the more I mess with it, the more confident I feel in Linux.  Sure beats booger sculpturing as a hobby!

  3. In the end it all comes down to: do you still want to do it if you take technical hassles and money constraints out of the equation. Or has it become more of a chore (I sure hope not)?

    If it has become a chore, call timeout for a few months or so. We will come back (though you might want to allow some sort of message email-list).

  4. I was going to say what ingolfson said, so I’ll put it in my own words – don’t go away for good, but if you’re gone for 6 months or a year or whatever, that’s fine smile. Just let us know when you’re back.

  5. believe it or not:

    ive seen mis managed projects with multi million dollar support
    budgets do far far far worse.

    miserable failure to the point of permanant customer dis-satisfaction
    and rufusal to use from everything from portals to crapplications.

    i have a feeling what has made SEB successful is the fact that you know what its like to operate on a budget and cold hard facts versus bullshit, buzzwords and vaporware.

    you actually give a fuck about what you have created.

    if SEBs EE and my SQL are running on a linux box
    you can ask them to ensure that its in runlevel 3 and not 5 in the /etc/inittab.

    this will ensure that X isn’t running and consuming
    9+% of system resources that could be being used for other things.

    often times linux boxes wont swap worth a fuck
    and even though 2G worth of swap is allocated it is most often unused entirely.

    a ‘top’ usually reveals whats up with consumption.

    I have seen latency issues with webservers before
    and have seen a marked improvement noticible to the end user
    (also eliminating downtime due to dns failure) if DNS is
    setup in a stripped down caching only mode and the system relies
    on external servers for its inital info. that way the trip times for
    little retarded backlink numeric resoloution
    and user ip resoloution take less time for frequent hits. little shit adds up

    but then again, im sure i preach to a choir.

    i like SEB, cant say why but i do. even if it has a pic of some poor bastard passed out with a cat
    i still like the fuckin thing anyway.

  6. if SEBs EE and my SQL are running on a linux box
    you can ask them to ensure that its in runlevel 3 and not 5 in the /etc/inittab.

    this will ensure that X isn’t running and consuming
    9+% of system resources that could be being used for other things.

    That assumes that X is installed in the first place. Rest assured that the VPS is uncontaminated by the presence of an X server.

    often times linux boxes wont swap worth a fuck
    and even though 2G worth of swap is allocated it is most often unused entirely.

    That’s a good thing, actually. If an application/web/database server needs to swap, you’ve lost. Always assuming there is a swap device in the first place; the VPS itself doesn’t have one, but it would be amusing to see if this could be worked around by means of a swap file. I’m not familiar enough with the Virtuozzo to know if the amount of RAM guaranteed to the VPS is locked in memory. One would hope so.

    By the way, remember one of the acronyms for the Emacs editor, eight megabytes and constantly swapping? Those were the days…

    a ‘top’ usually reveals whats up with consumption.

    Just about all of the CPU goes to apache (which includes PHP).

    I have seen latency issues with webservers before
    and have seen a marked improvement noticible to the end user
    (also eliminating downtime due to dns failure) if DNS is
    setup in a stripped down caching only mode and the system relies
    on external servers for its inital info. that way the trip times for
    little retarded backlink numeric resoloution
    and user ip resoloution take less time for frequent hits. little shit adds up

    That would matter if apache were configured to perform DNS lookups.

    but then again, im sure i preach to a choir.

    Yes, you do wink

    The simple fact is that EE as deployed for this site is both a memory and a CPU hog. If somebody can suggest remedies, I’m all ear.

  7. apologies for the preachy pants.

    i guess its different if you got you hands on it for sure.

    i am just so used to system administration by braile sometimes i cant turn it off..

    I often find myself thinking stuff like:
    maybe a PPP direct connect to the mysql server would speed things up a bit..?

    I wonder if full duplex is on and the autoneg was turned off vi ethtool on that system?

    is this asshole developer going to use java like he was hired for or is he gonna ask me to setup my own fucking CPAN mirror just so he can run his java wrapped perl crapplet with a jillion module dependancies- again?

    who blabbed like they knew what the fuck they were talking about to mangement and got em all rialed up? i am going to kill them when i find out..

  8. The simple fact is that EE as deployed for this site is both a memory and a CPU hog. If somebody can suggest remedies, I’m all ear.

    dont use BEA weblogic, even though its free with redhat (extras CD) and easy to set up. im pretty sure its compatable with mySQL but i dont know how the users would port over.. its also a memory hog
    killed a sun v120 w/ 1 gig of ram. but that was more of a ssl encrypted portal than a blog. is capable of deploying
    multiple ears and having many domains users. apache compatible.

    might be worth a look

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.