what did you learn today?

Status
You're currently viewing only Paladin's posts. Click here to go back to viewing the entire thread.
Not open for further replies.

Paladin

Ars Legatus Legionis
33,629
Subscriptor
I don't really take my phone in the shower/bathroom either. Not like I'm going to reach out of the shower to answer it, so having it close by is not a real attractive ides to me.

After hearing how my company's CFO managed to send both his cell phone and 2 way pager (this was 5+ years ago) into a just used, unflushed toilet, I tend to keep portable electronics out of the kitchen/bathroom areas if I can help it.
 

Paladin

Ars Legatus Legionis
33,629
Subscriptor
That when someone says to you, "Hi, We have 10 class C blocks of IP addresses allocated from you for our dedicated server. Can you swap them out for 10 different class C blocks?", you should ask, "Why?"

And when they answer, "No reason, we just want different ones." Ask "Why?" again.

Eventually you will discover that they have been sending mass amounts of spam carefully targeted against ISPs that typically don't report abuse to SpamCop etc. and are now blacklisted by those same large ISPs and want new IPs to start again.

I said, "No."
 

Paladin

Ars Legatus Legionis
33,629
Subscriptor
I learned this morning that the one time you get a consultant who decides to buy his customer's firewall on his own instead of letting me order and pre-config it for him (at the same price), he will be at a location 90+ miles away from me and magically end up with an ASA 5510 that has been sitting in someone's warehouse for long enough to still have 7.0 code on it and all its network interfaces defaulted to shutdown. He, of course, has no computer with a serial port to console into it.

Why are they paying him exactly?

I also learned, after 2 hours plus trying to get into the firewall over the network, that if you straight up tell him that there is no way to get in without a serial port, he will finally get off his can and walk over to the conference room where they have an old PC with a serial port on it (after telling me 4 times that no PC in the 8-plus story building had one).

Grr.
 

Paladin

Ars Legatus Legionis
33,629
Subscriptor
Originally posted by Shades of Grey:
Fulgan - those 2 options are fairly bizarre. I can see where one might not be totally sure.

1 - install client sw on the server...
2 - install server sw on the client.

That's got to be some kind of (unfunny) joke.
If I was on a client, I would be looking for

3 - install client sw on the client.

Or maybe it's just too early for me -- :p --


Same here. I read it three times before realizing that it was worded completely backwards. Would you like a bowl of cake? Or a slice of soup?
 

Paladin

Ars Legatus Legionis
33,629
Subscriptor
Today I learned (again) that it doesn't matter how many times you try to nicely phrase it to some people. Sometimes you just have to come out and say it, "I think you're asking for something that will ultimately be a bad solution and you should either consider alternative X, or explain to me in real detail why you want me to do the things you have asked me to do."

Usually the answer comes back, "I don't know, we just always done it that way."

Well, since you're changing crap around anyway, isn't this a great opportunity to do it the right way instead of just propagating yesterday's failures?
 

Paladin

Ars Legatus Legionis
33,629
Subscriptor
Originally posted by Fink_Ployd:
It is frigging amazing how much faster thunderbird is than outlook, on multi gigabyte mailboxes. Disentangling ourselves from exchange has been more painful than I'd like but appears to have been well worth it. The real test will be tomorrow.

AD can stay, at least for now, but fuck Exchange right in the ear.

I keep hearing moaning from our sales department about wanting Exchange or an equivalent despite the fact we have perfectly good interaction tools and CRM tools that are all web based and don't require client lockin etc. I'll fight tooth and nail to keep exchange out of our environment as long as I can.
 

Paladin

Ars Legatus Legionis
33,629
Subscriptor
Just put a cap of like 70MB on there and tell them Microsoft rolled it out as a new default with a patch. Then shake your fist at the MS gods on high Olympus.

When someone finally hits the cap you can tell them that Bill Gates did it.

In other words, never ask permission to implement basic sanity rules on the systems you admin. Users will fall in line with almost any reasonable restriction but implementing it after the fact is like taking a cookie from a 3 year old.
 

Paladin

Ars Legatus Legionis
33,629
Subscriptor
Oh, ouch. I was putting one of those in a rack last week and they didn't seem to flimsy (I had the face cover off at the time). I never lift servers that way though, the leverage never works. I always use something like a hand on the lower front left corner and a hand underneath toward the rear right corner. Those are heavy suckers though. Any obvious damage?
 

Paladin

Ars Legatus Legionis
33,629
Subscriptor
It's always like that isn't it? You know if you had had 20 minutes to think about things before hand and an option to just do 'something else' you really would have taken the other option, but when it comes down to it, the pressures of the situation cause you to make what turns out to be a bad choice in the interest of saving time/money/face or whatever.

Sometimes it is a bit depressing how much businesses come to rely on technology that didn't even exist a few years ago and rely on relatively few people to know how to hold it all together.
 

Paladin

Ars Legatus Legionis
33,629
Subscriptor
Originally posted by socoj2:
Originally posted by ronelson:
Brandon, what kind of device did you install that in? Know of any firewalls that do that? Being without a firewall is bad; having it take down a gig+ of traffic is worse.

This is more of an IPS feature a lot of the Proventia and Tipping point do this.

Yeah, generally the firewall is a routing device so if it fails in some way that prevents it from functioning at a normal level (access control for traffic fails or completely brain dead) simply bridging the network won't do anything for you since you have lost NAT and gateway features anyway. There are bridging or transparent firewalls but I think most people would consider it worse to have your stuff suddenly exposed to the world than to go offline. That is why most firewalls of a certain size or larger support pretty good failover instead of 'fail-through' or whatever.
 

Paladin

Ars Legatus Legionis
33,629
Subscriptor
Originally posted by jarends:
Originally posted by Accellion:
mod edit - quoted spam removed

Yeah that was kind of a bone head move. At least he/she didn't try to hide the association. All it would have taken was 'Hey, I work for the company in question. I can help you out if you want. Send me a PM.'

Could be worse, but could be a lot better too.
 

Paladin

Ars Legatus Legionis
33,629
Subscriptor
<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by ronelson:<BR><BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">I learned that it is in fact possible for a server to still be (well, somewhat) functional with a load average of over 3000 </div></BLOCKQUOTE>Wow. The highest I have seen in a "functioning" server was a 3-legged mailserver, circa 1997, serving up 27k students and faculty who all had their client set to check email every 1 minute (ah, POP3, how I do NOT miss you!). It ran in the high 120's regularly and never really choked. In recent times, I notice most anything that gets to 20+ is nearly dead, unless it drops. Once the 15 minute column gets up to 20+, it is time to reboot or deal with REALLY pissy customers. </div></BLOCKQUOTE><BR><BR>Can't remember what the cause of this was, probably multiple users with crons scheduled to do log analysis and backups at the same time plus a bunch of other crap all coinciding to turn the server into a grinding box of gravel:<BR><BR> <IMG SRC="http://pix.shovelfu.com/thread/crazyload.png"> <BR><BR>I was eventually able to SSH in and rescue the box by killing off a few of the bigger disk hog processes. This thing runs CPanel so it routinely has similar issues where users schedule conflicting processes that make it busy at night or whatever but this one was pretty epic so I took a screenshot of the monitoring output.
 

Paladin

Ars Legatus Legionis
33,629
Subscriptor
<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by M. Jones:<BR><BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by akro:<BR>Just becareful when you paste a large amount of data into a terminal window as it won't buffer data normally if the console is to slow for the USB dongle. I have seen a few people bitten by this pasting a switch\router config into a cisco switch where they would lose characters.... ouch... </div></BLOCKQUOTE><BR><BR>That's not kawai <I>at all</I>. </div></BLOCKQUOTE><BR><BR>YEah I've learned to go in increments of about 10 lines at most.<BR><BR>If I can get the minimal config in there and get the device online I just do the rest over ssh. Buffering works properly there.<BR><BR>And I assume you're looking for <I>Kakkoii</I> (cool) not Kawaii (cute).
 

Paladin

Ars Legatus Legionis
33,629
Subscriptor
<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by m0k4t2:<BR>I learned you can no longer buy Cisco switch modules (either Catalyst or Nexus) for the Dell M1000e Blade Chassis.<BR><BR>The page says "call to order" but when we talk to our sales guy, he says they "aren't in the system anymore." </div></BLOCKQUOTE><BR><BR>Awesome, I think we just bought one of those. I think we got the dell modules though because they appeared to be functionally very similar to the cisco ones but cheaper/more ports, etc.
 

Paladin

Ars Legatus Legionis
33,629
Subscriptor
<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by M. Jones:<BR><BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by kragbax:<BR>It looks like things back up, but they really didn't. </div></BLOCKQUOTE><BR>Lack of license is no excuse for a silent fail. There is no excuse for a silent fail when it comes to backups. It's hard enough to get backups right every time without license ninjas waiting to ambush you from the shrubbery.<BR><BR>Today I was reminded how much pain it can be to do peering and NAT debugs when a webserver doesn't respond to ICMP echo request. It was funny in 1996, guys, but now it's a dangerous combination of stubborn and stupid and it's preventing the adults from getting their work done. </div></BLOCKQUOTE><BR><BR>No kidding. Why block ping to something as public as a web server for heaven's sake?! Or anything connected to the internet in a public way for that matter (mail server, whatever)?
 

Paladin

Ars Legatus Legionis
33,629
Subscriptor
Today I learned (yet again) that I will always be the one who gets to deal with problems once they get to the stage where everyone else who was supposed to know what they were doing has given up and can't figure it out.<BR><BR>As usual, my answer in this case is, 'You sort of skipped the step of figuring things out before you spent the money, didn't you?'<BR><BR>Dell m1000e with 2 different kinds of switch modules in it (2 m6220 and 2 m6348) and now that they have it all sitting in a rack, they ask me "How do we make the switch modules all into one big switch so we can do bonding across all 4?"<BR><BR> <img src="http://episteme.meincmagazine.com/groupee_common/emoticons/icon_mad.gif" alt="Mad" width="15" height="15"> <img src="http://episteme.meincmagazine.com/groupee_common/emoticons/icon_eek.gif" alt="Eek" width="15" height="15"><BR><BR>You can't stack dissimilar switch modules dude!
 

Paladin

Ars Legatus Legionis
33,629
Subscriptor
ronelson":3bzf2g4b said:
I learned that it's possible for a Linux server to get confused enough about free space that "df" actually returns a negative value for used space. Extremely rare, but possible.
It is possible that it was not confused. Nokia's IPSO (BSD based) uses a ramdrive to extend partitions somehow, so you can write a few extra meg without the server keeling over. Helpful in theory, but in reality, when has a resource hog that fills the drive ever stopped when you just gave it a few extra meg?

You can also run into a situation where small files eat up huge block sizes or the number of inodes and some versions of df will report the usage differently than du (df -h shows 70G used and 20G free, du -h shows 10G used, you add another 2G of small files and the drive is filled). Lots of fun things you can do to fuck up a filesystem :)

Note: NEVER reboot when you are using 101% of the drive. EVER.

Many distro's actually report drive space as less than it really is because they reserve a bit which can only be written to by 'root'. Of course a lot of things write logs as root so that can fill more than 100% of the drive. So it is actually no that uncommon to overfill the drive that way if you let your logs go unhandled.

More info.
 

Paladin

Ars Legatus Legionis
33,629
Subscriptor
ronelson":1kdtlm85 said:
These are excellent explanations for why free space can be negative, the reserved space in particular is a pretty common case. My post, though, was about used space showing as negative, which is just plain weird
Duh, reading comprehension is gud!

I have seen that on Solaris, something somewhere got fragged and it took a reboot to clear it. In some ways, I wanted to find out what did it, in other ways I was afraid, very afraid :)

Wow, yeah I completely missed that one. Doh. Negative used space huh? I think I might have seen that once but I attributed it to a messed up partition that suffered a very intense FSCK on reboot. After that it was cool. Un/misallocated inodes or something?
 

Paladin

Ars Legatus Legionis
33,629
Subscriptor
finni":3pylxzjn said:
Heresiarch":3pylxzjn said:
Sorry, I mean it assumes everything has 255.255.255.0 and that anything else doesn't exist.
Wow. So it really is broken. That's awful.
I've run into a few fairly modern devices like that, a few DSL modems and such. It can be a real pain in the buttocks for sure. Basically, these days, it is the result of extremely lazy and uninformed programming. Heck just plain stupidity really since you can use all kinds of network frameworks and dev kits that properly implement networking and routing protocols from open source stuff.
 

Paladin

Ars Legatus Legionis
33,629
Subscriptor
Dr. Xing":33ucgfyo said:
I learned to never again trust a 3750 switch stack to keep the same switch as master after adding another stack to the switch and rebooting the stack.

startup-config : gone
vlan database: gone

Fortunately, I had backed up the config before starting the whole process and the vlan database was easily recreated after re-entering the VTP information and letting it propagate from another stack.

I will manually set the switch priority on all 3750 stacks from this point forward!

I think I have yet to encounter a switch stack system that gets it right 100% of the time if you don't do it manually. I have been burned by a variety of manufacturers and models when it comes to stacking and I always do manual now days if I remember to/have the ability to do so.
 

Paladin

Ars Legatus Legionis
33,629
Subscriptor
Yeah I should have been more specific. I have rarely lost config, but I have often encountered stack setups that will randomly pick a new master when a new switch comes online if you haven't manually prepared either the stack with a non-default priority for your desired master and secondary or prepped the new switch to be lower priority. Depending on the manufacturer or software, this usually seems to result in all the switches dropping their ports just long enough for people to get pissed. :mad: :rolleyes:
 

Paladin

Ars Legatus Legionis
33,629
Subscriptor
mkg":16olles1 said:
I've learned what "multiplexing" means in the Microsoft SQL licensing world. For all of you who've never heard of this little chestnut, I suggest you find it and learn all about it. :scared:

Why don't you let us in on it then? This is a great place (not this thread specifically if you want to go into a real detailed post, but this forum) for you to do a little write up on what you learned and share it with others.
 

Paladin

Ars Legatus Legionis
33,629
Subscriptor
Le sigh*

Once again, I have to do everything myself. I don't know how many times I asked my engineers working on a vps platform development if they had updated all the firmware and software on the various Dell pieces of gear involved (Equalogic san, blade chassis etc.). Answer was always, 'yes right after we got it.'

After 4 months of banging our collective head on a network instability problem when under high IO load on the SAN, a Dell engineer says, "Hey you guys should give the updated firmware a shot, yours is the factory 1.0 release." :mad: :mad: :mad:

Yeah, everything works great now. 4 months behind where we should be because even though I asked, I guess no one really thought they should check and tell me the truth. They just assumed someone else had done it.
 

Paladin

Ars Legatus Legionis
33,629
Subscriptor
WingMan":2ecito3u said:
Paladin":2ecito3u said:
Le sigh*

Once again, I have to do everything myself. I don't know how many times I asked my engineers working on a vps platform development if they had updated all the firmware and software on the various Dell pieces of gear involved (Equalogic san, blade chassis etc.). Answer was always, 'yes right after we got it.'

After 4 months of banging our collective head on a network instability problem when under high IO load on the SAN, a Dell engineer says, "Hey you guys should give the updated firmware a shot, yours is the factory 1.0 release." :mad: :mad: :mad:

Yeah, everything works great now. 4 months behind where we should be because even though I asked, I guess no one really thought they should check and tell me the truth. They just assumed someone else had done it.

I think I'd be writing up some engineers for that. Time and project money lost for something stupid and easily verified with a little effort.

Yeah, technically they are not under my purview and the big boss knows about it so I will leave it up to him to deal with. It's his pet project really so hopefully he is going to do some swatting on the nose stuff or something.
 

Paladin

Ars Legatus Legionis
33,629
Subscriptor
Grrr.

Once again: When I ask you [customer] for specific information, please provide me the correct response. Do not provide me what you think I *actually* need, or what *you think I can get by with*. If you are not comfortable giving me the info I asked for, just say so and I can find another way to do things the way you like. It's not a big deal really.

If you're paying me to create a dump of your database and import it to the new server you had me build for you, and you give me a short window at 11PM to do it, it seems like kind of a waste to then provide me a username and password that is not the root password I asked for specifically, nor does it have access to about 77% of the databases and tables on the server you want me to duplicate!

Of course, mysqldump doesn't happen to even mention that it can't dump the data I want, it just goes merrily on its way, generating a plausibly large dump file which I then import and get... NOTHING OF VALUE! :mad:
 

Paladin

Ars Legatus Legionis
33,629
Subscriptor
I know a guy who has struggled along with MS NLB for something like 6 years now. He even bought a coyote point load balancer to replace it at one point (not much of a step up but something) and then decided that it was too much like work to figure out the load balancer device interface. Took me about 20 minutes of fiddling to pretty much get it but he just couldn't be bothered.

So he limps along with random performance issues, uneven server load distribution, random server dropouts and hangs as well. What a mess.
 

Paladin

Ars Legatus Legionis
33,629
Subscriptor
I can't even count the number of times people have asked me to put various things in DNS that can't be done (email forwards, and autoresponders, web page rewrites, web page content aware load balancing :eek: etc.). It really doesn't help that godaddy et al seem to be bent on making all of those kind of features 'appear' to be part of DNS (as long as you are willing to pay a couple bucks more).

EDIT:

Man, I remember the Cobalt RAQs. We had one for testing and it was so slow and useless we returned it after a week. We eventually got a RAQ3 I think when we acquired some other company. It was still slow and buggy and didn't really do what you wanted most of the time it seemed.
 

Paladin

Ars Legatus Legionis
33,629
Subscriptor
I'm convinced that kind of stuff is more or less intentional, or at least intentionally left broken. Just like cell phone companies, they count on lazy/uninformed customer to just pay whatever is on the bill and hope that on average, the company is billing more than they should be (most of the time the errors are in favor of the telco obviously). Errors in the customer's favor seem to get fixed much more quickly. ;)
 

Paladin

Ars Legatus Legionis
33,629
Subscriptor
On that same line... how is it that people always know exactly when to stop procrastinating and bring all their needs to me at about 7 minutes before I am leaving for the day? Seems like every day I get my stuff lined up so I am ready to take off on time and suddenly everyone else realizes they have been sitting around doing nothing all day and suddenly everything is a top priority to get it done now! :rolleyes:

Fortunately, people still have yet to grasp the idea that I will just tell them to open a ticket on the issue and I will get to it so I can just tell them that on the way out the door. Eventually they will learn that they can open the ticket 2 minutes before walking over to my desk and eliminate my excuse. :( :D
 

Paladin

Ars Legatus Legionis
33,629
Subscriptor
bigmikebrooklyn":3iqz0nh5 said:
Fulgan":3iqz0nh5 said:
Today, I learned than it's still possible in 2010 for a major residential ISP to be clueless enough to force a reverse DNS check on their SMTP gateway.

This was a bad idea in 1995, it's unforgivable 15 years later.

okay, i'm gonna bite, i'm not ashamed, why is this bad?
I have my ideas, but they don't seem so earth shattering, so i assume i'm missing something.

It's pretty standard practice really. Anyone with IP addresses with incorrect/conflicting reverse DNS that is not a customer of that ISP (and therefor not entitled to relay through the server for any reason) probably should not be connecting to it. If the host in question is another mail server, it should have proper DNS/reverse DNS from day one. Whinging about people's security implementations like, especially when it is a pretty bare minimum like that is well, just that: whinging. ;) :D

Enigmaunbound":3iqz0nh5 said:
Don't send the password audit reset list to the Helpdesk on the day before a holiday weekend. Whoops.
Not seeing why?
They reset everyone's passwords? How does the helpdesk have access to do that? Or did they start calling people to have them reset their passwords or something?
 

Paladin

Ars Legatus Legionis
33,629
Subscriptor
Fulgan":29az54b8 said:
It's bad because:
1/ It breaks multi-hosting
2/ It's a real pain to setup and even worse to manage if you don't control your reverse
3/ It helps little with spam filtering since all it takes is a single reverse by the spammer to get the expected name
4/ it actually tells you nothing about the nature of the mail, only the way the DNS is setup.

It's pretty standard practice really. Anyone with IP addresses with incorrect/conflicting reverse DNS that is not a customer of that ISP (and therefor not entitled to relay through the server for any reason) probably should not be connecting to it.

And prey, how is an external party supposed to send mail TO them if not by connecting to the machine listed as the domain MX through SMTP ?

If the host in question is another mail server, it should have proper DNS/reverse DNS from day one.

Only in order to follow these rules dating from the day IPv4 was a waste desert and anyone askign for an IP to connect a webcam pointing at a coffee pot would get a class C by default. Nowadays, it's only a legacy rule that serves no useful purpose and that almost every serious player has abandoned. I don't see gmail, hotmail oand yahoo doing it, do you ? that's because they have a real spam filter in place that will actually TRY to avoid false positives by checking the SPF record instead of the PTR.

And I DO have a proper PRT record. It just doesn't equal the mail server's host name.

It doesn't break anything to do with shared IP hosting. As long as the server reliably uses a single hostname as its name when it makes outgoing connections for SMTP, just make sure your reverse DNS is the same (or vice versa). Make sure your forward DNS agrees. All done. Additionally, if you need an SMTP binding to use a separate hostname for outgoing mail you probably need another IP address (and attendant rDNS PTR) anyway so you can bind another port 25. Otherwise your MTA has to do some funky trickery which may result in unreliable results. Nobody cares though if your mail server says, 'Hi, my name is mail.toolboxservers.com and I want to send a message from tim@monkeypantsonline.com to you!' as long as the server's reverse DNS really is mail.toolboxservers.com

If you don't control the DNS it can be a bit of a pain to get it setup right but if your ISP won't/can't do it, that's a great reason to find a new one. It's a really trivial thing to change.

As for how much it impacts spam these days... probably not that much as you say, but it is a really easy baseline indicator of whether the host in question that is trying to send your server mail is 'serious' or not. It's also a really common practice in my opinion/experience.

And by day one, I meant day one you turn on your new mail server, not day one of the internet. :D

As for large mail hosts doing it? I know at least they used to. I used to see that stuff in people's logs all the time about yahoo and aol and other big mail providers requiring valid reverse DNS. Gmail does have very good spam filtering so I couldn't say for sure whether they do this reverse dns stuff or not, never run into it as an issue. Yahoo and some of the others you mention, terrible spam filtering. Basically they seem to take the tack of 'if you pay us to let it through, we will. If users mark your messages as not spam, we might whitelist you, otherwise, you're probably going in the spam folder unless we get enough blowback from a concerned mail admin somewhere'. :(

I've dealt with them and comcast and the like so many times on issues like this. I mean, I spent 2 months trying to get AOL and Yahoo to believe that IP addresses that had never been used before (new allocations from ARIN, and I mean new not reallocations) were not supposed to be on their blacklists 'just because'. :mad:
 

Paladin

Ars Legatus Legionis
33,629
Subscriptor
Soko":o0fc796n said:
That the hottest day of the year so far means two of our three AC units will invariably DIE. This lead to an EVA8000 screaming "OMFG IT'S HOT IN HERE!!". We started shutting down non-essential systems, but that only slowed the temperature rise.

A jury rigged air flow system (Hey - you have a fan - I need it NOW *nab* kthxbai!) to help the remaining 5 ton kept the temps to a reasonable level while we waited for the AC fix-it guys to show up. They did, fixes in place, all is moving back to normal.

I need a b33r now. A nice, very COLD, very large b33r.

Ah, I'm so glad to be moving past this point in my datacenter experience these days (I think/hope). We just last week put some of the finishing touches on our second 100+ ton water chiller system install for cooling our stuff. We are going from 1 carrier aquasnap to 2 carrier aquasnaps (the new one being nicer and just a bit bigger) so we are redundant.

Of course, during the transition we managed to have the plumbers shut off the cooling and the original chiller a couple times by accident (heart attack!) but only for a couple minutes. The best was when they threw the wrong valve lever and closed the flow from the chiller to the datacenter completely instead of closing an isolation valve on just one cooling/fan unit. Doh: pump on chiller shuts down when the flow sensors say, "Hey, where my water!?"

Now that it is more or less done, I can sleep much better. Not to mention the fact that utilization on the old chiller was reaching 89% pretty much every day and one of the 5 compressors on it was/is sounding like a '72 VW Bug engine. :scared: Being under 40% on 2 units is much more conducive to my peace of mind. Of course, the install/turn up is only 3 months overdue from the original schedule from the engineers but the CFO is holding them to their bid/quote like a steel+cement anchor. :D
 

Paladin

Ars Legatus Legionis
33,629
Subscriptor
As I mentioned up the page a bit... we're finishing up the install and transition to a new chiller for our AC system. Part of that was to install a new chilled water loop in the datacenter. The new loop is some PVC like super plastic stuff that is supposed to be super neato (cheap, strong, non-corroding etc.). The old loop was copper.

Yeah, a lot of copper pipe.

Copper pipe = $$

Of course, even though it is not written anywhere in the project specs, contracts, estimates etc. The engineer group doing the transition was about 2 seconds from uninstalling the old pipe and walking off with it. About $70,000 worth of copper pipe in lengths that can easily be sold and reused. LOL. Boss shut that down quick.

Of course, the engineer's response is: If we can't have it, we're not pulling it out. Even though it is written in the project bid. :rolleyes:
 

Paladin

Ars Legatus Legionis
33,629
Subscriptor
Spatula":q42fp7le said:
Identical hardware, identical software. Why does each of the Adobe Pro 9 updates want to reboot on some computers but not others? I am not actually looking for an answer, but it is still frustrating.
Probably depends on what was running when the update ran/was started. And yes, that kind of thing is frustrating.
 
Status
You're currently viewing only Paladin's posts. Click here to go back to viewing the entire thread.
Not open for further replies.