Unix, Linux, Database, Monitoring and other I.T. Issues
Finally decided to fix the order of my apache directory listings because I was frustrated by the case of the first letter affecting the listing order, and people not knowing how to use a “unix sort” listing. To fix it, the following line was added in the “<IfModule dir_module>” paragraph :
IndexOptions IgnoreCase FancyIndexing FoldersFirst NameWidth=* DescriptionWidth=* XHTML HTMLtable SuppressRules IconHeight=16 IconWidth=16
Finally replaced the now dead WRT55ag which died on me about 3 nights ago. I tried to resurrect it, but it really wont even power on. It can either be the power supply, or the unit - but no point in trying to save a 5 year old wireless device. After all, 5 years is pretty much longer than any other I’ve in a while. What I did was to use a currently unused wireless device, and largest waste of money ever. Its the WRT54g3g-vn wireless mobile broadband router. It takes an EVDo card from Verizon, and uses that as the internet connection. Then you have a proper router/firewall behind it, with a four port switch. Great idea - no doubt. But I thought I saw that my EVDo was a supported card, but as it turns out, its not. And these days, you dont even get the PCMCIA cards anymore. So there’s really no way this thing would ever get used again. So its now serving as an access point. Sure, its overkill, but at least I’m back in business, and will finally get some use out of this thing.
Well, I dont need to tell anyone about the heat wave we’re having, but its starting to cost me now. My smoke alarm in the attic melted and caused all the others to off insanely loud - so that was no fun. My “Hard Rockin” music server overheated and when I got home from work the other day, that was melting down and stunk pretty bad - so its removed from service. And this time, its the WRT55ag, which is no longer functioning. Its got a bit of a burnt smell to it, and after a good 5 years of good service, it quit. I guess I cant really complain about that many years of service, but now I have to setup a new one, to take over the duty that it was serving. Fun fun …
Had a Comcast Internet outage today, for a little over 2.5 hours. It was down from 07:18:19 to 09:52:46 but at this point its back up.
Doing a bit of DNS work tonite, and while looking for a better name server, came across Google’s public dns server which is awesome. See more about it here.
Updated my sockso server from 1.2.0 to 1.2.4 tonite. Its supposed to be more efficient, and more stable. Has new scheduling features too but I’m not using those yet.
Had some problems with my firewall today and yesterday .. and the day before. Its been running really really heavy, and thats not at all normal. I saw some odd behavior. It started on the 18th, and peaked on the 20th. In the end, I found that at least one of the two configured Comcast DNS servers had changed. So I changed from using
nameserver 68.87.64.146
nameserver 68.87.75.194
to using :
nameserver 68.87.64.150
nameserver 68.87.75.198
nameserver 68.87.29.164
nameserver 68.87.85.132
nameserver 68.87.76.228
nameserver 68.87.72.244
nameserver 68.87.68.244
After doing that, its recovered and working much better.
Had a power hit tonite :
May 20 23:01:47 vm380num2 hpasmd[3741]: NOTICE: hpasmd: System Power Supply: General Failure (Power Supply 1) has been repaired
May 20 23:01:47 vm380num2 hpasmd[3741]: NOTICE: hpasmd: System Power Supplies Not Redundant has been repaired
… but the nice new UPS’s that a friend got me, really pulled through like a champ. In the past, these caused server hits, but this time, it was without a glitch !
Just completed the repair of my FreeNas box. Turns out that the drive reporting all the S.M.A.R.T. errors may not have been bad after all. There were hardware issues on that AthlonXP 1.8Ghz server. So I ended up pulling the drive, and fitting it into an HP D510SFF 1.6Ghz host that I had. I was able to first GHOST the drive off for safe keeping, and then mounted it and tested it in the HP. Worked good, though it took like 6 minutes before the first network packet was received. But I carried the USB card over from FreeNAS (athlonXP) to FreeNAS (Pentium 4 HP), as well as the drives and memory. But now, its all good again. Hopefully it stays up more than 24 hours this time.
Had some problems with the 120Gb drive in my FreeNAS server over the last few days. So tonite, I copied all the data from that share onto my “mirrored share” partition. The ZFS 120Gb partition was my original “test” partition anyway - even though I did end up using it for some backup/archive data. It was a single partition, single point of failure, and was generating a lot of S.M.A.R.T. errors. So I copied all the data off it before it was toast, disabled all the NFS/Samba share stuff, removed the ZFS device and pool info, and then finally the device. In the BIOS I set it to not-used for now, and will eventually get the disk physically out of the machine, but there wasnt time for the “physical” work tonite.
Ran into some issues with my VunbuntuS instance, where dansguardian is reporting old and no longer usable. I had this originally listed in my monitoring, so it started to come up 2 days ago when the new signature updates force the upgrade. Since I dont use that portion of my VubuntuS service, I’m just going to stop monitoring it, and leave it down. It wont affect what I use it for.
Had to edit my zimbra config to allow relaying for my local network only. Funny enough, this wasnt needed until now, but its time has come, so it was setup. The documentation is internal to my site, and not sharable publicly, but it does work and its not much effort. All you have to do is run the “zmprov modifyServer” command with the “zimbraMtaMyNetworks ” option.
I should have posted this a while ago, because it comes up over and over for me. In setting up a few vista machines lately, and having trouble connecting to one of my samba shares, on my Fedora8 server - and here is the fix.
Had a problem with my aging AthlonXP workstation tonite. It locked up, then the fan started screaming, so I shut it down. Its been “complaining” for some time. That was why I originally move the HP a712n next to it, to copy all the data off it, but now I’ve gotten used to using two workstations. So the death of the AXP box presented a problem. So I moved the idle HP dc7600 convertible tower into place tonite. Its a P4 3.2Ghz with hyperThreading, 3Gb RAM, and has vista 64bit on it, and runs really well. So its hooked up with a 1600×1200 resolution and will let me do all that I need. So I’ll add AXP’s drive as a second drive and start using the data from it that way. May even try to do some sort vmware workstation on it, so that I can access the apps from AXP too - but that will take some time.
Had a momentary power flicker tonite but it was enough to bounce a few servers. Took this opportunity to bump my FreeNAS box from 512Mb RAM to 768Mb. The ZFS filesystem I’m using continually complains about there not being enough memory since the ZFS wants 512 Mb by itself. Hopefully this will stop the daily (or more) bounces of the box which I never really got to look into in order to find the root cause. Suspicion was the RAM, so we’ll see.
This is another one of those posts that probably wont help anyone else out, but is more for my own records. Tonite I shutdown the RoundCube server that I had running. I had been using the QmailToaster appliance as my mail server (front and back end, as well as antivirus and spam blocker), but there were parts of the front end interface I thought felt “dated”. Though it was awesome as far as stability and reliability, the look and feel was just a bit “old”. So I had found a product called RoundCube that had a very slick ajax interface, allowing web based dragging of emails, dropping into folders, etc. It was nice, very slick. A little heavy on the infra-requirements, but nice none the less. Well, in recent time, I’ve migrated over to Zimbra, and absolutely love it. Its so much richer, works fantastic, has calendaring, instant message, and so many other features - PLUS it blocks all the spam that was getting through my QmailToaster. So with this Zimbra instance up almost 2 months now, I can see there is no going back. As a result, I no longer need RoundCube at all, hence the power down. QmailToaster is next, now that I have proper relaying working on Zimbra.
Not much value for anyone reading this, more for my own records. But moved the T23 out of the server room today and put an R40 in its place. The R40 is a much faster laptop and will be my workstation for the server room.
Ok, so I’m currently working on a personal test project that requires me to have four different user sessions in a browser based application. The problem is that I cant use the same browser for more than one session, because each new tab just resumes the account already logged on. So I was using Firefox and Flock for two of the browsers, and Chrome and Safari for the other two. But since doing this, the memory on my machine has been taking a beating. The biggest culprits were Chrome and Safari. So I needed more Firefox type browsers and found two more. One is called CometBird, and the other is the OrcaBrowser. Both are good. CometBird is more like Firefox, whereas Orca takes a different spin on design and options menu’s. But all in all, this is working out much better.
Finally figured out what was stopping my Zimbra host from relaying mail for my other servers. I have a half dozen other hosts on my network that need to get mail out - whether for monitoring purposes, as part of an application’s communications, or otherwise. Until now, I’ve needed to keep my Qmail server up and relaying until I could figure it all out. And today I finally did. So my documentation on the Zimbra setup has been updated with a new section on how to relay through Zimbra, for my network only.
While trying to make my new mail server client “portable device friendly”, I was in search of a cheap iPhone or iTouch that I could use to test the mobile device browsing. But while searching for that, I found something even better. Its an iPhone simulator application, as reported here too.