xaminmo: Josh 2016 (Default)
### Change to EXT4 mount mode (OKAY before conversion)
vi /etc/fstab

### Reboot into single user mode
shutdown -r now
LILO: linux S

### Unmount or read-only every filesystem
umount -a
mount -oremount,ro /usr
mount -oremount,ro /

### Convert all ext4 into new metadata
Read more... )
xaminmo: Josh 2016 (Default)
Every repo was giving signature errors in apt:
Err:6 http://security.debian.org stretch/updates InRelease
At least one invalid signature was encountered.

This was pretty recent. My updates in May were fine.
This ONLY affected apt* update. Not clean, install, purge, etc.

I could bypass
Read more... )
xaminmo: Josh 2016 (Default)
QEMU on Windows will run ppc64 and ppc64le emulation.
It emulates the same as what PowerKVM on an S812L would provide.
It's kind of slow because there is no KVM module, AND Intel vs PPC,
AND emulator mode is single-core/proc/thread.

You can get Windows installer
Read more... )
xaminmo: (Josh 2014)
Reminder to self:
Debian packages are stored in library archive format.

ar -xv file.deb
This returns three files, in this specific order:
debian-binary # A small text file. Always "2.0\n" for now.
data.tar.gz # All of the filesystem bits that get deployed
control.tar.gz # control, md5sums, and pre/post scripts

Note also that data.tar can be .xz format as well.

There are dpkg-build tools for this, but all of this can be done manually for more control if desired.

. )
xaminmo: (Josh 2014)
Hrm. I read that Evolution could connect to Exchange. Supposedly, old versions would do ExtendedMAPI, and new versions will do EWS.

It's all lies. It doesn't handle local MailDirs properly (but I run IMAP, so that's not TOO bad), and when adding an EWS connection, nothing happens. Tests, apply, nothing added.

Maybe I can try DavMail again, but running multiple daemons (multiple accounts) seems bothersome.
xaminmo: Josh 2016 (Default)
I'm using Ubuntu Precise, which is linux 3.2.0, and various DebUntu packages.

The Amazon Machine ID is ami-82fa58eb. The root device is an 8g block on EBS (required for free tier).

It runs fine under a micro image (also required for free tier).

But I like JFS, and it doesn't have JFS.
Also, I wanted /home to be separate from the jumbo root filesystem.

So, I did this:
* Add the new device (22GB to stay within 30G free tier) as /dev/sdf (first free slot)
root@aws:/root# dmesg | tail
[161800.993075] blkfront device/vbd/2128 num-ring-pages 1 nr_ents 32.
[161801.895396] blkfront: xvdf: barrier or flush: disabled
[161801.905416]  xvdf: unknown partition table

* Create the volume group:
root@aws:/root# apt-get update
root@aws:/root# apt-get install lvm2
root@aws:/root# pvcreate /dev/xvdf
  Physical volume "/dev/xvdf" successfully created
root@aws:/root# vgcreate -Ay -M2 -p32 -s256M rootvg /dev/xvdf
  Volume group "rootvg" successfully created
root@aws:/root# lvcreate -Ay -L 20G -n hd1 -prw -rauto rootvg
  Logical volume "hd1" created

* Create the filesystem
root@aws:/root# apt-get install jfsutils
root@aws:/root# mkfs.jfs -L /home /dev/rootvg/hd1
mkfs.jfs version 1.1.15, 04-Mar-2011
Warning!  All data on device /dev/rootvg/hd1 will be lost!

Continue? (Y/N) y

Format completed successfully.

20971520 kilobytes total disk space.
root@aws:/root# cp -p /etc/fstab /etc/fstab.bak
root@aws:/root# vi /etc/fstab
echo 'LABEL=cloudimg-rootfs   /        ext4   defaults        1 1
LABEL=/home             /home    jfs    defaults,nofail 0 2
' > /etc/fstab

* Update the kernel to support JFS
root@aws:/root# aptitude safe-upgrade
root@aws:/root# apt-get install linux-image-extra-virtual jfsutils 
root@aws:/root# reboot
xaminmo: Josh 2016 (Default)
Years ago, I lost 36G due to an LVM issue. I couldn't even figure out what it was. The LVM 1.0 config just "went away" off of my RAID-5.

Today, I lost one spindle out of my RAID-6. It wasn't even a hard loss. It was a retry timeout. I decided to reboot.

I rebooted. Not a crash. No problems. Just a reboot. The system didn't come up. I had Erica power the machine on, and it came up to a ^D prompt for one filesystem. This filesystem was marked in fstab with "nofail", so it shouldn't have stopped the boot.

Whatever. ^D to continue and I'll mess with it in normal mode. Well, it was mounted in normal mode.

So I dismount it and run fsck. INFINITE ERRORS! fsck.ext4 begins scrubbing my filesystem of any useful information. It won't use any of the superblock backups either.

So I gave up after about 20 minutes of errors scrolling by. I mount it up and now it says 1% used. BAH!

It *looks* like all of my files are actually there, so I'm trying to copy them off to other places for inspection/salvage.

But this is unacceptable. RAID FUCKING SIX, one drive down. No other system problems, and I had no problems before the reboot. I just wanted to reset the spindle.

There is no reason for fsck.ext4 to scrub the entire filesystem of data. it looks like fsck.ext4 is the same program as for all other fsck.ext*, and it looks like it has no idea what ext4 really is.

So, not only did fsck not complete in a few minutes as it should with ext4, but it munged my filesystem.

THIS SHOULD NEVER HAPPEN, especially when I still had N+P in the RAID.

Of course, if I had run e2image prior to fsck, then I could restore the image; however, who is in the habit of running that? I just ran fsck and walked away. *sigh*

I cannot express how displeased I am with this.
xaminmo: Josh 2016 (Default)
Linux's compcache (compressed cache) became ramzswap.
This is like "Active Memory Sharing" for pSeries Hypervisor.
Linux's implementation is as an LZO compressed block device in kernel RAM.

As development went on, there was no need to limit this to paging space. Now, it's called zram. It's not super super stable, but it seems to work well enough when you leave it alone.

Linux already supports "priority" for paging spaces (ie, hierarchically used in order of best to worst). As such, backing devices are moot anyway. We'll just store non-compressible junk in RAM. It's rare, and not any worse than if there were no LZO.

zram is single-threaded, which is disappointing; however, its' performance is about 50% of RAM, and about 800% of disk. It's also allowed to run multiple devices at the same priority and each device will sit on its own core.

my foray into zram )
xaminmo: Josh 2016 (Default)
Today, I avoided my assignments from 11am until 7pm.

The fruits of this slack is a status script for Linux-HA
full text of script here )

This is fairly simple but should be drop-in for most environments.
I tested it on an OpenFiler 2.3r2, 2-node cluster.
xaminmo: (Gloomy Angry)

Why does internet mail have to be so fucking complicated!?!?!

Exim, by itself, is just a configurable filter.

In order to actually get proper filtering, you have to tie in spamassassin (SLOOOOWWW) and clamav (FLAKY!!!!) or something similar.

So, spamassassin is tweaky and kludgy on Debian. Well, all exim stuff on Debian is kludgy. "We like Debian-exim rather than mail for historical reasons... blah blah blah".

Anyway, right now, my direct, immediate problem is that clamav is not recreating its socket.
The dir is owned by clamav.adm.
The dir is 775
clamd.conf shold fix this:
LocalSocket /var/run/clamav/clamd.ctl
FixStaleSocket true
I even uninstalled and reinstalled it. The log shows no errors:
clamav.log here )
BUT! Still no /var/run/clamav/clamd.ctl

Even worse, there's no way to have debian check for the socket and skip antivirus scanning if it's missing.

Enabling/disabling isn't just a simple switch. It's multiple lines of config.

I am so pissed right now.

A) Things should generally work.
B) When things don't work, they should report WHY they choose to not work.


Oct. 28th, 2007 03:11 pm
xaminmo: Josh 2016 (Default)
This makes me want to retch, but is promising for those susceptible to MRSA wounds:

Congress protects whistleblowers by including ALL of their email addresses to everyone who had submitted a web comment:

Photoblogging all day long to protect himself from the FBI:

Linux on Laptops (compatibility):


Jun. 30th, 2007 11:00 am
xaminmo: Josh 2016 (Default)
Having PPTPD running on your iptables firewall is really difficult to get working.
There are plenty of sites about PPTPD, or iptables, or PPTP THROUGH iptables.
There's not much about having them on the same box.

I've added all the rules I could think to add, and all that I could find that seemed to not be redundant.

It probably also complicates things that my internet feed is actually a PPPoE connection too.
xaminmo: Josh 2016 (Default)
I configured fastestmirror and protectbase as per:

Since I couldn't get the sysfsutils/lm_sensor issue resolved, I updated with this:
# yum update -x \*sysfs\* -x \*sensor\*

Note that CentOS-5's kernel is 2.6.18-8.1.3.el5 wheras Fedora Core 6, before yum, is 2.6.20. FC is not enterprise class, so I'll take a heavily debugged 2.6.18 over a moderately debugged 2.6.20 when it comes to corporate infrastructure.
xaminmo: Josh 2016 (Default)
OK, so with Corsair RAM, the box seems very stable.

CentOS installed like a champ, and even detected sound/video (nForce-MCP550 motherboard, GeForce 7100 PCI-Express video).

I downloaded, but didn't install the nVidia branded drivers for video, since that doesn't auto-update, and the livna drivers seem to work well.

I did add a pile of repositories in order to get "yum update" to be happier. Some of the KDE stuff still depends on i386 versions of libraries rather than x86_64. (lm_sensors and sysfsutils). Unfortunately, this is a bit of a hang-up.

Repos I added )
xaminmo: Josh 2016 (Default)
Shuttle XPC SN27P2 with nForce-MCP55 chipset
GeForce 7100 PCI Express
two OCZ DDR2 5-5-5-15 1GB DDR4 DIMMs
two Athlon X2 4600+
two Maxtor (Seagate) DiamondMax 500GB SATA drives
one DVD burner
Unk rated, generic power supply.

CentOS 5 x64 hangs during verifying packages, or when warm, it will hang during Anaconda start-up.

Ok, so Fedora 6 went on, which is also RHEL5.

Even with fans turned up, if I chew up both CPUs with gzip -9, it will hang within 15 mins.


OCZ memory is aparently no stable in these, but lots of people have stability issues with this box.

My company bought it to replace our hosted webserver.


Plus 4 hours of Gary's time.

Plus 9 hours of my time.

I think it would have been better to buy server grade system. I'm fond of the Dell Power Edge line for its pricepoint and reliability.
xaminmo: Josh 2016 (Default)
SCO was served notice on Apr 23 by NASDAQ of noncompliance. Basically, they've been under $1.00/share for a month.

This means they have 180 days to shape up...

And if they don't, then they would be served a 180 day notice...

During that second notice period, they could move to the cash market for another 180 days grace period...

And at the end of that, they would be delisted from NASDAQ.

Now, this isn't bankruptcy. This just means they wouldn't be listedon NASDAQ. They could still be purchased in the OTC market, or through direct trades with a stockholder of record.

But it's a bad sign. Their revenue per share has dropped substantially over the last 4 years, and their share price has dropped substantially as well.
xaminmo: Josh 2016 (Logo Linux)
Ok, so the mail issue was actually a firewall issue, but in the troubleshooting, I did a clean reinstall of exim4.

Mail seems to be really happy after fully reconfiguring it. Virus scanning seems to mostly be working, though there were 6 failures lastnight. I think maybe a rate limit issue. Will look into it..

Procmail has been disabled since exim has good filtering in .forward files and procmail seemed to be not happy on my server.

Spam scanning seems to be working, but I don't have bayesian filtering enabled. It's too heavy for my little ole mail server. Everyone who doesn't have an external forward set up has it set to automatically send spam as tagged by spamassassin to go to their junk mail folder automatically.


This means that spammers who do not retry temporary rejection will never be received. Since the premise of spamming is to use as little time on their end as possible, this should remove a good chunk of their traffic without having to spend much CPU time on it. This should not affect messages sent by normal, valid, full-time mail servers, which some spammers use. It's not perfect, but it's a cheap way to cut spam counts.
xaminmo: Josh 2016 (Default)
I'm also tired of Debian.

I mean, mail works for 2 weeks, and then just stops.

I get errors all the time.

There's not a default, working config for anything other than simple, basic mail. If you want SSL, that's all sorts of custom shit that is unmaintainable through upgrades.

Same for spam and virus scanning.

I'm tired of shit not working. Always. This is aged, antiquated technology. There is no valid reason for it to be complicated and require constant maintenance just to operate. I'm not even talking about adding new features. I'm talking about just normal, daily operations.

Even if it is my config, or whatever, I can't just uninstall and reinstall. You can't uninstall the MTA on Debian. And then, there's all these stray config files that will be left around, even after --purge.

There's no easy solution though. It's no easier anywhere.

I'm just tired of wasting time on this.
xaminmo: Josh 2016 (Default)
Note to self: no matter how many times I try, "frep" is not a valid command.

Anyway, I'm trying to build a metadevice, and I get this:
# mdadm --create /dev/md0 -f -v -b internal --level=1 --raid-devices=2 /dev/sdi1 /dev/sdj1
mdadm: Cannot open /dev/sdi1: Device or resource busy
mdadm: Cannot open /dev/sdj1: Device or resource busy
mdadm: create aborted

fuser said nothing owned sdi1 or sdj1.
sdi and sdj were open by multipathd.
multipath -l showed both of these disks and 2 other disks as duplicate paths for the same disk.
WTF!?!?! OK. Why when their serial numbers listed by multipath are different? And why not the other disks that are all the same model and revision?

So, apt-get remove multipath-tools and it's all cleared up, but I still get the same errors.
I tried manually modprobing the raid modules.
I tried rebooting.
I tried using a partition size under 100g.

UPDATE Ok, so multipathd didn't unload on uninstall. Another reboot got me going.


xaminmo: Josh 2016 (Default)

August 2017

  1 2 345
1314 151617 1819
27 28293031  


RSS Atom

Most Popular Tags

Style Credit

Expand Cut Tags

No cut tags
Page generated Sep. 23rd, 2017 06:15 pm
Powered by Dreamwidth Studios