|
Mike Fedyk
[Recent Entries][Archive][Friends][User Info]
Below are the 11 most recent journal entries recorded in the "Mike Fedyk" journal:
|
06:22 am
[Link] |
VMware Server 1.0.3 on Ubuntu Feisty 7.04
Following "How To Install VMware Server On Ubuntu 7.04 (Feisty Fawn)" at howtoforge, getting VMware server installed and configured was a snap. :)
Tags: feisty, ubuntu, vmware
|
|
09:57 am
[Link] |
Virtual Networking
"Host only networks" in VMware are cool. If you want 50 networks on your system, you don't have to worry about having the physical network hardware (or VLANs) to separate the traffic.
Now I just need to find where vmware stores their configuration so I don't have to edit it through the vmware-config.pl program and find out how to turn off dhcp service to host-only-networks.
That's the one hole I see in this setup. The dhcp service for host-only-networks is from the host, and that can be an attack vector. So far I've just been killing them since I don't need them and don't want dhcp on those networks.
Tags: dhcp, networking, virtualization, vmware
|
|
09:42 pm
[Link] |
Routing, VMware & Wireless
I spent today hacking with my friend Charles Wyble in addition to returning and borrowing some books. :)
We setup a Linux router in a VMware guest. His machine now has five ethernet ports (one built-in and four in a multi-port card), but the tulip driver in Ubuntu Dapper Server has a weird probing quirk. The ports that physically are ordered 1, 2, 3, 4 were detected as eth4, eth1, eth2, eth3. No problem. Just reorder the network interface names in /etc/iftab and we're set.
Next we configured the vmware networking so that each vmnet is associated with the correspondingly named ethX number. So eth1 -> vmnet1, etc. Then we added a "host only" interface, but it took over vmnet1 and that was fun debugging. "Why am I not seeing my public IP address on this interface?!"
After getting packets flowing, installed shorewall and setup a dmz. A cursory look at this setup shows that there is a price to pay in resource usage. What would take 2-7% cpu usage (a torrent of FC5 at 600Kbyte/s) to do on the the host system takes about 25% processor usage within the VM guest on an AMD Athlon XP 2200+. The security and flexibility of this setup is unquestionable, but it is good to know the cost.
Also Charles had a problem with Windows in a VMware guest on his Ubuntu Feisty 64bit laptop not being able to get out to the network. It turned out not being a windows problem, but a ndiswrapper problem with his bcm43xx chipset wireless card. Plugging in a wired ethernet connection worked around the problem and getting a Ralink mini-PCI card will be a more permanent solution.
I guess it was a windows problem, since a windows NDIS driver caused the problem. Heh.
Tags: linux, ralink, shorewall, vmware
|
|
05:46 pm
[Link] |
Quad core CPU support in Centos 4.5
The machine I saw the error in with Vmware Server 1.0.2 (Yes, I need to upgrade...) was running on a 2xquad core xeon server (2 sockets). Now centos 4.5 has support for quad core.
I wonder if that has anything to do with the error I was seeing in the centos5 guest...
Tags: centos4, centos5, quad core, vmware
|
01:04 pm
[Link] |
CPU errors with SMP Centos5 on VMware Server
Looks like I'm running into the "Centos beta 5 shows cpu errors on vmware" with Vmware Server 1.0.2 on a Centos4 host. Switching the VM to one CPU seems to avoid the problems. We'll see.
Tags: centos5, smp, vmware, vmware server
|
10:06 am
[Link] |
Vmware server and LVM
I wanted to store my VMs on LVM with each partition on a separate Logical Volume (LV) and the MBR stored in the vmx file. It looks like I can do this with Xen, but vmware doesn't even allow you to select LVM devices. I tried putting in a full path, but it just says "Unable to determine the bus type of device /dev/mapper/VolGroup00-base_guest". This is on a Centos4 host. Boo vmware.
The nice thing about putting this on LVM is that each LVM LV is treated as a separate drive by the disk accounting subsystem in Linux so that you can see how much activity each VM is generating. I guess I'll just have to monitor each VM independently... :-/
Tags: centos, lvm, vmware, xen
|
|
06:22 pm
[Link] |
Xen or VMware or OpenVZ?
Of course openVZ is not mutually exclusive with either Xen or VMware. They can be used together, but that is not the purpose of this post...
I found an interesting link to a forum post on the centos mailing list talking about OpenVZ today.
It says that once the memory limits are reached in an OpenVZ Virtual Environment (VE) it kills the process that hit the limit.
<speculation>I suspect (since I haven't researched and found the details yet) that after reaching the memory limit in a VE, OpenVZ simply refuses allocating memory when processes request more memory. That can be bad because a lot of apps don't handle this very well and die. You would see this also if you simply limited the memory allowed to a process group or user.</speculation>
What happens differently when a virtualized kernel (like in Xen or VMware) hits its physical memory limit is that it starts using swap and if swap is full or unavailable[1] it will discard as many memory pages as possible, disk cache, exeutables mmmap()ed, etc. This can allow more memory to be allocated, though with a serious drop in performance.
But the kernel tries hard to satisfy the allocation before refusing.
If my speculation above is correct and OpenVZ hasn't made any modifications to the linux Virtual Memory Manager so it will simulate "memory pressure" in each VE independently, then OpenVZ (and presumably the generic Linux resource limits code) will simply reach the limit and refuse further allocations without trying hard to find space for allocations and have serious problems with applications that don't handle memory allocation failures properly (that would be most programs).
This needs to be looked into further...
1. Some people have gotten into the habbit of configuring systems without swap. This mostly started because of the terrible VMM in the early 2.4.x kernel series. If you're running a kernel from kernel.org after 2.4.16 or any distro kernel, you should configure swap space on your system(s).
Tags: centos5, linux, openvz, virtualization, vmware, xen
|
|
07:06 pm
[Link] |
Fun with xen on centos5
(15:52:39) mm: ping
(16:18:03) mfedyk: pong
(16:18:54) mm: these boxes don't like me, I asked rails-02 to reboot to test the init stuff and xenstored appears to be hung
(16:19:13) mm: can't even get an xm list from it
(16:19:32) mfedyk: hmm
(16:19:52) mfedyk: oyga
(16:20:41) mm: it's been like this for about 20 minutes
(16:23:34) mfedyk: xenstored is stuck in a loop
(16:25:09) mm: figures
(16:25:30) mfedyk: you doing anything disk intensive?
(16:25:38) mm: not doing anything at all
(16:25:57) mfedyk: gah
(16:29:02) mfedyk: ok, restarted xend
(16:29:14) mfedyk: I can ssh into rails
(16:29:24) mfedyk: it's been up for 30 mins
(16:29:25) mm: what's the uptime?
(16:29:27) mm: haha
(16:29:42) mm: sounds like it rebooted, but something was unhappy
(16:29:50) mfedyk: yes
(16:29:58) mfedyk: let me investigate
(16:30:05) mfedyk: tell me if you get more hiccups
(16:40:27) mfedyk: oh, this is a nice read: http://lists.xensource.com/archives/html/xen-devel/2006-10/msg00487.html
(16:41:47) mm: hahah
(16:41:47) mm: GET on /xend/domain/demo is resulting in 16 copies being made.
(16:42:04) mfedyk: and writes...
(16:43:26) mm: hopefully that's something that deadrat backports
(16:52:56) mm: how the heck can rails-02 have a load of 1, it's not doing anything
(16:53:28) mfedyk: probably has a process stuck in D state
(16:53:44) mfedyk: yup, xenwatch
(16:54:52) mm: I give up
(16:54:58) mfedyk: lol
(16:55:05) mfedyk: we should consider using vmware
Tags: centos5, vmware, xen
|
|
05:52 pm
[Link] |
vmware doesn't like cpufreq on my laptop
After thinking about vmware's problem with cpufreq it makes perfect sense.
It's just weird watching ping that normally sends a packet out every second only sending one packet every approx. 4 seconds because the cpu is running at ~500Mhz instead of ~2200Mhz.
Current Mood:  blah
Tags: cpufreq, vmware
|
08:58 am
[Link] |
Uselessness of the FC6 recovery cd
I recently received a drive with an entire FC3 system ready and working and needed to get it operational on one of my systems. Since the last time I used Fedora was back when FC4 was current and before that, FC2, I didn't happen to have any FC3 ISOs on my systems. So I headed to the friendly fedora torrent page and downloaded the FC6 rescue cd.
Surely the installer in FC6 should be the most refined after fielding all of the bug reports from previous versions. Right? Don't hold your breath.
My buslogic scsi controller was correctly detected by knoppix. Why wouldn't the FC6 "rescue" disk?
(Yes, I am going through the process of converting this FC3-system-on-a-drive into a vmware image. Apparently if you tell vmware to use a usb drive directly (I stuck it into an external usb enclosure), it doesn't give you the option of choosing whether the VM sees a symbios or buslogic scsi controller on vmware 1.0.1)
The solution was to use the FC5 rescue cd. Instead of just telling me I don't have any drives (like in FC6), it will prompt me to try different modules from an available list. At that point I rebooted the VM into knoppix to find out what module to use (this blog entry is not in chronological order within itself) and then knew what module to choose in the FC5 module list. In retrospect I could've just checked the output of lspci since the module is named "BusLogic"...
Ok, my copy is finished. Now just have to get grub setup and I'll have a working testing/development environment for a new project. :)
Tags: fc6, fedora, recovery, vmware
|
|
12:31 am
[Link] |
Switching from Debian/Ubuntu to Centos/Fedora
"Why?!! You're going backwards. Most people start with Red Hat and then smarten up and go to Debian."
You're right, I am going backwards. But that's how I usually do things. That's also how I'm learning Active Directory. By knowing LDAP and Kerberos inside and out first. My first distro was Debian "hamm" 2.0 back in December 1998. Not long after that first install I promptly hosed it (chmod -R 777 /), but my second install is still going strong today with the great debian upgrades. The system has survived three computers, and two hard drives. Just about the only thing that is the same as back then is some of the data files.
So why am I switching to Centos on my servers and Fedora on my desktop? Because I already know all of the quirks of a Debian system and whenever I get on a rpm based system I end up spending too much time reading man pages and searching around for simple things. Though I don't expect it to be a painless process.
<rant>
Another thing I usually do when I get on a rpm system is scream (internally) "Where did all of my packages go?!!". Why do they have to call apache httpd? It's apache damn it! And what half-wit came up with the bright idea to put the ever changing version number in the /usr/share/doc directory? Why not just use the package name and be done with it? You can't have two packages with the same name... Oops, you can with kernel packages. Did I already use the word half-wit? Damn. Oh and yum, it's great. I love having 1/3 of the features and 1/10 the speed and 3 times the memory usage of apt-get.
Though they do have multi-arch which will allow for 32-bit firefox until a 64bit version of flash comes out. Debian may have that by 2010, we'll see. Also while installing security updates is *always* painless on Debian, the rigidity of their policies can be a problem. When they say security updates, they mean security updates *only*. No bug fixes, etc. Some things like broken gaim packages, old kernels not supporting key hardware and other issues come out of that rigidity, so running on Debian isn't entirely a bunch of roses.
</rant>
Lack of packages brings me to the external repositories available for RH based systems. Half of them are incompatible with each other, and most of the time you want/need packages from multiple conflicting repositories. rpmforge seems to be trying to bring people together but have excluded the guy(s) from atrpms. I intend to find out why.
It seems to me that the rpm world needs a "universal 3rd-party repo". A repository that doesn't duplicate any packages found in the main distribution, but provides a central place for packages that work together with what's available in the main distro and that 3rd-party repository. Also they need all of the packages available in debian that aren't already in rpm repositories. Maybe it should be called RPMian (.deb is to Debian what .rpm is to RH), but maybe rpmforge or atrpms will be a good project to join instead of starting *another* repository.
I'm all for joining an existing project instead of starting a new one.
Tags: atrpms, centos, deb, debian, fedora, rpm, rpmforge, ubuntu, vmware
|
|
![[info]](http://l-stat.livejournal.com/img/userinfo.gif){kind=small}
mfedyk's journal