Thread: [SSI-users] trying to build OpenSSI from source
Brought to you by:
brucewalker,
rogertsang
Menu
▾
▴
ssic-linux-users
|
From: Scott W. <sc...@sl...> - 2010-07-03 22:46:23
|
Hi all,
I spent some time tracking down security fixes to the Linux kernel and
patching 2.6.14-ssi.
Trying to build the linux-source-2.6.14-ssi.tar.bz2 kernel from
`apt-get install linux-source-2.6.14-ssi`, I get:
CC init/main.o
In file included from include/asm/mpspec.h:5,
from include/asm/smp.h:18,
from include/linux/smp.h:19,
from include/linux/sched.h:26,
from include/linux/module.h:10,
from init/main.c:16:
include/asm/mpspec_def.h:78: warning: 'packed' attribute ignored for
field of type 'unsigned char[6]'
In file included from include/linux/blkdev.h:10,
from include/linux/ide.h:14,
from include/cluster/synch.h:35,
from include/linux/vproc.h:16,
from init/main.c:76:
include/linux/pagemap.h: In function 'fault_in_pages_readable':
include/linux/pagemap.h:236: warning: passing argument 1 of
'ssi_rcopy_from_user' discards qualifiers from pointer target type
include/linux/pagemap.h:236: error: assignment of read-only variable '__gux_val'
include/linux/pagemap.h:242: warning: passing argument 1 of
'ssi_rcopy_from_user' discards qualifiers from pointer target type
include/linux/pagemap.h:242: error: assignment of read-only variable '__gux_val'
make[1]: *** [init/main.o] Error 1
make: *** [init] Error 2
I get this same fatal in the patched and unpatched source trees.
Is this a toolchain problem? Are there known-good versions of things
I should be using to build this? Is building from source particularly
tricky?
Also:
Trying to check out CI from `cvs
-d:pserver:ano...@cv...:/cvsroot/ci-linux login`, the
(empty) password is rejected:
scott@debian:~/ssi$ cvs
-d:pserver:ano...@cv...:/cvsroot/ci-linux login
Logging in to :pserver:ano...@cv...:2401/cvsroot/ci-linux
CVS password:
cvs login: authorization failed: server cvs.openssi.org rejected
access to /cvsroot/ci-linux for user anonymous
scott@debian:~/ssi$
Thanks for any suggestions anyone can offer.
Cheers,
-scott
|
|
From: Mulyadi S. <mul...@gm...> - 2010-07-04 02:06:15
|
Hi Scott... On Sun, Jul 4, 2010 at 05:46, Scott Walters <sc...@sl...> wrote: > Hi all, > > I spent some time tracking down security fixes to the Linux kernel and > patching 2.6.14-ssi. > > Trying to build the linux-source-2.6.14-ssi.tar.bz2 kernel from > `apt-get install linux-source-2.6.14-ssi`, I get: > > CC init/main.o > In file included from include/asm/mpspec.h:5, > from include/asm/smp.h:18, > from include/linux/smp.h:19, > from include/linux/sched.h:26, > from include/linux/module.h:10, > from init/main.c:16: > include/asm/mpspec_def.h:78: warning: 'packed' attribute ignored for > field of type 'unsigned char[6]' > In file included from include/linux/blkdev.h:10, > from include/linux/ide.h:14, > from include/cluster/synch.h:35, > from include/linux/vproc.h:16, > from init/main.c:76: > include/linux/pagemap.h: In function 'fault_in_pages_readable': > include/linux/pagemap.h:236: warning: passing argument 1 of > 'ssi_rcopy_from_user' discards qualifiers from pointer target type > include/linux/pagemap.h:236: error: assignment of read-only variable '__gux_val' > include/linux/pagemap.h:242: warning: passing argument 1 of > 'ssi_rcopy_from_user' discards qualifiers from pointer target type > include/linux/pagemap.h:242: error: assignment of read-only variable '__gux_val' > make[1]: *** [init/main.o] Error 1 > make: *** [init] Error 2 either you get buggy patched kernel, or this version of kernel is somewhat compiled with old gcc .... uhm, you use gcc 2.x? gcc < 3.2? (I just throwing ideas to ease bug hunting here)... -- regards, Mulyadi Santosa Freelance Linux trainer and consultant blog: the-hydra.blogspot.com training: mulyaditraining.blogspot.com |
|
From: John H. <john@Calva.COM> - 2010-07-05 10:57:32
|
Scott Walters wrote: > Hi all, > > I spent some time tracking down security fixes to the Linux kernel and > patching 2.6.14-ssi. > > Trying to build the linux-source-2.6.14-ssi.tar.bz2 kernel from > `apt-get install linux-source-2.6.14-ssi`, I get: > > CC init/main.o > In file included from include/asm/mpspec.h:5, > from include/asm/smp.h:18, > from include/linux/smp.h:19, > from include/linux/sched.h:26, > from include/linux/module.h:10, > from init/main.c:16: > include/asm/mpspec_def.h:78: warning: 'packed' attribute ignored for > field of type 'unsigned char[6]' > In file included from include/linux/blkdev.h:10, > from include/linux/ide.h:14, > from include/cluster/synch.h:35, > from include/linux/vproc.h:16, > from init/main.c:76: > include/linux/pagemap.h: In function 'fault_in_pages_readable': > include/linux/pagemap.h:236: warning: passing argument 1 of > 'ssi_rcopy_from_user' discards qualifiers from pointer target type > include/linux/pagemap.h:236: error: assignment of read-only variable '__gux_val' > include/linux/pagemap.h:242: warning: passing argument 1 of > 'ssi_rcopy_from_user' discards qualifiers from pointer target type > include/linux/pagemap.h:242: error: assignment of read-only variable '__gux_val' > make[1]: *** [init/main.o] Error 1 > make: *** [init] Error 2 > > I get this same fatal in the patched and unpatched source trees. > > Is this a toolchain problem? Are there known-good versions of things > I should be using to build this? Is building from source particularly > tricky? > > Yes, yes and yes. You need gcc-3.3 or gcc-3.4 > Also: > > Trying to check out CI from `cvs > -d:pserver:ano...@cv...:/cvsroot/ci-linux login`, the > (empty) password is rejected: > > scott@debian:~/ssi$ cvs > -d:pserver:ano...@cv...:/cvsroot/ci-linux login > Logging in to :pserver:ano...@cv...:2401/cvsroot/ci-linux > CVS password: > cvs login: authorization failed: server cvs.openssi.org rejected > access to /cvsroot/ci-linux for user anonymous > scott@debian:~/ssi$ > Using the command shown on the http://ci-linux.sourceforge.net/ page works for me: cvs -d:pserver:ano...@ci...:/cvsroot/ci-linux login (Sourceforge's CVS configuration is a never ending mystery. There are innumerable commands that look like they might work, but you have to give exactly the right one). |
|
From: Scott W. <sc...@sl...> - 2010-07-17 22:25:36
|
W00t! There are a heck of a lot of local privilege escalation exploits for Linux. Maybe hundreds. Most of them were introduced in later kernels or require obscure devices to be loaded. So far, I've only bitten on the highest profile ones. I couldn't find any neatly organized list of Linux exploits anywhere. redhat.com directs every attempt to find exploits to whitepapers about how great security in RedHat is. Debian was the best, but it still took going through their list, then on to CVEs, to Linux changelongs, to github linux-mirror blame logs, to the actual commit behind that codechange, then sticking '.diff' on the end of that URL and feeding it to wget in order to get the patches to fix these things. At least in most cases. Newer exploits reference the git commit hash as part of the kernel.org's gitweb. A surprising number of patches applied cleanly. Vulnerability scanners for Linux look like a project. I couldn't actually get Nessus to go and it looks like it mostly or entirely is interested in scanning the network, not the local machine. This makes sense. Any tips about a repository for exploits would be appreciated but I'm really not expecting anything I need to ever be in one neat little pile. I took the low road for now of building from the source package as delivered by apt rather than checking out from CVS. Perhaps I'll try to get that running later. I hate grub with a profound, eternal hatred. I've yet to actually successfully use it. Despite finding dozens of problems (problems that exist for no other bootloader), it plunges on forth and ruins everything, aborting half way through. It can't read the second stage bootblock... but hexdump can. The parens aren't balanced in the config file... yes they are. This is not specific to grub under OpenSSI. Thank bloody $diety that extlinux/syslinux actually does seem to work with OpenSSI. It took me a little bit to find ssi-ksync-network and regenerate /tfpboot. The new kernel booted without problems... but so did the old one. Heh. I need to wade through these CVEs some more. I think I still have work to do in that department. Alright, pardon my verb(i?)age. I just had to share. Thanks, -scott On 7/5/10, John Hughes <jo...@ca...> wrote: > Scott Walters wrote: >> Hi all, >> >> I spent some time tracking down security fixes to the Linux kernel and >> patching 2.6.14-ssi. >> >> Trying to build the linux-source-2.6.14-ssi.tar.bz2 kernel from >> `apt-get install linux-source-2.6.14-ssi`, I get: >> >> CC init/main.o >> In file included from include/asm/mpspec.h:5, >> from include/asm/smp.h:18, >> from include/linux/smp.h:19, >> from include/linux/sched.h:26, >> from include/linux/module.h:10, >> from init/main.c:16: >> include/asm/mpspec_def.h:78: warning: 'packed' attribute ignored for >> field of type 'unsigned char[6]' >> In file included from include/linux/blkdev.h:10, >> from include/linux/ide.h:14, >> from include/cluster/synch.h:35, >> from include/linux/vproc.h:16, >> from init/main.c:76: >> include/linux/pagemap.h: In function 'fault_in_pages_readable': >> include/linux/pagemap.h:236: warning: passing argument 1 of >> 'ssi_rcopy_from_user' discards qualifiers from pointer target type >> include/linux/pagemap.h:236: error: assignment of read-only variable >> '__gux_val' >> include/linux/pagemap.h:242: warning: passing argument 1 of >> 'ssi_rcopy_from_user' discards qualifiers from pointer target type >> include/linux/pagemap.h:242: error: assignment of read-only variable >> '__gux_val' >> make[1]: *** [init/main.o] Error 1 >> make: *** [init] Error 2 >> >> I get this same fatal in the patched and unpatched source trees. >> >> Is this a toolchain problem? Are there known-good versions of things >> I should be using to build this? Is building from source particularly >> tricky? >> >> > Yes, yes and yes. > > You need gcc-3.3 or gcc-3.4 > >> Also: >> >> Trying to check out CI from `cvs >> -d:pserver:ano...@cv...:/cvsroot/ci-linux login`, the >> (empty) password is rejected: >> >> scott@debian:~/ssi$ cvs >> -d:pserver:ano...@cv...:/cvsroot/ci-linux login >> Logging in to :pserver:ano...@cv...:2401/cvsroot/ci-linux >> CVS password: >> cvs login: authorization failed: server cvs.openssi.org rejected >> access to /cvsroot/ci-linux for user anonymous >> scott@debian:~/ssi$ >> > Using the command shown on the http://ci-linux.sourceforge.net/ page > works for me: > > cvs -d:pserver:ano...@ci...:/cvsroot/ci-linux > login > > (Sourceforge's CVS configuration is a never ending mystery. There are > innumerable commands that look like they might work, but you have to > give exactly the right one). > > |
|
From: Cumberland, L. <lon...@ni...> - 2010-08-04 14:17:54
|
Greetings Scott and John (and others that may have some insight on the installation process), I hope that all is well with you today. As you know, I have tried to some time to get the OpenSSI to install on our 64-bit machines that have the E1000 network cards and SCSI hotswap drives with little to no success. With this in mind, I decided to fall back to the 32-bit version of OpenSSI and try it from there after installing the base "Lenny" Debian. I grabbed the "Small CD" i386 from: http://www.debian.org/distrib/netinst and installed the basic setup in "Graphical Install" followed by setting up the source locations and network interfaces (eth0 for Internet and eth1 for cluster-interconnect on 192.168.xx.xx subnet) After the basic Debian install was completed, I have gone through and tried to follow the instructions once from: http://deb.openssi.org/openssi/dists/1.9.6-lenny-preview/ which failed and also once more from the "install log" at: http://deb.openssi.org/openssi/dists/1.9.6-lenny-preview/install-log.html in each case, I did exactly what was said wit the additions of in: Step 10: Edited the /etc/mkinitrd/modules and added the network driver: e1000 Followed by: mkinited -o /boot/initrd.img-2.6.12-ssi-686-smp 2.6.12-ssi-686-smp 11: The listed modprobe ext2 modprobe ext3 modprobe loop modprobe e1000 now it seems that the system, once rebooted cannot mount the filesystem. I have to ask what exact procedure you or others have used and if you have the log files from an install that works since I am at a loss as to why this procedure is not working as it seems like it should. Can you please let me know what might be happening and how I can proceed to get OpenSSI installed on a fresh "Lenny" Debian system which I am now setting up? Thanks and have a great day, Lonnie Cumberland, Prof. Physicist National Institute of Standards and Technology Ionizing Radiation Division (846) Radiation Physics Group (245), Room C106 ADDRESS: 100 Bureau Drive, Stop 8462 Gaithersburg, MD 20899-8462 EMAIL: lon...@ni... http://physics.nist.gov/Divisions/Div846/div846.html |
|
From: John H. <john@Calva.COM> - 2010-08-05 09:01:05
|
Cumberland, Lonnie wrote: > I have to ask what exact procedure you or others have used and if you have the log files from an install that works since I am at a loss as to why this procedure is not working as it seems like it should. > The log files are the exact procedure I followed. I'm sorry that I haven't been able to find any time to work on this for the moment, I'll try again this weekend. (Current tasks, not necessarily in priority order: Port a huge application from SCO UnixWare to Debian Linux Finish remodeling daughters bedroom Replace broken fridge Clean up mess left after cutting down tree Work on OpenSSI Hobby project - Porting software to obsolete ICL mainframe computer. (I don't have enough space for a model railway set)). |
|
From: Cumberland, L. <lon...@ni...> - 2010-08-05 12:42:21
|
Greetings All, I just tried to use "ssi-ksync" on the main node to sync up things for the ramimage that is being sent over via PXE boot to the nodes, but it seems that the dhcp client is missing from the ram image. The message that I am getting on node 2 is: "Gathering Cluster info DHCP client application not found Add dhcp client application to /etc/mkinitrd/exe and rebuild ramdisk image ERROR: Could not find a NIC with node configuration. Halting." I think that I need to edit the ramdisk so that it has the dhck client application, but am not sure how to do that. Any advice would be greatly appreciated. Thanks and have a great day, Lonnie Cumberland, Prof. Physicist National Institute of Standards and Technology Ionizing Radiation Division (846) Radiation Physics Group (245), Room C106 ADDRESS: 100 Bureau Drive, Stop 8462 Gaithersburg, MD 20899-8462 EMAIL: lon...@ni... http://physics.nist.gov/Divisions/Div846/div846.html > -----Original Message----- > From: John Hughes [mailto:john@Calva.COM] > Sent: Thursday, August 05, 2010 5:19 AM > To: Cumberland, Lonnie > Cc: Scott Walters; Openssi users > Subject: Re: [SSI-users] trying to build OpenSSI [with 2.6.14 kernel] > from source > > Cumberland, Lonnie wrote: > > I have to ask what exact procedure you or others have used and if you > have the log files from an install that works since I am at a loss as > to why this procedure is not working as it seems like it should. > > > The log files are the exact procedure I followed. > > I'm sorry that I haven't been able to find any time to work on this for > the moment, I'll try again this weekend. > > (Current tasks, not necessarily in priority order: > > Port a huge application from SCO UnixWare to Debian Linux > Finish remodeling daughters bedroom > Replace broken fridge > Clean up mess left after cutting down tree > Work on OpenSSI > Hobby project - Porting software to obsolete ICL mainframe computer. > (I > don't have enough space for a model railway set)). |
|
From: Cumberland, L. <lon...@ni...> - 2010-08-05 15:16:27
|
Hello All, I just found the "Node Hang at boot" over at: http://wiki.openssi.org/go/Debian which describes part of my problem but even though I tried it as well as the solution in the "Node hang at boot, variant 2", I still get the same message. What's strange is that the e1000 driver is included in the ramdisk image as I mounted it to take a look. I also, added the /sbin/dhclient to the /etc/mkinitrd/exe before I used "ssi-ksync" to rebuild the ramdisk image. No luck so far on getting Node 2 to complete the booting even though it receives the kernel from the tftpd server on Node 1. Any ideas? Thanks and have a great day, Lonnie Cumberland, Prof. Physicist National Institute of Standards and Technology Ionizing Radiation Division (846) Radiation Physics Group (245), Room C106 ADDRESS: 100 Bureau Drive, Stop 8462 Gaithersburg, MD 20899-8462 EMAIL: lon...@ni... http://physics.nist.gov/Divisions/Div846/div846.html > -----Original Message----- > From: Cumberland, Lonnie [mailto:lon...@ni...] > Sent: Thursday, August 05, 2010 8:42 AM > To: John Hughes > Cc: Openssi users > Subject: Re: [SSI-users] trying to build OpenSSI [with 2.6.14 kernel] > from source > > Greetings All, > > I just tried to use "ssi-ksync" on the main node to sync up things for > the ramimage that is being sent over via PXE boot to the nodes, but it > seems that the dhcp client is missing from the ram image. > > The message that I am getting on node 2 is: > > "Gathering Cluster info > > DHCP client application not found > > Add dhcp client application to /etc/mkinitrd/exe and rebuild ramdisk > image > > ERROR: Could not find a NIC with node configuration. Halting." > > I think that I need to edit the ramdisk so that it has the dhck client > application, but am not sure how to do that. > > Any advice would be greatly appreciated. > > Thanks and have a great day, > Lonnie Cumberland, Prof. > Physicist > > National Institute of Standards and Technology > Ionizing Radiation Division (846) > Radiation Physics Group (245), Room C106 > ADDRESS: > 100 Bureau Drive, Stop 8462 > Gaithersburg, MD 20899-8462 > > EMAIL: lon...@ni... > http://physics.nist.gov/Divisions/Div846/div846.html > > > > -----Original Message----- > > From: John Hughes [mailto:john@Calva.COM] > > Sent: Thursday, August 05, 2010 5:19 AM > > To: Cumberland, Lonnie > > Cc: Scott Walters; Openssi users > > Subject: Re: [SSI-users] trying to build OpenSSI [with 2.6.14 kernel] > > from source > > > > Cumberland, Lonnie wrote: > > > I have to ask what exact procedure you or others have used and if > you > > have the log files from an install that works since I am at a loss as > > to why this procedure is not working as it seems like it should. > > > > > The log files are the exact procedure I followed. > > > > I'm sorry that I haven't been able to find any time to work on this > for > > the moment, I'll try again this weekend. > > > > (Current tasks, not necessarily in priority order: > > > > Port a huge application from SCO UnixWare to Debian Linux > > Finish remodeling daughters bedroom > > Replace broken fridge > > Clean up mess left after cutting down tree > > Work on OpenSSI > > Hobby project - Porting software to obsolete ICL mainframe computer. > > (I > > don't have enough space for a model railway set)). > > > ----------------------------------------------------------------------- > ------- > The Palm PDK Hot Apps Program offers developers who use the > Plug-In Development Kit to bring their C/C++ apps to Palm for a share > of $1 Million in cash or HP Products. Visit us here for more details: > http://p.sf.net/sfu/dev2dev-palm > _______________________________________________ > Ssic-linux-users mailing list > Ssi...@li... > https://lists.sourceforge.net/lists/listinfo/ssic-linux-users |
|
From: Mulyadi S. <mul...@gm...> - 2010-08-10 09:09:46
|
Hi Greg Didn't realize you were here too...glad to know... On Tue, Aug 10, 2010 at 02:55, Greg Freemyer <gre...@gm...> wrote: > Hey Mulyadi, > > I think you're ahead of the current status of OpenSSI / Lustre support,: > > http://wiki.openssi.org/go/Lustre > > It's not clear to me if full Luster support it part of the roadmap for 1.9: > > http://wiki.openssi.org/go/Roadmap Whoops, sorry, seems like I didn't update myself properly. But anyway, distributed fs is really fascinating today. @Lonnie: perhaps, just an idea, you would like to consider Dragonfly BSD too? AFAIK they implement distributed fs too. -- regards, Mulyadi Santosa Freelance Linux trainer and consultant blog: the-hydra.blogspot.com training: mulyaditraining.blogspot.com |
|
From: Scott W. <sc...@sl...> - 2010-08-10 09:33:10
|
> @Lonnie: perhaps, just an idea, you would like to consider Dragonfly > BSD too? AFAIK they implement distributed fs too. > > Mulyadi Santosa Dragonfly is still reworking kernel facilities and filesystem primitives to be distributed friendly. They have a good deal of work left. I'm hoping for them. Harddrives in each machine to create one large filesystem might just not be realistic right now. A NAS design might be the way to go. -scott |
|
From: Cumberland, L. <lon...@ni...> - 2010-08-10 13:57:15
|
Thanks to everyone for the wonderful information and I am making really good progress here at the moment. I have just modified my fstab to mount the swaps and drives for each node. ----------------------------------------------------- # /etc/fstab: static file system information. # # <file system> <mount point> <type> <options> <dump> <pass> proc /proc proc defaults,node=* 0 0 UUID=333de891-33d7-46eb-8aff-640abd6133fc / ext3 chard,errors=remount-ro,node=1 0 1 /dev/sda5 none swap sw,node=1 0 0 /dev/sda1 none swap sw,node=2 0 0 /dev/sda1 none swap sw,node=3 0 0 /dev/sda1 none swap sw,node=4 0 0 /dev/sda1 none swap sw,node=5 0 0 /dev/sda2 /etc/mnt/mnt02 ext3 rw,node=2 0 2 /dev/sda2 /etc/mnt/mnt03 ext3 rw,node=3 0 2 /dev/sda2 /etc/mnt/mnt04 ext3 rw,node=4 0 2 /dev/sda2 /etc/mnt/mnt05 ext3 rw,node=5 0 2 /dev/scd0 /media/cdrom0 udf,iso9660 user,noauto,node=1 0 0 /dev/fd0 /media/floppy0 auto rw,user,noauto,node=* 0 0 --------------------------------------------------- Also noticed that doing a "mount -a" from the head node does not seem to load up new fstab and mount the drives. I had to manually do: Onnode 2 mount /dev/sda2 /mnt/mnt02 For each node. Why is that? I will be moving things around a bit, but for now, just wanted to test out some things. Also, if I can mount things through the fstab on the head node then why do I need the command which is discussed in the documentation. Onnode 2 ./clusterfstab --nodenum=2 --mountpoint=<some mount point> Thanks and have a great day, Lonnie Cumberland, Prof. Physicist > -----Original Message----- > From: Scott Walters [mailto:sc...@sl...] > Sent: Tuesday, August 10, 2010 5:33 AM > To: Mulyadi Santosa > Cc: Openssi users > Subject: Re: [SSI-users] Debian Lenny OpenSSI with LVM and RAID 1 > > > @Lonnie: perhaps, just an idea, you would like to consider Dragonfly > > BSD too? AFAIK they implement distributed fs too. > > > > Mulyadi Santosa > > Dragonfly is still reworking kernel facilities and filesystem > primitives > to be distributed friendly. They have a good deal of work left. I'm > hoping for them. > > Harddrives in each machine to create one large filesystem might just > not > be realistic right now. A NAS design might be the way to go. > > -scott > > ----------------------------------------------------------------------- > ------- > This SF.net email is sponsored by > > Make an app they can't live without > Enter the BlackBerry Developer Challenge > http://p.sf.net/sfu/RIM-dev2dev > _______________________________________________ > Ssic-linux-users mailing list > Ssi...@li... > https://lists.sourceforge.net/lists/listinfo/ssic-linux-users |
|
From: Cumberland, L. <lon...@ni...> - 2010-08-05 15:15:49
|
Hello All, I just found the "Node Hang at boot" over at: http://wiki.openssi.org/go/Debian which describes part of my problem but even though I tried it as well as the solution in the "Node hang at boot, variant 2", I still get the same message. What's strange is that the e1000 driver is included in the ramdisk image as I mounted it to take a look. I also, added the /sbin/dhclient to the /etc/mkinitrd/exe before I used "ssi-ksync" to rebuild the ramdisk image. No luck so far on getting Node 2 to complete the booting even though it receives the kernel from the tftpd server on Node 1. Any ideas? Thanks and have a great day, Lonnie Cumberland, Prof. Physicist National Institute of Standards and Technology Ionizing Radiation Division (846) Radiation Physics Group (245), Room C106 ADDRESS: 100 Bureau Drive, Stop 8462 Gaithersburg, MD 20899-8462 EMAIL: lon...@ni... http://physics.nist.gov/Divisions/Div846/div846.html > -----Original Message----- > From: Cumberland, Lonnie [mailto:lon...@ni...] > Sent: Thursday, August 05, 2010 8:42 AM > To: John Hughes > Cc: Openssi users > Subject: Re: [SSI-users] trying to build OpenSSI [with 2.6.14 kernel] > from source > > Greetings All, > > I just tried to use "ssi-ksync" on the main node to sync up things for > the ramimage that is being sent over via PXE boot to the nodes, but it > seems that the dhcp client is missing from the ram image. > > The message that I am getting on node 2 is: > > "Gathering Cluster info > > DHCP client application not found > > Add dhcp client application to /etc/mkinitrd/exe and rebuild ramdisk > image > > ERROR: Could not find a NIC with node configuration. Halting." > > I think that I need to edit the ramdisk so that it has the dhck client > application, but am not sure how to do that. > > Any advice would be greatly appreciated. > > Thanks and have a great day, > Lonnie Cumberland, Prof. > Physicist > > National Institute of Standards and Technology > Ionizing Radiation Division (846) > Radiation Physics Group (245), Room C106 > ADDRESS: > 100 Bureau Drive, Stop 8462 > Gaithersburg, MD 20899-8462 > > EMAIL: lon...@ni... > http://physics.nist.gov/Divisions/Div846/div846.html > > > > -----Original Message----- > > From: John Hughes [mailto:john@Calva.COM] > > Sent: Thursday, August 05, 2010 5:19 AM > > To: Cumberland, Lonnie > > Cc: Scott Walters; Openssi users > > Subject: Re: [SSI-users] trying to build OpenSSI [with 2.6.14 kernel] > > from source > > > > Cumberland, Lonnie wrote: > > > I have to ask what exact procedure you or others have used and if > you > > have the log files from an install that works since I am at a loss as > > to why this procedure is not working as it seems like it should. > > > > > The log files are the exact procedure I followed. > > > > I'm sorry that I haven't been able to find any time to work on this > for > > the moment, I'll try again this weekend. > > > > (Current tasks, not necessarily in priority order: > > > > Port a huge application from SCO UnixWare to Debian Linux > > Finish remodeling daughters bedroom > > Replace broken fridge > > Clean up mess left after cutting down tree > > Work on OpenSSI > > Hobby project - Porting software to obsolete ICL mainframe computer. > > (I > > don't have enough space for a model railway set)). > > > ----------------------------------------------------------------------- > ------- > The Palm PDK Hot Apps Program offers developers who use the > Plug-In Development Kit to bring their C/C++ apps to Palm for a share > of $1 Million in cash or HP Products. Visit us here for more details: > http://p.sf.net/sfu/dev2dev-palm > _______________________________________________ > Ssic-linux-users mailing list > Ssi...@li... > https://lists.sourceforge.net/lists/listinfo/ssic-linux-users |
|
From: John H. <john@Calva.COM> - 2010-08-06 08:46:44
|
Cumberland, Lonnie wrote: > What's strange is that the e1000 driver is included in the ramdisk image as I mounted it to take a look. > Ok, this is better. I could not understand why on earth the e1000 driver would not be in the ramdisk (other than by user error). So now it's looking like the version of the e1000 driver we have doesn't understand your card. Could you please try dropping into the shell and seeing if: 1. the e1000 module is loaded into memory 2. does it show up in ifconfig 3, what does dmesg show. |
|
From: Cumberland, L. <lon...@ni...> - 2010-08-05 17:37:21
|
Thanks for getting back to me Scott, Just tried your suggestion and am still getting the exact same messages to where Node 2 starts the kernel and then go to the point of halting like before. Not sure what is happening here. Do you think that I should have to start all over again with a fresh install of Debian "Lenny" on Node 1 and to a complete re-install? Node 1 seems to be working from what I can tell and even the "cluster -v" shows it as "UP", but no luck with getting Node 2 all of the way up yet. Thanks and have a great day, Lonnie Cumberland, Prof. Physicist (301) 975-6869 (Office) (313) 333-2935 (Cell) (301) 926-7416 (Fax) National Institute of Standards and Technology Ionizing Radiation Division (846) Radiation Physics Group (245), Room C106 ADDRESS: 100 Bureau Drive, Stop 8462 Gaithersburg, MD 20899-8462 EMAIL: lon...@ni... http://physics.nist.gov/Divisions/Div846/div846.html > -----Original Message----- > From: Scott Walters [mailto:sc...@sl...] > Sent: Thursday, August 05, 2010 1:14 PM > To: Cumberland, Lonnie > Cc: John Hughes; Openssi users > Subject: Re: [SSI-users] trying to build OpenSSI [with 2.6.14 kernel] > from source > > Shouldn't have to explicitly add dhcp. > > It sounds to me -- taking a guess here -- that your hand-rolled > mkinitrd has > survived and has never been replaced by OpenSSI's. Node 2 gets the > right kernel > but the wrong initrd. > > Remove /tftpboot/kernel and initrd first just to make sure that they're > being rebuilt and replaced with fresh copies. ssi-ksync-network is a > shell > script and error reporting is never great in shell scripts. I remember > having to tweak things a bit in my work of building and installing > a patched kernel. The initnode happily booted the patched kernel > but then the other nodes came up on the old, original kernel until > I cleaned out that directory and muddled with things until I could get > ssi-ksync-network to go. ssi-ksync calls ssi-ksync-network. If stuff > in /tftpboot isn't rebuilt, step through the script one line at a time > (perhaps just run commands at the prompt) and make sure nothing errors > out or comes up with null data where it shouldn't. Though honestly > my problems were probably related to not using the prescribed > bootloader =) > > Cheers, > -scott > > > > > On 0, "Cumberland, Lonnie" <lon...@ni...> wrote: > > Hello All, > > > > I just found the "Node Hang at boot" over at: > > > > http://wiki.openssi.org/go/Debian > > > > which describes part of my problem but even though I tried it as well > as the solution in the "Node hang at boot, variant 2", I still get the > same message. > > > > What's strange is that the e1000 driver is included in the ramdisk > image as I mounted it to take a look. > > > > I also, added the /sbin/dhclient to the /etc/mkinitrd/exe before I > used "ssi-ksync" to rebuild the ramdisk image. > > > > No luck so far on getting Node 2 to complete the booting even though > it receives the kernel from the tftpd server on Node 1. > > > > Any ideas? > > > > > > Thanks and have a great day, > > Lonnie Cumberland, Prof. > > Physicist > > > > National Institute of Standards and Technology > > Ionizing Radiation Division (846) > > Radiation Physics Group (245), Room C106 > > ADDRESS: > > 100 Bureau Drive, Stop 8462 > > Gaithersburg, MD 20899-8462 > > > > EMAIL: lon...@ni... > > http://physics.nist.gov/Divisions/Div846/div846.html > > > > > > > > > -----Original Message----- > > > From: Cumberland, Lonnie [mailto:lon...@ni...] > > > Sent: Thursday, August 05, 2010 8:42 AM > > > To: John Hughes > > > Cc: Openssi users > > > Subject: Re: [SSI-users] trying to build OpenSSI [with 2.6.14 > kernel] > > > from source > > > > > > Greetings All, > > > > > > I just tried to use "ssi-ksync" on the main node to sync up things > for > > > the ramimage that is being sent over via PXE boot to the nodes, but > it > > > seems that the dhcp client is missing from the ram image. > > > > > > The message that I am getting on node 2 is: > > > > > > "Gathering Cluster info > > > > > > DHCP client application not found > > > > > > Add dhcp client application to /etc/mkinitrd/exe and rebuild > ramdisk > > > image > > > > > > ERROR: Could not find a NIC with node configuration. Halting." > > > > > > I think that I need to edit the ramdisk so that it has the dhck > client > > > application, but am not sure how to do that. > > > > > > Any advice would be greatly appreciated. > > > > > > Thanks and have a great day, > > > Lonnie Cumberland, Prof. > > > Physicist > > > > > > National Institute of Standards and Technology > > > Ionizing Radiation Division (846) > > > Radiation Physics Group (245), Room C106 > > > ADDRESS: > > > 100 Bureau Drive, Stop 8462 > > > Gaithersburg, MD 20899-8462 > > > > > > EMAIL: lon...@ni... > > > http://physics.nist.gov/Divisions/Div846/div846.html > > > > > > > > > > -----Original Message----- > > > > From: John Hughes [mailto:john@Calva.COM] > > > > Sent: Thursday, August 05, 2010 5:19 AM > > > > To: Cumberland, Lonnie > > > > Cc: Scott Walters; Openssi users > > > > Subject: Re: [SSI-users] trying to build OpenSSI [with 2.6.14 > kernel] > > > > from source > > > > > > > > Cumberland, Lonnie wrote: > > > > > I have to ask what exact procedure you or others have used and > if > > > you > > > > have the log files from an install that works since I am at a > loss as > > > > to why this procedure is not working as it seems like it should. > > > > > > > > > The log files are the exact procedure I followed. > > > > > > > > I'm sorry that I haven't been able to find any time to work on > this > > > for > > > > the moment, I'll try again this weekend. > > > > > > > > (Current tasks, not necessarily in priority order: > > > > > > > > Port a huge application from SCO UnixWare to Debian Linux > > > > Finish remodeling daughters bedroom > > > > Replace broken fridge > > > > Clean up mess left after cutting down tree > > > > Work on OpenSSI > > > > Hobby project - Porting software to obsolete ICL mainframe > computer. > > > > (I > > > > don't have enough space for a model railway set)). > > > > > > > > > ------------------------------------------------------------------- > ---- > > > ------- > > > The Palm PDK Hot Apps Program offers developers who use the > > > Plug-In Development Kit to bring their C/C++ apps to Palm for a > share > > > of $1 Million in cash or HP Products. Visit us here for more > details: > > > http://p.sf.net/sfu/dev2dev-palm > > > _______________________________________________ > > > Ssic-linux-users mailing list > > > Ssi...@li... > > > https://lists.sourceforge.net/lists/listinfo/ssic-linux-users > > > > --------------------------------------------------------------------- > --------- > > The Palm PDK Hot Apps Program offers developers who use the > > Plug-In Development Kit to bring their C/C++ apps to Palm for a share > > of $1 Million in cash or HP Products. Visit us here for more details: > > http://p.sf.net/sfu/dev2dev-palm > > _______________________________________________ > > Ssic-linux-users mailing list > > Ssi...@li... > > https://lists.sourceforge.net/lists/listinfo/ssic-linux-users |
|
From: John H. <john@Calva.COM> - 2010-08-06 08:48:46
|
Cumberland, Lonnie wrote: > Thanks for getting back to me Scott, > > Just tried your suggestion and am still getting the exact same messages to where Node 2 starts the kernel and then go to the point of halting like before. > > Not sure what is happening here. > > Do you think that I should have to start all over again with a fresh install of Debian "Lenny" on Node 1 and to a complete re-install? > > Node 1 seems to be working from what I can tell and even the "cluster -v" shows it as "UP", but no luck with getting Node 2 all of the way up yet. > Wait, node1 is up? With the e1000 driver and all? It shows up in the output of "ifconfig all"? |
|
From: Cumberland, L. <lon...@ni...> - 2010-08-06 13:18:12
|
Hi John and Scott, I now succeed in figuring out an efficient process to bet more nodes up and running while now having 5 nodes up and operational. I do have one question about "ssi-addnode". In our physical design, I have a server rack with 30 nodes. 15 of the nodes are connected to 1 switch while the other 15 nodes are connected to a second switch and both switches are cascaded and mounted into the top of the rack. When I run the ssi-addnode command, it only allow me to add a maximum of 15 nodes and am suspecting that it is not seeing beyond the first switch for some reason. How can I go about making sure that the ssi-addnode can see the second switch? Thanks and have a great day, Lonnie Cumberland, Prof. Physicist National Institute of Standards and Technology Ionizing Radiation Division (846) Radiation Physics Group (245), Room C106 ADDRESS: 100 Bureau Drive, Stop 8462 Gaithersburg, MD 20899-8462 EMAIL: lon...@ni... http://physics.nist.gov/Divisions/Div846/div846.html > -----Original Message----- > From: John Hughes [mailto:john@Calva.COM] > Sent: Friday, August 06, 2010 5:10 AM > To: Cumberland, Lonnie > Cc: Scott Walters; Openssi users > Subject: Re: [SSI-users] trying to build OpenSSI [with 2.6.14 kernel] > from source > > Cumberland, Lonnie wrote: > > Thanks for getting back to me Scott, > > > > Just tried your suggestion and am still getting the exact same > messages to where Node 2 starts the kernel and then go to the point of > halting like before. > > > > Not sure what is happening here. > > > > Do you think that I should have to start all over again with a fresh > install of Debian "Lenny" on Node 1 and to a complete re-install? > > > > Node 1 seems to be working from what I can tell and even the "cluster > -v" shows it as "UP", but no luck with getting Node 2 all of the way up > yet. > > > > Wait, node1 is up? > > With the e1000 driver and all? > > It shows up in the output of "ifconfig all"? |
|
From: Cumberland, L. <lon...@ni...> - 2010-08-06 14:59:42
|
Hello All, I was finally able to get our cluster up and running this morning with 5 nodes online for testing. All was going well until I tried to install FreeNX server which seems to have replaced some libraries and my Xserver is no longer working from what I can tell. I used the Aptitude install freenx. Is there a way to roll-back those changes to get the Xserver back up and running as the cluster still seems to be active, but just no Xserver running. Thanks and have a great day, Lonnie Cumberland, Prof. Physicist National Institute of Standards and Technology Ionizing Radiation Division (846) Radiation Physics Group (245), Room C106 ADDRESS: 100 Bureau Drive, Stop 8462 Gaithersburg, MD 20899-8462 EMAIL: lon...@ni... http://physics.nist.gov/Divisions/Div846/div846.html |
|
From: Scott W. <sc...@sl...> - 2010-08-06 18:46:49
|
Unix made it safe and easy to install multiple versions of applications and libraries. Then Debian came along and made it unsafe to install even one version of an application. Welcome to Debian. I don't have any suggestions for recovering what you had. Fixing apt muck-ups like that involves a lot of iterative fussing and has no assurance of success. A lot of people check /usr, the appropriate parts of /var, etc into version control so they can undo apt muck-ups. I just run full backups before attempting any apt operation. In the future, use the --no-upgrade flag as much as you can get away with. apt loves to try to upgrade the universe every time you want to install one simple app. That tells it not to. If you did want to try to muck around with it, I suggest getting on #debian on FreeNode IRC and asking for help. Here's a list of the versions (apt-cache showpkg output) for my packages matching xorg: http://slowass.net/~scott/tmp/xorg.versions.txt I'm not even going to talk about how to downgrade packages as I don't want to be responsibile. Doing anything with apt is playing with fire. Wish I had better news for ya. Regards, -scott On 0, "Cumberland, Lonnie" <lon...@ni...> wrote: > Hello All, > > I was finally able to get our cluster up and running this morning with 5 nodes online for testing. > > All was going well until I tried to install FreeNX server which seems to have replaced some libraries and my Xserver is no longer working from what I can tell. > > I used the Aptitude install freenx. > > Is there a way to roll-back those changes to get the Xserver back up and running as the cluster still seems to be active, but just no Xserver running. > > Thanks and have a great day, > Lonnie Cumberland, Prof. > Physicist > > National Institute of Standards and Technology > Ionizing Radiation Division (846) > Radiation Physics Group (245), Room C106 > ADDRESS: > 100 Bureau Drive, Stop 8462 > Gaithersburg, MD 20899-8462 > > EMAIL: lon...@ni... > http://physics.nist.gov/Divisions/Div846/div846.html > |
|
From: Cumberland, L. <lon...@ni...> - 2010-08-09 15:13:58
|
Hello All, I hope that you are doing well today. I have abandoned the FreeNX idea over a much simpler one for our particular need. It uses XRDP which can be installed from apt-get (also from http://xrdp.sourceforge.net/) and as most our connection users will be coming from a Windows machine, this will make things easier. So far it works well, but I have to adjust the keymap still. Now I am looking at the disk drives that are located on each node for mapping into a single filespace so that all space appears to be a single drive to the users. For this, I was reading that OpenSSI appears to be using DRBD (if I understand correctly) which allows it to mount the filespace on each node. If this is so then I am guessing that perhaps I could use LVM to bring all of the drives together into a volume. If that is not possible then perhaps I am utilize a stackable files system from FUSE like XtreemFS, gfs, or some other. Any ideas? I am also going to be adding RAID 1 to the system so that the main drive has a complete mirror and failover incase the main drive goes down then the secondary will pick up. Thanks and have a great day, Lonnie Cumberland, Prof. Physicist National Institute of Standards and Technology Ionizing Radiation Division (846) Radiation Physics Group (245), Room C106 ADDRESS: 100 Bureau Drive, Stop 8462 Gaithersburg, MD 20899-8462 EMAIL: lon...@ni... http://physics.nist.gov/Divisions/Div846/div846.html |
|
From: John H. <john@Calva.COM> - 2010-08-10 16:31:12
|
Cumberland, Lonnie wrote: > Now I am looking at the disk drives that are located on each node for mapping into a single filespace so that all space appears to be a single drive to the users. For this, I was reading that OpenSSI appears to be using DRBD (if I understand correctly) which allows it to mount the filespace on each node. If this is so then I am guessing that perhaps I could use LVM to bring all of the drives together into a volume. > > If that is not possible then perhaps I am utilize a stackable files system from FUSE like XtreemFS, gfs, or some other. Any ideas? > > I am also going to be adding RAID 1 to the system so that the main drive has a complete mirror and failover incase the main drive goes down then the secondary will pick up. > Your problem with linux raid and lvm wil be making sure that things get failed over from node to node in the event of a node crash. This works: use DRDB to mirror between the nodes use CFS to make the filesystem on the DRBD device available to all the nodes. This also works: Use shared disk hardware (SAS, SCSI, Fibre Channel, whatever) to make the disks available to all the nodes. Use CFS on the disks to make the filesystem to all the nodes. This has worked in the past, I have no practical experience: Use a cluster aware filesystem (Lustre, even NFS) and mount the filesystem on all the nodes. |
|
From: John H. <john@Calva.COM> - 2010-08-11 11:11:43
|
Some explanation of how OpenSSI interacts with filesystems and so on. When talking about using various filesystems with OpenSSI it's worth keeping two ideas in mind: 1. Cross node access 2. Failover. By "cross node access" I mean the ability for a process on one node to access files on another node. This is good for making the cluster look like one machine and necessary for migration of processes from one node to another. Cross node access can be made possible in two ways: 1. CFS - the I/O requests to a filesystem are handled on behalf on all nodes by a server node. 2. parallel mounts - the filesystem is mounted on all nodes. CFS is easy, the node that mounts the actual Linux filesystem stacks a CFS layer on top of it, all other nodes send their I/O requests to the CFS server. Parallel mounts needs a "physical" filesystem that can be accessed by multiple nodes. A simple example is NFS - each OpenSSI node directly accesses the remote NFS server. More complicated examples of parallel mounts are Lustre and so on. CFS Failover is necessary if we want to use CFS in a fault tolerant cluster. If the CFS server node goes down some other node has to take over its job. In order for this to work the other node needs to have physical access to the disks the filesystem was stored on - either by having used DRBD to make the data available to both the CFS primary and secondary nodes, or by actually having a physical path from both nodes to the disks (SAS, SCSI, iSCSI, FC or whatever). Also the filesystem under the CFS mount needs to be journaled, or the CFS failover will be forced to wait for a fsck before the filesystem is available again. It's failover that makes handling things like RAID and LVM exciting. Both the primary and secondary node need to access the RAID/LVM setup, but you they need to co-ordinate this access very carefully. for LVM there is CLVM (cluster LVM) which could probably be ported to OpenSSI. For RAID you'd need to modify OpenSSI to activate the RAID volumes on the secondary node during the failover. It should be possible, but it's not going to be an easy job. I'd really spend some time with the basic system, trying various failure scenarios, seeing how things work before taking on a big job like this. |
|
From: John H. <john@Calva.COM> - 2010-08-11 11:17:07
|
Cumberland, Lonnie wrote: > Now I am looking at the disk drives that are located on each node for mapping into a single filespace so that all space appears to be a single drive to the users. You can't really do this. What you could do is have each node mount its disk on a different directory - possibly under /cluster/nodeX/ - then you'd have a whole mass of space available to all nodes. The downside would be that each node would then be a single point of failure. You could reduce the chances of failure by using pairs of nodes, setting up DRBD between pairs, then making failover CFS configurations. Better have a lot of bandwidth between the nodes. |
|
From: Cumberland, L. <lon...@ni...> - 2010-08-11 12:04:07
|
Hi John, Thanks for getting back to me and I also liked you post on the OpenSSI filesystem as well. In the other post, I noticed that you mentioned " Linux filesystem stacks a CFS layer " which I assume that OpenSSI is using FUSE or FiST to stack the filesystem, but I had problem to install the "fuse-tools" which seems to be some problem with the udev as I know that we will have single-point failure with this topology, but will also have a redundant backup over the nodes as well using this approach. Additionally, I have the cluster Interconnect on a Gigabit network so that local bandwidth is not really a problem for the moment. http://romanrm.ru/en/mhddfs http://svn.uvw.ru/mhddfs/trunk/README I was trying to load it into the main server but seem to get errors and as being afraid of crashing the cluster, any ideas on what I should do to install the Fuse-Tools so that I can mount stackable filesystems? ------------------------------------------------- spartan:/mnt# apt-get install mhddfs Reading package lists... Done Building dependency tree Reading state information... Done The following extra packages will be installed: fuse-utils The following NEW packages will be installed: fuse-utils mhddfs 0 upgraded, 2 newly installed, 0 to remove and 2 not upgraded. Need to get 0B/38.2kB of archives. After this operation, 213kB of additional disk space will be used. Do you want to continue [Y/n]? WARNING: The following packages cannot be authenticated! fuse-utils mhddfs Authentication warning overridden. Selecting previously deselected package fuse-utils. (Reading database ... 102736 files and directories currently installed.) Unpacking fuse-utils (from .../fuse-utils_2.7.4-1.1+lenny1_i386.deb) ... Selecting previously deselected package mhddfs. Unpacking mhddfs (from .../mhddfs_0.1.12-1_i386.deb) ... Processing triggers for man-db ... Setting up fuse-utils (2.7.4-1.1+lenny1) ... creating fuse group... udev active, skipping device node creation. invoke-rc.d: WARNING: Service udev has no entry in rc.nodeinfo invoke-rc.d: Starting only on initnode Usage: /etc/init.d/udev {start|stop|restart|force-reload} invoke-rc.d: initscript udev, action "reload" failed. dpkg: error processing fuse-utils (--configure): subprocess post-installation script returned error exit status 1 dpkg: dependency problems prevent configuration of mhddfs: mhddfs depends on fuse-utils; however: Package fuse-utils is not configured yet. dpkg: error processing mhddfs (--configure): dependency problems - leaving unconfigured Errors were encountered while processing: fuse-utils mhddfs E: Sub-process /usr/bin/dpkg returned an error code (1) spartan:/mnt# ------------------------------------------------- I will also be working to get DRBD installed and running so that I can establish a good failover and use pairs of nodes in the CFS configuration. Thanks and have a great day, Lonnie Cumberland, Prof. Physicist > -----Original Message----- > From: John Hughes [mailto:john@Calva.COM] > Sent: Wednesday, August 11, 2010 7:19 AM > To: Cumberland, Lonnie > Cc: Scott Walters; Openssi users > Subject: Re: [SSI-users] Debian Lenny OpenSSI with LVM and RAID 1 > > Cumberland, Lonnie wrote: > > Now I am looking at the disk drives that are located on each node for > mapping into a single filespace so that all space appears to be a > single drive to the users. > You can't really do this. > > What you could do is have each node mount its disk on a different > directory - possibly under /cluster/nodeX/ - then you'd have a whole > mass of space available to all nodes. > > The downside would be that each node would then be a single point of > failure. > > You could reduce the chances of failure by using pairs of nodes, > setting > up DRBD between pairs, then making failover CFS configurations. > > Better have a lot of bandwidth between the nodes. |
|
From: Cumberland, L. <lon...@ni...> - 2010-08-11 13:08:30
|
Hello All, Yesterday, I edited the fstab on the main node and added the swap points and local drive mounts for each of my test nodes. Today, I rebooted the system and found that the swaps and mounts from the nodes were not loaded. Am I doing something wrong here to mount the swaps and node drives? I could mount with the simple command Onnode x mount /dev/sda2 /mnt/mnt02 And Swapon Previously, John mentioned: ".... > > What you could do is have each node mount its disk on a different > directory - possibly under /cluster/nodeX/ - then you'd have a whole > mass of space available to all nodes. > > The downside would be that each node would then be a single point of > failure. " Which is what I want to do, but now I am not sure how if the fstab is not working for some reason. Thanks and have a great day, Lonnie Cumberland, Prof. Physicist |
|
From: John H. <john@Calva.COM> - 2010-08-11 14:21:41
|
Cumberland, Lonnie wrote: > Hello All, > > Yesterday, I edited the fstab on the main node and added the swap points and local drive mounts for each of my test nodes. > > Today, I rebooted the system and found that the swaps and mounts from the nodes were not loaded. > It should work, assuming the fstab is set up right. For example I have: # /etc/fstab: static file system information. # # <file system> <mount point> <type> <options> <dump> <pass> proc /proc proc defaults,node=* 0 0 UUID=9e7eb8c6-a7ae-4db6-a8c6-da037e298ca0 / ext3 chard,errors=remount-ro,node=1 0 1 /dev/hda1 /boot ext3 defaults,node=1 0 2 /dev/sda2 none swap sw,node=10:11 0 0 /dev/hda2 none swap sw,node=1:3:4:5:6:7:8:9 0 0 and on boot all the nodes get their swap space (some on their IDE drive, /dev/hda2, some on their SATA drive, /dev/sda2). /boot gets mounted on node 1. |
Thanks for helping keep SourceForge clean.
X