[SLUG] Strange NFS server wanting to make connections from port 2049 to 910 on client
Hi I'm getting an interesting message is my firewall logs on my NFS server. The NFS server 10.0.0.1 is blocking these packets outgoing DENY IN= OUT=eth0 SRC=10.0.0.1 DST=10.0.0.100 LEN=1236 TOS=0x00 PREC=0x00 TTL=64 ID=34527 DF PROTO=TCP SPT=2049 DPT=910 WINDOW=2003 RES=0x00 ACK PSH URGP=0 why? The NFS client 10.0.0.100 is still able to access and read from the nfs shares. The client is running Ubuntu Gutsy and the server Ubuntu Feisty One the server rpcinfo -p program vers proto port 102 tcp111 portmapper 102 udp111 portmapper 1000241 udp 32765 status 1000241 tcp 32765 status 132 udp 2049 nfs 133 udp 2049 nfs 134 udp 2049 nfs 1000211 udp 32768 nlockmgr 1000213 udp 32768 nlockmgr 1000214 udp 32768 nlockmgr 1000211 tcp 32768 nlockmgr 1000213 tcp 32768 nlockmgr 1000214 tcp 32768 nlockmgr 132 tcp 2049 nfs 133 tcp 2049 nfs 134 tcp 2049 nfs 151 udp 32767 mountd 151 tcp 32767 mountd 152 udp 32767 mountd 152 tcp 32767 mountd 153 udp 32767 mountd 153 tcp 32767 mountd The ports are all set so they will not change when I restart nfs and can be very strict with my iptables rules created with fwbuilder. -- Steve __ This email has been scanned by the MessageLabs Email Security System. For more information please visit http://www.messagelabs.com/email __ -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] Reminder to reply all.
Hi Tony, Blindraven wrote: .. snip .. Just a heads up to everyone who has been replying to me re: my Gentoo post, and in general because it keeps occurring. If you forget to hit reply all you are sending emails directly to the person and not the list. The whole purpose of the list is that everyone gets to read through, share their opinions and people can kind of sift through stuff and hopefully find something that is useful to them, I know I have on many occasions. What actually led me to this email was the pure extensiveness and effort that has gone in to the replies I have received. Not only do these people miss out on the unsaid list credit of helping, others will miss out on the help I was lucky enough to receive. I'm ranting, in short.* Reply All. ..snip.. I'm guessing people avoided replying to all to avoid one of these http://ubuntuforums.org/archive/index.php/t-231240.html by not offending Gentoo and non-Gentoo distro users. I'm of similar experience to yourself and got put into the deep end when I had to maintain and then upgrade/recreate a Gentoo server. You could install it with UML (user-mode-linux) on your Ubuntu system to start off with to begin with as it will make you famliar with install procedure. My best advice to you is to go at Gentoo alone and follow their wonderful documentation , Stage3 install, http://www.gentoo.org/doc/en/handbook/ by meticulously reading through every line of it to get your system installed and up and running to experience what Gentoo is truly about. This process can be considered a wonderful learning process which makes you more intimate with the workings of Linux system with many hours of compiling or a pain staking process which is filled with annoying gotcha's, many hours of What/Where-TF? and tedious hours of compiling. My experience was half way in between, YMMV. ;) Make sure you read Working with Portage section of the handbook to get your head around the package management system. Eventually you'll end up spending a lot of your time lurking in this forum Portage Programming http://forums.gentoo.org/viewforum-f-8.html?sid=2e886ee6b2040aafb246c6036b26919b The support forums for Gentoo are amazing being on par with, maybe slightly better that Ubuntu's. Gentoo taught me a lot in regards to Systems Administration * Package upgrades need to be well planned, tested on dummy systems, will most likely require downtime and require a rollback solution * Read the release notes for the software packages you are looking to upgrade (e.g. Samba, Apache) as old configuration files may not work with the new software * Package/Build servers are a great option if you are looking to maintain the performance and uptime of live machines. As a Debian/Ubuntu user (Desktop/laptop) I now run a gentoo server at home as it makes me go about things the hard way (I feel) and makes me very aware of any discrepancies I haven't taken into account. I (and you might of) had the impression that Gentoo was running all of the latest software compared to the ancient Debian and not quite so ancient Ubuntu... this is not the case as the latest version of certain software can be masked, listed but kind of unavailable for Ubuntu users mindsets, due to bugs and incomplete testing so you still may have the feeling that you are being held back somewhat by the distribution. Your better off learning how to build your own packages/repositories and working them into the distribution rather than changing distro's. This has made appreciate Ubuntu's and especially Debian's packaging time lines as I will know what I'm getting has a very good chance of working together with everything else. -- Steve __ This email has been scanned by the MessageLabs Email Security System. For more information please visit http://www.messagelabs.com/email __ -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] Windows and Ubuntu
It's a bug (found with The Google, 'ubuntu nvidia version module') Discussion in forums http://ubuntuforums.org/showthread.php?t=471882 Bug https://bugs.launchpad.net/ubuntu/+source/linux-restricted-modules-2.6.20/+bug/106217 remove the hidden file /lib/linux-restricted-modules/.nvidia_new_installed and then install the appropriate nvidia-glx package Just replying to this thread. Also relates to 'Nvidia problem - The continuing saga' -- Steve Heracles wrote: Thanks for all the help so far everyone but still no joy. I have stuffed my daughters Ubuntu trying to set up the nvidia drivers. I can get them to work if I load them using the NVIDIA binary and then use startx without rebooting. When I reboot X wont start and it tells me there is a driver mismatch. I have removed all of the possible conflicting drivers that I can find and edited the /etc/modules file to remove all reference to nvidia as Jeff suggested. Still no joy! I know I can start the machine and let X fail - run sudo NVIDIA-.. and then login each time the machine is restarted, but this seems a tedious way to use it. She needs to use windows for her games and linux for the internet so reboots are not uncommon. I must be missing something. But what??? NOTE: I have a very similar card (same GeForce model - 6200 - but different maker and it worked perfectly when I changed drivers. This is why I used the same procedure on hers. Heracles __ This email has been scanned by the MessageLabs Email Security System. For more information please visit http://www.messagelabs.com/email __ -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] troubles installing mediawiki1.7 on Debian Etch - extensions?
Hi Jeremy Jeremy Visser wrote: Steve Granger wrote: ref name=LVM Howto{{cite web | url = http://www.tldp.org/HOWTO/LVM-HOWTO/index.html | title = LVM HOWTO | date=[[November 27]], [[2006]] | accessdate = 2007-06-15}}/ref {{cite web}} is a template used on Wikipedia. http://en.wikipedia.org/wiki/Template:Cite_web You most likely need to copy the code from here http://en.wikipedia.org/w/index.php?title=Template:Cite_webaction=edit onto your wiki. that's exactly what I needed ! -- Steve __ This email has been scanned by the MessageLabs Email Security System. For more information please visit http://www.messagelabs.com/email __ -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] troubles installing mediawiki1.7 on Debian Etch - extensions?
On a related issue, I've installed the mediawiki-extensions, linked the available extensions in /etc/mediawiki-extensions/extensions-available to /etc/mediawiki-extensions/extensions-enabled /etc/mediawiki-extensions/extensions-enabled# ls -la total 0 drwxr-xr-x 2 root root 232 2007-06-15 15:23 . drwxr-xr-x 4 root root 160 2007-06-15 13:49 .. lrwxrwxrwx 1 root root 32 2007-06-15 15:22 Cite.php - ../extensions-available/Cite.php lrwxrwxrwx 1 root root 33 2007-06-15 15:22 geshi.php - ../extensions-available/geshi.php lrwxrwxrwx 1 root root 36 2007-06-15 15:22 inputbox.php - ../extensions-available/inputbox.php lrwxrwxrwx 1 root root 39 2007-06-15 15:22 NewestPages.php - ../extensions-available/NewestPages.php lrwxrwxrwx 1 root root 32 2007-06-15 15:23 poem.php - ../extensions-available/poem.php lrwxrwxrwx 1 root root 48 2007-06-15 15:23 SpecialLastUserLogin.php - ../extensions-available/SpecialLastUserLogin.php but when I try to reference something in the wiki ref name=LVM Howto{{cite web | url = http://www.tldp.org/HOWTO/LVM-HOWTO/index.html | title = LVM HOWTO | date=[[November 27]], [[2006]] | accessdate = 2007-06-15}}/ref it just displays ref name=LVM HowtoTemplate:Cite web/ref any pointers? I might then be able to make off an item in the TODO list /usr/share/doc/mediawiki-extensions/TODO * Add documentation for this extensions - Cite.php - geshi.php - poem.php - SpecialLastUserLogin.php * Add documentation about ext enabling -- Steve __ This email has been scanned by the MessageLabs Email Security System. For more information please visit http://www.messagelabs.com/email __ -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] mdadm
Hi David, Removing mdadm is okay, no hotplug devices use it and it only required if you have a software RAID device/setup. From the bug report that you posted, this is a new feature that gets you to make sure the mdadm.conf file is the same as the out of a mdadm scan. This causes the failure of your initrd creation as initrd builds depend on mdadm being happy (because you can include modules related to mdadm to be built into your initrd to bring up hardware devices before mdadm tries to create md devices) before they are built. -- Steve (sorry, just hit reply before thinking.. must be from my intense three day session on forgetting all about work) :-} david wrote: I'm getting error messages about mdadm. There are suggestions that this can be removed if not using RAID (I don't). [1] Are there any other issues with removing it? EG: hotplug devices like USB storage devices etc. Google isn't really making this clear. thanks... David p[1] https://bugs.launchpad.net/ubuntu/+source/initramfs-tools/+bug/98911 __ This email has been scanned by the MessageLabs Email Security System. For more information please visit http://www.messagelabs.com/email __ -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] mdadm
david wrote: On Tue, 2007-06-12 at 17:11 +1000, Steve Granger wrote: Hi David, Removing mdadm is okay, no hotplug devices use it and it only required if you have a software RAID device/setup. From the bug report that you posted, this is a new feature that gets you to make sure the mdadm.conf file is the same as the out of a mdadm scan. This causes the failure of your initrd creation as initrd builds depend on mdadm being happy (because you can include modules related to mdadm to be built into your initrd to bring up hardware devices before mdadm tries to create md devices) before they are built. yes... I removed, but then re-installed because of all the nasty warnings I was getting. I'm not entirely sure where I'm at now - not quite game to re-boot just yet. Nothing is simple :( So you've removed mdadm? what happens when you run dpkg-configure -a? does the install process finish? run update-initramfs -k all -u to get your initrd to build again, it should do this without complaining now. Seems like a few people in the bug thread ignored the error, rebooted and didn't have their systems come up... Good way to learn how to use a rescue CD... -- Steve (sorry, just hit reply before thinking.. must be from my intense three day session on forgetting all about work) :-} david wrote: I'm getting error messages about mdadm. There are suggestions that this can be removed if not using RAID (I don't). [1] Are there any other issues with removing it? EG: hotplug devices like USB storage devices etc. Google isn't really making this clear. thanks... David p[1] https://bugs.launchpad.net/ubuntu/+source/initramfs-tools/+bug/98911 __ This email has been scanned by the MessageLabs Email Security System. For more information please visit http://www.messagelabs.com/email __ __ This email has been scanned by the MessageLabs Email Security System. For more information please visit http://www.messagelabs.com/email __ -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] mdadm
... snip update-initramfs -k all -u Lots of warnings (see below). I sort of think these are not problems, because I don't have any raid devices but I'm still bothered that I might reboot and die! Some of the messages are pretty cryptic if you don't actually know. God help a beginner. Maybe I shouldn't worry so much - just re-boot and hope. I really did read /usr/share/doc/mdadm/README.upgrading-2.5.3.gz as suggested, but I'm not a lot wiser. At least I've got a vague idea what initramfs IS now. I still can't guess what md is administering.Multi Device? so you have rm -f /var/lib/mdadm/CONF-UNCHECKED got rid of the errors for me.. on a system running software raid, not that it should matter. [EMAIL PROTECTED]:~/test $ sudo update-initramfs -k all -u Password: update-initramfs: Generating /boot/initrd.img-2.6.20-16-386 W: mdadm: unchecked configuration file: /etc/mdadm/mdadm.conf W: mdadm: please read /usr/share/doc/mdadm/README.upgrading-2.5.3.gz . W: mdadm: no arrays defined in configuration file. W: mdadm: falling back to emergency procedure in initramfs. update-initramfs: Generating /boot/initrd.img-2.6.20-15-386 W: mdadm: unchecked configuration file: /etc/mdadm/mdadm.conf W: mdadm: please read /usr/share/doc/mdadm/README.upgrading-2.5.3.gz . W: mdadm: no arrays defined in configuration file. W: mdadm: falling back to emergency procedure in initramfs. update-initramfs: Generating /boot/initrd.img-2.6.17-10-386 W: mdadm: unchecked configuration file: /etc/mdadm/mdadm.conf W: mdadm: please read /usr/share/doc/mdadm/README.upgrading-2.5.3.gz . W: mdadm: no arrays defined in configuration file. W: mdadm: falling back to emergency procedure in initramfs. update-initramfs: Generating /boot/initrd.img-2.6.15-23-386 W: udev hook script requires at least kernel version 2.6.17 W: not generating requested initramfs for kernel 2.6.15-23-386 ... snip -- Steve __ This email has been scanned by the MessageLabs Email Security System. For more information please visit http://www.messagelabs.com/email __ -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] Debian Sid USB/SCSI Drive
.. snip The USB drive is a 250GB drive thats sole purpose is for samba sharing that doesn't require backups (for example, freely distributable programs from the net) Not in the array.. okay :) I've got a similar set up in that I have 2 sata hard drives and a Dell 2405 monitor plugged in which names all the flash drive slots /dev/sd*, a usb drive and usb key drive plugged in. Udev makes life a lot more predictable. Its less then ideal to use sda as a usb drive for obvious reasons, and I am wondering what is the best way to fix this? udevinfo -a -p /sys/block/sdb/sdd Will list the attributes for the drive here's my usbdisk udev rules /etc/udev/rules.d/95-usbdisk.rules ATTR{size}==586072368, ATTRS{model}==00JB-00KFA0 , ATTRS{vendor}==WDC WD30, NAME{all_partitions}=usbdisk My concern isn't about removal and insertion of the disk on the running OS, its that sometimes it may be present on bootup, and other times it may not. So therefore the SCSI drives will be assigned different letters if the usb drive is not present, which would degrade the array, as they would now be sda, b and c, and not sdb, c and d. On looking at your udev rules, could I apply this to both udev on the OS, and udev on initrd, and instead of the USB drive being assigned /dev/sda, it will be assigned /dev/usbdisk, and the other SCSI disks would be assigned /dev/sda, b and c? I'm thinking you will need to set it up for for the udev on the OS and initrd. If you built your kernel with make-kpkg --initrd it's easy to update the initrd with update-initramfs -k all -u Only just picked up on 'update-initramfs' after working with Debian for 5 years... Thanks again. Scott Got a bit carried away with the software raid stuff... it's been my bread and butter for the last two weeks and seems to have to implanted itself into my subconscious and gets triggered the instance I here raid, disk, linux. -- Steve __ This email has been scanned by the MessageLabs Email Security System. For more information please visit http://www.messagelabs.com/email __ -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] Debian Sid USB/SCSI Drive
Hi Scott, Scott Ragen wrote: Hi Guys, I recently upgraded our server to Sid the latest .deb kernel, and I have a problem that I'm not sure of the best way to tackle. We have 3 scsi disks in a software raid, and an external usb drive. Since the upgrade of the kernel, the USB drive has become sda, with the scsi drives being sdb,c and d. What was the old device name of the USB drive? (I'll assume /dev/sdd) Do you have only the one md device? (I'll assume /dev/md0) What level of raid do you have? (I'll assume raid 5 seeing you have so many devices) Its less then ideal to use sda as a usb drive for obvious reasons, and I am wondering what is the best way to fix this? You'll need to remove the usbdisk from the array and re-add it as another device. You can assign the device name through the use of udev rules. For an Ubuntu/Debian scenario regarding udev and usb drives see http://www.debian-administration.org/articles/126 udevinfo -a -p /sys/block/sdb/sdd Will list the attributes for the drive here's my usbdisk udev rules /etc/udev/rules.d/95-usbdisk.rules ATTR{size}==586072368, ATTRS{model}==00JB-00KFA0 , ATTRS{vendor}==WDC WD30, NAME{all_partitions}=usbdisk This isn't a very unique set of attributes, if you can find a serial number that would be better. Fail the usbdisk and remove it from the array. This is a bit tricky as from my experience mdadm will still require the failed device to exist in /dev/ for it to be removed and you will still want /dev/sda to exist as one of the SCSI drives in the array. Try shutting down the machine, remove the usb drive, power up the machine, reattach the usb drive and it should come up as /dev/sdd (if not, set it to appear with udev rules), will probably start resync-ing into the array. (off top of head, be sure to double check) mdadm --manage /dev/md0 --fail /dev/sdd mdadm --manage /dev/md0 --remove /dev/sdd set up the udev rules for the usb drive and then add the /dev/usbdisk partition into the array mdadm --manage /dev/md0 --add /dev/usbdisk it should then start doing it's resync watch it resync watch -n1 cat /proc/mdstat Reboot and test and test and test and TEST! and then reboot again to make sure it comes up okay :) I'm still a newbie with linux software raid so I may have missed some points, may need to set the superblock to 0 on the usbdisk before adding it back into the array. Thanks, Scott No worries, -- Steve __ This email has been scanned by the MessageLabs Email Security System. For more information please visit http://www.messagelabs.com/email __ -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html