[opensuse] Co-existing in a dynamic network Howto?
Hello - I could use some advise on how to configure our SuSE systems to work in a dynamic network. I have several computers running SuSE and use fstab to do a number of automounts of file systems that exist on other computers within our network. Trouble is that these other computers may come and go and I do not have control over the exact configuration of the network at any given time. I use fstab to define what I want automounted, so as to be able to run some scripts that access available remote file systems during boot up. (That is one reason why I cannot use the KNetworkManager as it is not active until after boot up has taken place and a second is that these scripts must be ran during the boot up process so as to set up some other server configurations automatically and not wait until a user has logged in.) Trouble is two fold. during boot up, if the automount process finds one computer that does not exist, it times out and seems to give up on mounting any of the rest of the computers in the list. I would like to get around this without having to endure a long timeout for each failed mount attempt also so that the boot up process can be accomplished in a reasonable amount of time. And, after bootup, this timeout problem also seems to plague the SuSE menu (in the kicker bar, i.e. the menus displayed when one clicks on the SuSE logo) When I click on the SuSE menu launcher, it seems to want to do an automount of all the file systems defined in the fstab file and this can result in a horrendously long time out process (for each failed attempt to automount a remote file system) during which the entire kicker becomes unresponsive and frozen. I probably am overlooking some simple configuration setting but so far have not been able to find a method of getting around these two issues. Any help and suggestions would be much appreciated! (Am mostly running SuSE 11.1 with KDE 3.5) Marc... -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
Well I didn't get a reply so guess I will submit my query again and hope some knowledgeable Linux/SuSE guru will notice it this time.. ;-) Thanks in advance for any advice and help offered... Marc..
Hello - I could use some advise on how to configure our SuSE systems to work in a dynamic network. I have several computers running SuSE and use fstab to do a number of automounts of file systems that exist on other computers within our network. Trouble is that these other computers may come and go and I do not have control over the exact configuration of the network at any given time.
I use fstab to define what I want automounted, so as to be able to run some scripts that access available remote file systems during boot up. (That is one reason why I cannot use the KNetworkManager as it is not active until after boot up has taken place and a second is that these scripts must be ran during the boot up process so as to set up some other server configurations automatically and not wait until a user has logged in.)
Trouble is two fold. during boot up, if the automount process finds one computer that does not exist, it times out and seems to give up on mounting any of the rest of the computers in the list. I would like to get around this without having to endure a long timeout for each failed mount attempt also so that the boot up process can be accomplished in a reasonable amount of time. And, after bootup, this timeout problem also seems to plague the SuSE menu (in the kicker bar, i.e. the menus displayed when one clicks on the SuSE logo) When I click on the SuSE menu launcher, it seems to want to do an automount of all the file systems defined in the fstab file and this can result in a horrendously long time out process (for each failed attempt to automount a remote file system) during which the entire kicker becomes unresponsive and frozen.
I probably am overlooking some simple configuration setting but so far have not been able to find a method of getting around these two issues. Any help and suggestions would be much appreciated! (Am mostly running SuSE 11.1 with KDE 3.5)
Marc...
-- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
On Sat, Mar 7, 2009 at 10:27 PM, Marc Chamberlin
Well I didn't get a reply so guess I will submit my query again and hope some knowledgeable Linux/SuSE guru will notice it this time.. ;-) Thanks in advance for any advice and help offered...
Trouble is two fold. during boot up, if the automount process finds one computer that does not exist, it times out and seems to give up on mounting any of the rest of the computers in the list. I would like to get around this without having to endure a long timeout for each failed mount attempt also so that the boot up process can be accomplished in a reasonable amount of time.
Marc: 1st: What Kai said. This is why people invented servers, the ad-hoc network structure just does not work well. 2nd: Assuming you have been forced into this and can't get around it, look at man mount and see if the -F parameter might speed the process. You will still get the wait, but only one wait, as all mounts will be done in parallel. You might also be able to foist the "hotplug" option with the -o option. -- ----------JSA--------- Someone stole my tag line, so now I have this rental. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
John Andersen wrote:
On Sat, Mar 7, 2009 at 10:27 PM, Marc Chamberlin
wrote: Well I didn't get a reply so guess I will submit my query again and hope some knowledgeable Linux/SuSE guru will notice it this time.. ;-) Thanks in advance for any advice and help offered...
Trouble is two fold. during boot up, if the automount process finds one computer that does not exist, it times out and seems to give up on mounting any of the rest of the computers in the list. I would like to get around this without having to endure a long timeout for each failed mount attempt also so that the boot up process can be accomplished in a reasonable amount of time.
Marc: 1st: What Kai said. This is why people invented servers, the ad-hoc network structure just does not work well.
2nd: Assuming you have been forced into this and can't get around it, look at man mount and see if the -F parameter might speed the process. You will still get the wait, but only one wait, as all mounts will be done in parallel.
You might also be able to foist the "hotplug" option with the -o option.
Many thanks John, Kai and Dave for your replies... Kai you probably hit it on the head, a Win's like automount would probably do the trick and yeah I have tried most of the mount options you guys mentioned... What would be really cool is if there was a way to automagically discover all the computers on a network, much like Windoz does, and then I could create a list to work from... Your suggestions however has got me thinking and I think you have put me on a path that might work, at least for Linux/SuSE systems.. (I will worry about Windoz separately....) I am now trying to write a script that will periodically run via a cron job. It will ping and test to see if each computer, on a list of known computers on our network, can be found. Depending on the results, this script will dynamically modify the contents of fstab within a specially comment delineated region. (I DO love sed, grep, vi and emacs macros!! How do Windoz admins survive without such tools?) Upon logout the script will clear this list from the fstab file. So in effect this script will run at boot up, periodically while each computer is active, and on shutdown... Part of the scripts job will also be to update the list of known/valid computers which it can get from a master list on one of the desktops... Hopefully this will also solve the complaint about long delays when someone clicks on the SuSE start menu in the kicker bar. I have discovered, rather empirically, that these delays are caused by the kicker task trying to do a full mount of all the mount points defined in fstab... Kai - to answer your questions, my network consists mostly of laptops that come and go and some desktops. All are dual booted to either Windoz or Linux, users choice. Our DHCP server does know their names and does assign a fixed static internal network address to each one. So that simplifies the problem a lot. Also we have a static file system structure that must be set up on each system so we can work with that. We use software that will allow each user to push files to other computers automatically, when they update a particular file, and not use a pull system, such as what rcs, cvs or svn databases offer.. And yes we manage synchronization issues by allowing user to only update a certain part of the common file structure that they have been given access to... So that is the basis for what I am trying to accomplish. As for network admins... guess who is it? ;-) (I am NOT a Linux guru either, just the best we have so am learning as I go...) Now to figure out where the hooks are to inject my scripts into the boot up and shutdown processes... I think I know - /etc/rc.d looks promising... Marc.. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
Marc Chamberlin wrote: <snip>
Kai - to answer your questions, my network consists mostly of laptops that come and go and some desktops. All are dual booted to either Windoz or Linux, users choice. Our DHCP server does know their names and does assign a fixed static internal network address to each one. So that simplifies the problem a lot. Also we have a static file system structure that must be set up on each system so we can work with that. <snip>
Marc, An even easier way for the dual-boot laptop issue, (if your users are smart enough to maintain the same hostname in Linux and windows) is to configure dhcpd.conf like: subnet 192.168.6.0 netmask 255.255.255.0 { option domain-name "3111skyline.com"; ddns-domainname "3111skyline.com"; option subnet-mask 255.255.255.0; range dynamic-bootp 192.168.6.110 192.168.6.150; default-lease-time 28800; max-lease-time 172800; zone 3111skyline.com. {primary 127.0.0.1; key DHCP_UPDATER; } zone 6.168.192.in-addr.arpa. { primary 127.0.0.1; key DHCP_UPDATER; } # # We want P35a to appear at the same fixed address regardless of # which OS is booted to prevent multiple dhcp leases belonging to the # same hardware address # host P35a.3111skyline.com { hardware ethernet 00:26:f5:17:2d:83; fixed-address 192.168.6.101; } That way no matter which OS they use, the IP dhcpd hands out is always the same for the same hostname/mac address of any given box. -- David C. Rankin, J.D.,P.E. Rankin Law Firm, PLLC 510 Ochiltree Street Nacogdoches, Texas 75961 Telephone: (936) 715-9333 Facsimile: (936) 715-9339 www.rankinlawfirm.com -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Marc Chamberlin wrote:
John Andersen wrote:
Many thanks John, Kai and Dave for your replies... Kai you probably hit it on the head, a Win's like automount would probably do the trick and yeah I have tried most of the mount options you guys mentioned... What would be really cool is if there was a way to automagically discover all the computers on a network, much like Windoz does, and then I could create a list to work from...
I thought that "avahi" was for that, but I don't know how.
Your suggestions however has got me thinking and I think you have put me on a path that might work, at least for Linux/SuSE systems.. (I will worry about Windoz separately....) I am now trying to write a script that will periodically run via a cron job. It will ping and test to see if each computer, on a list of known computers on our network, can be found. Depending on the results, this script will dynamically modify the contents of fstab within a specially comment delineated region.
You could simply have the remote mounts marked "noauto". These should not be mounted automatically during boot, but will have to be manually mounted by a command. You ping, it is up, then mount. Or, instead of automated by cron, another script to mount shares, which first pings the machine, then mounts the share. The user gives the name of the machine or the directory that he wants as a parameter. If the other machine does not respond, give the appropriate message. Mounting manually on request might have the advantage if the user happens to know how long the other machine is going to be up. Mmm... I wonder if machines that are about to be unplugged could broadcast some kind of message so that the users connected to its shares have time to umount safely.
Now to figure out where the hooks are to inject my scripts into the boot up and shutdown processes... I think I know - /etc/rc.d looks promising...
Read the man page and the admin book! SuSE is different in this respect, and simply writing scripts there, and adding links manually, will not work. There is a readme there and an skeleton sample script, too - -- Cheers / Saludos, Carlos E. R. (from 11.1-ex-factory) -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.9 (GNU/Linux) Comment: Using GnuPG with SUSE - http://enigmail.mozdev.org iEYEARECAAYFAkm5FVkACgkQU92UU+smfQU9bgCfUnLpwVyiQIsJMzxzzUg2j1Kr 1VIAn0s6blyTDy2enwEKOHs/4X4pWJcm =qoWs -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
Marc Chamberlin wrote:
Hello - I could use some advise on how to configure our SuSE systems to work in a dynamic network. I have several computers running SuSE and use fstab to do a number of automounts of file systems that exist on other computers within our network. Trouble is that these other computers may come and go and I do not have control over the exact configuration of the network at any given time. So, what you'd like is a mounting solution similar to a WINS-based system?
I use fstab to define what I want automounted, so as to be able to run some scripts that access available remote file systems during boot up. (That is one reason why I cannot use the KNetworkManager as it is not active until after boot up has taken place and a second is that these scripts must be ran during the boot up process so as to set up some other server configurations automatically and not wait until a user has logged in.)
Trouble is two fold. during boot up, if the automount process finds one computer that does not exist, it times out and seems to give up on mounting any of the rest of the computers in the list. I would like to get around this without having to endure a long timeout for each failed mount attempt also so that the boot up process can be accomplished in a reasonable amount of time. And, after bootup, this timeout problem also seems to plague the SuSE menu (in the kicker bar, i.e. the menus displayed when one clicks on the SuSE logo) When I click on the SuSE menu launcher, it seems to want to do an automount of all the file systems defined in the fstab file and this can result in a horrendously long time out process (for each failed attempt to automount a remote file system) during which the entire kicker becomes unresponsive and frozen.
I probably am overlooking some simple configuration setting but so far have not been able to find a method of getting around these two issues. Any help and suggestions would be much appreciated! (Am mostly running SuSE 11.1 with KDE 3.5) It seems like the problem is with your network admins. I've read over your message several times during the past week and cannot understand quite what you expect.
Basically, if you have dependencies on other machines to run scripts, those machines should have static addresses and be "always on" for your systems to access. Thinking back over the years in my experiences with Novell 3.x/4.x systems, Windows NT systems, OS/2 systems and Linux-based systems they all require the remote computer/server to be available at the time it is requested or it goes into some sort of wait state until a timeout is achieved. Since it appears your network staff are reluctant to provide stable computers, then maybe you can run these scrips at a periodic basis instead of rebooting your computers? -- kai www.perfectreign.com | www.ecmplace.com www.twitter.com/PerfectReign -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
2009/3/1 Marc Chamberlin
Hello - I could use some advise on how to configure our SuSE systems to work in a dynamic network. I have several computers running SuSE and use fstab to do a number of automounts of file systems that exist on other computers within our network. Trouble is that these other computers may come and go and I do not have control over the exact configuration of the network at any given time.
I use fstab to define what I want automounted, so as to be able to run some scripts that access available remote file systems during boot up. (That is one reason why I cannot use the KNetworkManager as it is not active until after boot up has taken place and a second is that these scripts must be ran during the boot up process so as to set up some other server configurations automatically and not wait until a user has logged in.)
Trouble is two fold. during boot up, if the automount process finds one computer that does not exist, it times out and seems to give up on mounting any of the rest of the computers in the list. I would like to get around this without having to endure a long timeout for each failed mount attempt also so that the boot up process can be accomplished in a reasonable amount of time. And, after bootup, this timeout problem also seems to plague the SuSE menu (in the kicker bar, i.e. the menus displayed when one clicks on the SuSE logo) When I click on the SuSE menu launcher, it seems to want to do an automount of all the file systems defined in the fstab file and this can result in a horrendously long time out process (for each failed attempt to automount a remote file system) during which the entire kicker becomes unresponsive and frozen.
I probably am overlooking some simple configuration setting but so far have not been able to find a method of getting around these two issues. Any help and suggestions would be much appreciated! (Am mostly running SuSE 11.1 with KDE 3.5)
Marc...
The picture is not quite clear for me.... Also, don't use the term "automounted", just mounted (resources to mount listed in fstab). Given your "dynamic network", you should use autofs instead, that way the resources will be mounted only when you try to access them. Can you elaborate about what you need to run from the mounted FS at boot time?, Did you specify the "_netdev" option in that FS entry?. Would help to understand the big picture, maybe some things can be done in a different/better way.... Regards, -- Ciro Iriarte http://cyruspy.wordpress.com -- -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
participants (6)
-
Carlos E. R.
-
Ciro Iriarte
-
David C. Rankin
-
John Andersen
-
Kai Ponte
-
Marc Chamberlin