[OpenAFS] OpenAFS 1.3.78 on Linux 2.6

2005-02-04 Thread Jason McCormick
First off, the 1.3.78 release is working VERY well for us. It's nice not to have to re-spin the kernel RPMs to export the syscall table. OpenAFS 1.3.78 also seems to have fixed a number of issues: * Our cache consistency problems seem to have been solved * Kernel panics under high I/O no

Re: [OpenAFS] OpenAFS 1.3.78 on Linux 2.6

2005-02-04 Thread Derrick J Brashear
On Fri, 4 Feb 2005, Jason McCormick wrote: First off, the 1.3.78 release is working VERY well for us. It's nice not to have to re-spin the kernel RPMs to export the syscall table. OpenAFS 1.3.78 also seems to have fixed a number of issues: * Our cache consistency problems seem to have been

Re: [OpenAFS] MacOSX with reliable AFS homedirs?

2005-02-04 Thread Everette Gray Allen
Has anyone gotten Krb5, ldap, and AFS homedirs working reliably? Yes I have had MacOS X with kerberos, openldap, and OpenAFS home dirs working for 2 full semesters here at NC State. I have had kerberos, OpenAFS Homes and some other kinds of directory service working for 2 years before that.

[OpenAFS] Oops in 2.6.11-rc1 ...

2005-02-04 Thread Kevin Coffman
... with citi NFS patches This happens immediately after starting up and cd /afs/citi.umich.edu. OpenAFS is cvs head as of yesterday. Anyone have an idea what I might have done wrong? Kernel is using Redhat FC3 config. configured with: ../openafs-cvs-head/configure '--enable-supergroups'

Re: [OpenAFS] Oops in 2.6.11-rc1 ...

2005-02-04 Thread Derrick J Brashear
Several variables. Using fakestat? Try without? On Fri, 4 Feb 2005, Kevin Coffman wrote: ... with citi NFS patches This happens immediately after starting up and cd /afs/citi.umich.edu. OpenAFS is cvs head as of yesterday. Anyone have an idea what I might have done wrong? Kernel is using Redhat

Re: [OpenAFS] Oops in 2.6.11-rc1 ...

2005-02-04 Thread chas williams - CONTRACTOR
In message [EMAIL PROTECTED],Kevin Coffman writes: Feb 3 18:16:15 rock kernel: EIP is at generic_delete_inode+0x2d/0x310 looks like 2.6.11 has a new inode entity called i_sb_list. it seems to be handled inside new_inode(), so we would need to do the same in afs_NewVCache(). i guess you could

Re: [OpenAFS] MacOSX with reliable AFS homedirs?

2005-02-04 Thread John Koyle
Troy Benjegerdes wrote: Has anyone gotten Krb5, ldap, and AFS homedirs working reliably? We've had to resort to setting up each individual users with a startup items script to run aklog. I've tried the 'kfm_aklog' plugin, but it doesn't seem to work, and none of the apple login hook stuff seems to

[OpenAFS] client hangs with 1.2.13 redhat 9 servers

2005-02-04 Thread Joe Buehler
We recently upgraded to 1.2.13 on our Redhat 9 fileservers and are now seeing performance problems on various clients. Operations on files will sometimes work just fine, other times they will hang for upwards of 30 seconds or so. I've seen this for 1.3.72 windows clients, a 1.2.10 Solaris client,

Re: [OpenAFS] Oops in 2.6.11-rc1 ...

2005-02-04 Thread Kevin Coffman
This does the trick. Thanks! In message [EMAIL PROTECTED],Kevin Coffman writes: Feb 3 18:16:15 rock kernel: EIP is at generic_delete_inode+0x2d/0x310 looks like 2.6.11 has a new inode entity called i_sb_list. it seems to be handled inside new_inode(), so we would need to do the same in

Re: [OpenAFS] MacOSX with reliable AFS homedirs?

2005-02-04 Thread Troy Benjegerdes
Also, I assume you know that you have to add login_logout_notification= aklog to the [libdefaults] section in /Library/Preferences/edu.mit.kerberos (If you have nat clients might also want to add noaddresses = true ) And I assume the kerberos is required for login in

[OpenAFS] Re: client hangs with 1.2.13 redhat 9 servers

2005-02-04 Thread Joe Buehler
A further point of information. Using tcpdump on the main two fileservers I see a lot of udp checksum errors. One of the machines has fragment reassembly timeouts, which probably explains the AFS delays. I am a bit puzzled by tcpdump claiming errors in the UDP checksum for *outgoing* packets on

Re: [OpenAFS] Re: client hangs with 1.2.13 redhat 9 servers

2005-02-04 Thread Jeffrey Hutzelman
On Friday, February 04, 2005 01:58:06 PM -0500 Joe Buehler [EMAIL PROTECTED] wrote: A further point of information. Using tcpdump on the main two fileservers I see a lot of udp checksum errors. One of the machines has fragment reassembly timeouts, which probably explains the AFS delays. I am a

[OpenAFS] kaserver core dumps (openafs-1.3.78, x86_64, FC3)

2005-02-04 Thread RTC
Hi, list! I am having some trouble with openafs-1.3.78. Simply put: The kaserver crashes with coredumps (bosserver happily spawns kaservers). Any idea? Regards, Arne config info: ./configure --with-afs-sysname=amd64_linux26 \ --with-linux-kernel-headers=/lib

Re: [OpenAFS] fine-grained incrementals?

2005-02-04 Thread Jeffrey B. Woodward
John S. Bucy wrote: On Wed, Feb 02, 2005 at 06:51:13PM -0500, Jeffrey B. Woodward wrote: Have you considered using my suggestion (private email reply) of trying Subversion (SVN) to manage the binary diffs and to attain a proper version management system at the same time? I don't have a

Re: [OpenAFS] kaserver core dumps (openafs-1.3.78, x86_64, FC3)

2005-02-04 Thread Jeffrey Hutzelman
On Friday, February 04, 2005 12:26:10 PM -0800 Arne Dietrich [EMAIL PROTECTED] wrote: Hi, list! I am having some trouble with openafs-1.3.78. Simply put: The kaserver crashes with coredumps (bosserver happily spawns kaservers). Hm. I bet no one has tried to run dbservers on x86_64 before. I

Re: [OpenAFS] kaserver core dumps (openafs-1.3.78, x86_64, FC3)

2005-02-04 Thread Jeffrey Hutzelman
On Friday, February 04, 2005 12:26:10 PM -0800 Arne Dietrich [EMAIL PROTECTED] wrote: Hi, list! I am having some trouble with openafs-1.3.78. Simply put: The kaserver crashes with coredumps (bosserver happily spawns kaservers). Any idea? Try adding #include afs/afsutil.h somewhere near the top

[OpenAFS] Using CPLEX for AFS server balancing

2005-02-04 Thread Russ Allbery
Hello folks, At various AFS get-togethers over the years, I've mentioned the AFS server balancing method that Neil Crellin developed here at Stanford that phrases the balancing problem as a linear programming problem and then uses a commercial linear programming optimizer to find a solution.