[OpenAFS] kernel crash linux-4.4.15 openafs-1.6.18.2
Dear All, 10 days ago I have reported a problem with OpenAFS on Linux to openafs-bugs (#133467). Now I have found more results and a solution to circumvent this problem. I have a simple program executing the following calls again and again: fopen afs-file mode append fwrite 64MiB fclose The machine has a disk cache with 16GiB and afsd is started with -afsdb -blocks 14862564 -stat 65536 -volumes 128 -chunksize 18 The first 40 64MiB writes took about 0.7s each, then the time to append 64MiB is 1.2s for block #41 and is growing to 11s for write #154, then I run a "fs flushall" and the write time drops to 0.8s, but is growing again. When I kill the process, I get a kernel crash most of the time. When I use afsd -dcache 32768, I see the following: for the first 128 writes (=128*64MiB=32768*chunksize) the time for each write is growing from 0.7s to 1.5s, then it jumps to 6.5s and is growing to ~10s for write #154 and ~19s for write #220 For our production machines I have set the chunksize to 2^21 and use the default dcache size (1?). Now we do not see any problems with our current jobs. It seems that the algorithm to look for free dcache entries is very slow. Any hints? Best regards, Hans-Werner smime.p7s Description: S/MIME Cryptographic Signature
Re: [OpenAFS] Re: accessing R/O volume becomes slow
On 11/27/2014 01:11 PM, Stephan Wiesand wrote: On 27 Nov 2014, at 11:26, Hans-Werner Paulsen wrote: Yesterday, on another machine I created and deleted 4 million files on AFS. The number of afs_inode_cache slabs grew from 1 million to 5 million. Today there are still 5 million entries. It should shrink when there's memory pressure. If you're still worried, there's the -disable-dynamic-vcaches switch for afsd. On my desktop PC (Linux 3.16.5 x86_64, OpenAFS 1.6.10) I set the -disable-dynamic-vcaches option, the -stat option has a value of 65536. When I create 100,000 files, I see 100,000 more afs_inode_cache slab objects. But, the fileserver is seeing this option, there are only 65253 nFEs, 65253 nCBs (4194304 nblks). Without -disable-dynamic-vcaches the number of CBs is about the number of created files. And if I try to create more files than nCBs on the fileserver, the fileserver (dafileserver) hangs for about 15 minutes (dafileserver 100-120% cpu!), and I get a "connection timeout" on the client. ___ OpenAFS-info mailing list OpenAFS-info@openafs.org https://lists.openafs.org/mailman/listinfo/openafs-info
Re: [OpenAFS] Re: accessing R/O volume becomes slow
On 11/26/2014 09:15 PM, Andrew Deason wrote: On Wed, 26 Nov 2014 10:51:00 +0100 Hans-Werner Paulsen wrote: Checking the machine I see more than 5 million of afs_inode_cache slab entries. Is this normal? Any hint how to proceed? That's not unusual if you are accessing a lot of files (say, about 5 million recently accessed). But having a lot of vcaches in memory can cause certain operations to be slow; there was a fix just added in 1.6.10 to improve speed for a background cleanup process with lots of files (well, and PAGs): 94f1d4. Yesterday, on another machine I created and deleted 4 million files on AFS. The number of afs_inode_cache slabs grew from 1 million to 5 million. Today there are still 5 million entries. ___ OpenAFS-info mailing list OpenAFS-info@openafs.org https://lists.openafs.org/mailman/listinfo/openafs-info
Re: [OpenAFS] accessing R/O volume becomes slow
On 11/26/2014 07:54 PM, Benjamin Kaduk wrote: On Wed, 26 Nov 2014, Hans-Werner Paulsen wrote: Hello, this is on Linux 3.14.8 x86_64, and OpenAFS 1.6.9. The machine is running normally for several months, and then accessing a specific R/O volume (e.g. ls -lR ) becomes slow. Accessing the R/W version of this volume works normally. Accessing other R/O volumes, which have the same size and number of files, works normally. Accessing the R/O version of the problem volume from other clients works normally. The command "fs flushall" does not solve the problem. The (second) "ls -lR" command needs 10 seconds on the R/O, and 2 seconds on the R/W version of this volume. Accessing the R/O version from other fileservers (using fs setserverprefs) does not change anything. Checking the machine I see more than 5 million of afs_inode_cache slab entries. Is this normal? Any hint how to proceed? Are accesses from the same client to other volumes on the same (slow) fileserver slow or fast? Maybe you are getting throttled for too many failed RPCs... I have 2 volumes A and B (similiar in size and number of files) with R/W on machine X, and R/O on machine X, Y and Z. With "fs setserverprefs" I am accessing only the fileserver on X. On the problem machine the first "ls -lR" takes about 1 to 2 minutes. The following "ls -lR"s take 4 to 5 seconds on A(R/W), B(R/W), B(R/O) and 15 seconds on A(R/O). On another client machine all four values are identical and much better ( 1.5 second). ___ OpenAFS-info mailing list OpenAFS-info@openafs.org https://lists.openafs.org/mailman/listinfo/openafs-info
Re: [OpenAFS] accessing R/O volume becomes slow
On 11/26/2014 11:07 AM, Jan Iven wrote: On 11/26/2014 10:51 AM, Hans-Werner Paulsen wrote: Hello, this is on Linux 3.14.8 x86_64, and OpenAFS 1.6.9. The machine is running normally for several months, and then accessing a specific R/O volume (e.g. ls -lR ) becomes slow. Accessing the R/W version of this volume works normally. Accessing other R/O volumes, which have the same size and number of files, works normally. Accessing the R/O version of the problem volume from other clients works normally. The command "fs flushall" does not solve the problem. The (second) "ls -lR" command needs 10 seconds on the R/O, and 2 seconds on the R/W version of this volume. Accessing the R/O version from other fileservers (using fs setserverprefs) does not change anything. Checking the machine I see more than 5 million of afs_inode_cache slab entries. Is this normal? Any hint how to proceed? suggest to check for local disk errors (incl SMART reallocation) on the partition hosting your AFS cache. Cheers jan Thank you for your help, but: No, there are no local disk errors. We had the same problem (but did not further analyse it) a few months ago. Rebooting solved that. And if there is a failure with the local disk, we should see slow performance with other volumes, too. HW ___ OpenAFS-info mailing list OpenAFS-info@openafs.org https://lists.openafs.org/mailman/listinfo/openafs-info
[OpenAFS] accessing R/O volume becomes slow
Hello, this is on Linux 3.14.8 x86_64, and OpenAFS 1.6.9. The machine is running normally for several months, and then accessing a specific R/O volume (e.g. ls -lR ) becomes slow. Accessing the R/W version of this volume works normally. Accessing other R/O volumes, which have the same size and number of files, works normally. Accessing the R/O version of the problem volume from other clients works normally. The command "fs flushall" does not solve the problem. The (second) "ls -lR" command needs 10 seconds on the R/O, and 2 seconds on the R/W version of this volume. Accessing the R/O version from other fileservers (using fs setserverprefs) does not change anything. Checking the machine I see more than 5 million of afs_inode_cache slab entries. Is this normal? Any hint how to proceed? Best regards, HW ___ OpenAFS-info mailing list OpenAFS-info@openafs.org https://lists.openafs.org/mailman/listinfo/openafs-info
Re: [OpenAFS] Re: No buffer space available
On 07/23/2013 09:09 PM, Andrew Deason wrote: On Tue, 23 Jul 2013 11:40:53 -0500 Andrew Deason wrote: On Tue, 23 Jul 2013 13:11:17 +0200 Hans-Werner Paulsen wrote: sometimes creating a file (using different programs) fails with the error message "No buffer space available". This is on amd64_linux26 with OpenAFS 1.6.2 (both client and server). Any idea? I don't see anywhere we'd be generating that error code (ENOBUFS), and I can't see how it would show up that way if we got it back from a socket or something. Do you have any idea if the machine is under a lot of load or memory pressure, or if the directory has a lot of entries in it? Oh, and if you wanted a "next step" involving less guesswork, get an fstrace dump if you can reproduce the problem. That's something like: fstrace clear cm fstrace setlog cmfx -buffers 1024 fstrace sets cm -active # do something to trigger the error, then: fstrace dump cm > /tmp/whatever fstrace sets cm -inactive Providing that to a developer will let us see a lower level trace of what's going on, and should say where that error is coming from. But it can contain some sensitive information like filenames. Unfortunately I cannot reproduce the problem. Thank you for the usage instructions of fstrace, anyway. ___ OpenAFS-info mailing list OpenAFS-info@openafs.org https://lists.openafs.org/mailman/listinfo/openafs-info
Re: [OpenAFS] Re: No buffer space available
On 07/23/2013 06:40 PM, Andrew Deason wrote: On Tue, 23 Jul 2013 13:11:17 +0200 Hans-Werner Paulsen wrote: sometimes creating a file (using different programs) fails with the error message "No buffer space available". This is on amd64_linux26 with OpenAFS 1.6.2 (both client and server). Any idea? I don't see anywhere we'd be generating that error code (ENOBUFS), and I can't see how it would show up that way if we got it back from a socket or something. Do you have any idea if the machine is under a lot of load or memory pressure, or if the directory has a lot of entries in it? The client was not under heavy load, but the fileserver, and the network. ___ OpenAFS-info mailing list OpenAFS-info@openafs.org https://lists.openafs.org/mailman/listinfo/openafs-info
[OpenAFS] No buffer space available
Hello, sometimes creating a file (using different programs) fails with the error message "No buffer space available". This is on amd64_linux26 with OpenAFS 1.6.2 (both client and server). Any idea? Best regards, HW -- Hans-Werner Paulsen h...@mpa-garching.mpg.de MPI für Astrophysik Tel 089-3-2602 Karl-Schwarzschild-Str. 1 Fax 089-3-2235 D-85741 Garching ___ OpenAFS-info mailing list OpenAFS-info@openafs.org https://lists.openafs.org/mailman/listinfo/openafs-info
Re: [OpenAFS] Re: Mixing 1.4 and 1.6 fileservers
On Mon, Sep 24, 2012 at 12:14:52PM -0500, Andrew Deason wrote: > ... You can mix any fileserver versions in a cell, and any client > version can use any fileserver version, etc etc. If the servers are > running database server processes (vlserver, ptserver, etc) then mixing > versions is not supported, but should still work with those versions. What does happen, when I run 1.4.14 and 1.6.1 db servers simultaneously? Is it save to upgrade one db server while the db servers with the old versions are still running, when I do NOT modify the databases during the upgrade process? Best regards, HW -- Hans-Werner Paulsen h...@mpa-garching.mpg.de MPI für Astrophysik Tel 089-3-2602 Karl-Schwarzschild-Str. 1 Fax 089-3-2235 D-85741 Garching ___ OpenAFS-info mailing list OpenAFS-info@openafs.org https://lists.openafs.org/mailman/listinfo/openafs-info
Re: [OpenAFS] sqlite on AFS will not work, even with whole-file locking
On Thu, Apr 22, 2010 at 11:19:57AM +0100, Simon Wilkinson wrote: > Okay, I understand now. And you're right, this is somewhat strange > behaviour, which has been there for years. And it won't help in the > cooperative locking case, sadly. > > When a file is opened RW, and is marked as being dirty, the Unix cache > manager stores all of that files details locally - it won't update > stat information in response to callback breaks, nor will it flush > pages that are already in memory, or invalidate chunks on disk. It > does this in an attempt to prevent locally made changes from being > overwritten by those on the server (because AFS is write-on-close, and > our conflict resolution strategy is last-closer-wins). If you are using "flock" to coordinate access to AFS files from programs running on different machines, it is necessary to get an up-to-date copy from the fileserver as last step of the "lock" call. And you have to flush the local cache to the fileserver before you "unlock" the file. When this is the first step of the flock(LOCK_UN) call, existing programs using "flock" should work without modifications. Best regards, HW -- Hans-Werner Paulsen h...@mpa-garching.mpg.de MPI für Astrophysik Tel 089-3-2602 Karl-Schwarzschild-Str. 1 Fax 089-3-2235 D-85741 Garching ___ OpenAFS-info mailing list OpenAFS-info@openafs.org https://lists.openafs.org/mailman/listinfo/openafs-info
Re: [OpenAFS] sqlite on AFS will not work, even with whole-file locking
On Wed, Apr 21, 2010 at 09:48:27AM -0400, Derrick Brashear wrote: > if you opened it O_RDWR on this client, it better have a valid callback. > > if it's modified and you still have it open, the callback is broken. > if the client doesn't refetch, it's a bug, and it has nothing to do > with locking particularly. I do not know the exact semantics of the AFS filesystem, and therefore I do not know that it is a bug. Is it really a bug? Running the following program on machine A fd = open("xxx",O_RDONLY); while (1) { ret = fstat(fd,&buf); printf("size: %d mtime: %s",(int)buf.st_size,ctime(&(buf.st_mtime))); sleep(1); } and modifying "xxx" on machine B (e.g. echo "J" >>xxx) reports me the up-to-date information on machine A. But, when I open the file with fd = open("xxx",O_RDWR); the stat-information is never updated. This is on i386_linux26 with 1.4.12, but I have seen this behavior forever. Best regards, HW -- Hans-Werner Paulsen h...@mpa-garching.mpg.de MPI für Astrophysik Tel 089-3-2602 Karl-Schwarzschild-Str. 1 Fax 089-3-2235 D-85741 Garching ___ OpenAFS-info mailing list OpenAFS-info@openafs.org https://lists.openafs.org/mailman/listinfo/openafs-info
Re: [OpenAFS] sqlite on AFS will not work, even with whole-file locking
On Wed, Apr 21, 2010 at 08:46:54AM -0400, Derrick Brashear wrote: > if you have a valid callback, the file better be up to date. uh Hm, I do not understand. I have the following code on one client: (1) fd = open("afs-file",O_RDWR) (2) flock(fd,LOCK_EX) ... When the file is modified on the fileserver after (1) and before (2) the copy on the client is NOT up to date (the file is opened O_RDWR). HW -- Hans-Werner Paulsen h...@mpa-garching.mpg.de MPI für Astrophysik Tel 089-3-2602 Karl-Schwarzschild-Str. 1 Fax 089-3-2235 D-85741 Garching ___ OpenAFS-info mailing list OpenAFS-info@openafs.org https://lists.openafs.org/mailman/listinfo/openafs-info
Re: [OpenAFS] sqlite on AFS will not work, even with whole-file locking
On Wed, Apr 21, 2010 at 12:49:39PM +0100, Simon Wilkinson wrote: > >When there are two processes (on different machines) executing that > >code, the (2) flock call has to update the local copy of the afs-file, > >otherwise locking is useless. And the (3) flock call has to sync the > >local copy with the fileserver. > >Writing a small test program I see that this synchronization isn't > >done. > >How can I use the flock(2) call on AFS files? > > Are you saying that the locks don't make it to the fileserver (so two > processes on different machines can flock() the same file). Or that > the file isn't flushed to the server when it is unlocked, so the > second machine doesn't see the changes that the first machine has made? > The second one. To be honest, today I checked only, that the local copy (the cache) is not updated from the fileserver on "flock(fd,LOCK_EX)". HW -- Hans-Werner Paulsen h...@mpa-garching.mpg.de MPI für Astrophysik Tel 089-3-2602 Karl-Schwarzschild-Str. 1 Fax 089-3-2235 D-85741 Garching ___ OpenAFS-info mailing list OpenAFS-info@openafs.org https://lists.openafs.org/mailman/listinfo/openafs-info
Re: [OpenAFS] sqlite on AFS will not work, even with whole-file locking
On Mon, Apr 12, 2010 at 12:34:23AM -0400, Derrick Brashear wrote: > On Sun, Apr 11, 2010 at 11:13 PM, Adam Megacz wrote: > > > > Brandon Simmons writes: > >> Thanks for the response. It seems like whole-file locking in sqlite > >> would be a good choice for me in any case, > > > >> In a situation where the whole-file locking scheme is used, would AFS > >> be an acceptable choice? Would it be better than NFS? > > > > I had the same idea, and tried it. It does not work. Your databases > > will get corrupted. I never figured out why, although I did confirm > > that sqlite was in fact requesting only whole-file locks. > > > > It would be nice if it worked, though. There are a lot of applications > > out there where writes to the database are extremely rare, so > > invalidating all the clients' caches is not a problem. > > do you happen to know what the corruption looked like (blocks of > zeroes, just not readable, something else) > > -- > Derrick On 27 Oct 2008 I had a question about flock on AFS, because I did not understand how flock should work on AFS at all. May I repeat this question: Hello, today I am totally confused how the flock(2) call should work on AFS files. Normally locking works in the following way: 1fd = open("afs-file",O_RDWR) do something 2flock(fd,LOCK_EX) do something with "afs-file" 3flock(fd,LOCK_UN) do something 4close(fd) When there are two processes (on different machines) executing that code, the (2) flock call has to update the local copy of the afs-file, otherwise locking is useless. And the (3) flock call has to sync the local copy with the fileserver. Writing a small test program I see that this synchronization isn't done. How can I use the flock(2) call on AFS files? Thank you for any help, -- Hans-Werner Paulsen h...@mpa-garching.mpg.de MPI für Astrophysik Tel 089-3-2602 Karl-Schwarzschild-Str. 1 Fax 089-3-2235 D-85741 Garching ___ OpenAFS-info mailing list OpenAFS-info@openafs.org https://lists.openafs.org/mailman/listinfo/openafs-info
[OpenAFS] Re: [OpenAFS-announce] OpenAFS 1.4.12 release candidate 4 available
Dear all, on i386_linux26 with linux-2.6.32.9 and gcc-4.4.3 I compiled openafs-1.4.12pre4 without any problems. The client runs fine for about 2 weeks. Server not tested. Kind regards, Hans-Werner Paulsen -- Hans-Werner Paulsen h...@mpa-garching.mpg.de MPI für Astrophysik Tel 089-3-2602 Karl-Schwarzschild-Str. 1 Fax 089-3-2235 D-85741 Garching ___ OpenAFS-info mailing list OpenAFS-info@openafs.org https://lists.openafs.org/mailman/listinfo/openafs-info
[OpenAFS] flock on AFS files
Hello, today I am totally confused how the flock(2) call should work on AFS files. Normally locking works in the following way: 1fd = open("afs-file",O_RDWR) do something 2flock(fd,LOCK_EX) do something with "afs-file" 3flock(fd,LOCK_UN) do something 4close(fd) When there are two processes (on different machines) executing that code, the (2) flock call has to update the local copy of the afs-file, otherwise locking is useless. And the (3) flock call has to sync the local copy with the fileserver. Writing a small test program I see that this synchronization isn't done. How can I use the flock(2) call on AFS files? Thank you for any help, HW -- Hans-Werner Paulsen [EMAIL PROTECTED] MPI für Astrophysik Tel 089-3-2602 Karl-Schwarzschild-Str. 1 Fax 089-3-2235 D-85741 Garching ___ OpenAFS-info mailing list OpenAFS-info@openafs.org https://lists.openafs.org/mailman/listinfo/openafs-info
[OpenAFS] 1.4.7pre2+patch success on i386
Hello, I added the following modifications to OpenAFS-1.4.7pre2 reported on the list: --- openafs-1.4.7pre2/src/afs/LINUX/osi_vnodeops.c.orig Wed Mar 26 05:17:32 2008 +++ openafs-1.4.7pre2/src/afs/LINUX/osi_vnodeops.c Thu Apr 3 14:30:36 2008 @@ -570,8 +570,10 @@ AFS_GLOCK(); -if (fp->f_flags | O_RDONLY) /* readers dont flush */ +if ((fp->f_flags & O_ACCMODE) == O_RDONLY) { +AFS_GUNLOCK(); return 0; +} credp = crref(); vcp = VTOAFS(FILE_INODE(fp)); and the OpenAFS client does not show any problems with linux-2.6.24.4 gcc-4.2.3 on a i386 system. HW -- Hans-Werner Paulsen [EMAIL PROTECTED] MPI für Astrophysik Tel 089-3-2602 Karl-Schwarzschild-Str. 1 Fax 089-3-2235 D-85741 Garching ___ OpenAFS-info mailing list OpenAFS-info@openafs.org https://lists.openafs.org/mailman/listinfo/openafs-info
[OpenAFS] Re: [OpenAFS-announce] OpenAFS 1.4.5 release candidate 1 available
On Sun, Oct 14, 2007 at 10:16:07PM -0400, Derrick J Brashear wrote: > Notable changes include updates for newer Linux kernels (including > 2.6.23) ... Does not compile on i386 and x86_64 with kernel 2.6.23. I think that linux-2623-support-20071004 is missing. With this patch applied I was able to compile OpenAFS and start the client on i386 and x86_64. HW -- Hans-Werner Paulsen [EMAIL PROTECTED] MPI für Astrophysik Tel 089-3-2602 Karl-Schwarzschild-Str. 1 Fax 089-3-2235 D-85741 Garching ___ OpenAFS-info mailing list OpenAFS-info@openafs.org https://lists.openafs.org/mailman/listinfo/openafs-info
Re: [OpenAFS] Strange access problems on one client
On Sun, Oct 07, 2007 at 01:15:00PM -0400, Marc Dionne wrote: > Anyone care to test out the attached patch for src/dir/dir.c > > The hashing code in the DirHash() function relies on integer overflow to > make the hval value turn into a negative value. gcc 4.2 assumes that > this value can never go negative and optimizes out the (hval < 0) test. > > Marc > > diff -u -r1.24 dir.c > --- src/dir/dir.c 13 Oct 2005 15:12:12 - 1.24 > +++ src/dir/dir.c 7 Oct 2007 17:10:37 - > @@ -478,8 +478,9 @@ > { > /* Hash a string to a number between 0 and NHASHENT. */ > register unsigned char tc; > -register int hval; > +unsigned long hval; > register int tval; > + > hval = 0; > while ((tc = (*string++))) { > hval *= 173; > @@ -488,7 +489,7 @@ > tval = hval & (NHASHENT - 1); > if (tval == 0) > return tval; > -else if (hval < 0) > +else if (hval >= 1<<31) > tval = NHASHENT - tval; > return tval; > } this patch is fine for architectures where the size of "unsigned long" is 4 bytes. But on the x86_64 architecture this will not work, because the size is 8 bytes. One can use "unsigned int". HW -- Hans-Werner Paulsen [EMAIL PROTECTED] MPI für Astrophysik Tel 089-3-2602 Karl-Schwarzschild-Str. 1 Fax 089-3-2235 D-85741 Garching ___ OpenAFS-info mailing list OpenAFS-info@openafs.org https://lists.openafs.org/mailman/listinfo/openafs-info
Re: [OpenAFS] gcc-4.2.1, afs-client not working
On Tue, Oct 09, 2007 at 12:23:44PM +0200, Hans-Werner Paulsen wrote: > blabla... Sorry, this problem is "Strange access problems on one client" and already solved. HW -- Hans-Werner Paulsen [EMAIL PROTECTED] MPI für Astrophysik Tel 089-3-2602 Karl-Schwarzschild-Str. 1 Fax 089-3-2235 D-85741 Garching ___ OpenAFS-info mailing list OpenAFS-info@openafs.org https://lists.openafs.org/mailman/listinfo/openafs-info
Re: [OpenAFS] gcc-4.2.1, afs-client not working
On Fri, Oct 05, 2007 at 10:12:10AM -0400, Derrick Brashear wrote: > If you can tell us which file or files when compiled with 4.2.1 are screwing > up, it would make it easier to be able to figure out what's going on. Of course, but I did not understand how the different Makefiles work to compile the kernel module, and how to specify different compiler flags for different source files. Now I have some results: Garance A Drosihn suggested to try the gcc option -fno-tree-vrp, and yes the kernel module compiled with this option runs fine. In addition only the source file src/libafs/MODLOAD-2.6.22.9-MP/afs_dir.c, which is src/dir/dir.c, needs to be compiled with this option. HW -- Hans-Werner Paulsen [EMAIL PROTECTED] MPI für Astrophysik Tel 089-3-2602 Karl-Schwarzschild-Str. 1 Fax 089-3-2235 D-85741 Garching ___ OpenAFS-info mailing list OpenAFS-info@openafs.org https://lists.openafs.org/mailman/listinfo/openafs-info
[OpenAFS] gcc-4.2.1, afs-client not working
Hello, on i386_linux26 I compiled the kernel 2.6.22.9 and OpenAFS 1.4.4 using the gcc 4.2.1. Now I get: === # ls -l /afs ls: cannot access /afs/mpa-garching.mpg.de: No such file or directory ls: cannot access /afs/ipp-garching.mpg.de: No such file or directory ls: cannot access /afs/world: No such file or directory ls: cannot access /afs/rzg.mpg.de: No such file or directory ls: cannot access /afs/andrew.cmu.edu: No such file or directory total 14 d? ? ?? ?? andrew.cmu.edu drwxr-xr-x 11 root root 6144 Sep 21 07:18 ipp ?? ? ?? ?? ipp-garching.mpg.de drwxr-xr-x 3 root root 2048 Mar 30 2007 mpa ?? ? ?? ?? mpa-garching.mpg.de drwxr-xr-x 11 root root 6144 Sep 21 07:18 rzg d? ? ?? ?? rzg.mpg.de ?? ? ?? ?? world === When I recompiled the OpenAFS software using gcc 4.1.2 everything is fine. Any idea or help? Hans-Werner -- Hans-Werner Paulsen [EMAIL PROTECTED] MPI für Astrophysik Tel 089-3-2602 Karl-Schwarzschild-Str. 1 Fax 089-3-2235 D-85741 Garching ___ OpenAFS-info mailing list OpenAFS-info@openafs.org https://lists.openafs.org/mailman/listinfo/openafs-info
Re: [OpenAFS] Maximum size /vicepX partition
On Tue, Jan 24, 2006 at 02:59:00PM +0100, Horst Birthelmer wrote: > On Jan 24, 2006, at 2:26 PM, Hans-Werner Paulsen wrote: > >the size of a /vicepX partition on an OpenAFS-1.4.0 fileserver seems > >to be limited to 4TByte. > > Are you sure, it's the fileserver and not your partitions filesystem > that is limited? > > Can you write data to the partition directly on the fileserver? > I mean writing directly on the filesystem of the fileserver (in this > case /vicepd). Yes. ls -l /vicepd total 0 drwx--2 root root 72 Jan 24 12:43 AFSIDat drwx--2 root root 72 Jan 24 12:40 Lock HW -- Hans-Werner Paulsen [EMAIL PROTECTED] MPI für Astrophysik Tel 089-3-2602 Karl-Schwarzschild-Str. 1 Fax 089-3-2235 D-85741 Garching ___ OpenAFS-info mailing list OpenAFS-info@openafs.org https://lists.openafs.org/mailman/listinfo/openafs-info
[OpenAFS] Maximum size /vicepX partition
Hello, the size of a /vicepX partition on an OpenAFS-1.4.0 fileserver seems to be limited to 4TByte. I can create volumes on this fileserver/partition, but when I try to create a file on this volume I get: $ touch xxx touch: creating `xxx': No space left on device The fileserver writes the following lines to the log file: Tue Jan 24 12:40:10 2006 File Server started Tue Jan 24 12:40:10 2006 Tue Jan 24 12:42:53 2006 Partition /vicepd that contains volume 536920876 is full Tue Jan 24 12:44:41 2006 Shutting down file server at Tue Jan 24 12:44:41 2006 Tue Jan 24 12:44:41 2006 Partition /vicepd: -5451620 available 1K blocks (minfree=0), Tue Jan 24 12:44:41 2006 overallocated by 10008912 blocks ... -- Hans-Werner Paulsen [EMAIL PROTECTED] MPI für Astrophysik Tel 089-3-2602 Karl-Schwarzschild-Str. 1 Fax 089-3-2235 D-85741 Garching ___ OpenAFS-info mailing list OpenAFS-info@openafs.org https://lists.openafs.org/mailman/listinfo/openafs-info
[OpenAFS] fileserver monitor
Hello, one of our OpenAFS fileserver is very busy, and I want to check who is doing the top I/O. Neither "afsmonitor" nor "scout" can tell me who (client) is using which data (volume). Are there any other tools around to identify the top usage? HW -- Hans-Werner Paulsen [EMAIL PROTECTED] MPI für Astrophysik Tel 089-3-2602 Karl-Schwarzschild-Str. 1 Fax 089-3-2235 D-85741 Garching ___ OpenAFS-info mailing list OpenAFS-info@openafs.org https://lists.openafs.org/mailman/listinfo/openafs-info
Re: [OpenAFS] Callback/Cache Issues with 1.3.82 on FC3
On Wed, May 04, 2005 at 08:41:19AM +0200, Stephan Wiesand wrote: > On Tue, 3 May 2005, chas williams - CONTRACTOR wrote: > > >i cant seem to duplicate this failure. deleting on the 1.2 client > >makes the file disappear on the 1.3 client as well. do you have > >a little more info? cache types for the clients? We have the same problem, one machine 1.3.81, the other one 1.3.82: 1.3.82:~ >echo a > a 1.3.82:~ >cat a a 1.3.81:~ >cat a a 1.3.81:~ >rm a 1.3.81:~ >cat a cat: a: No such file or directory 1.3.82:~ >cat a a On both machines we are using identical software, compiled using gcc-3.3.2, only OpenAFS versions are different. linux 2.6.11.7 glibc 2.3.2 cache ext2 -- Hans-Werner Paulsen [EMAIL PROTECTED] MPI für Astrophysik Tel 089-3-2602 Karl-Schwarzschild-Str. 1 Fax 089-3-2235 D-85741 Garching ___ OpenAFS-info mailing list OpenAFS-info@openafs.org https://lists.openafs.org/mailman/listinfo/openafs-info
Re: [OpenAFS] Re: Problem with openafs-1.3.81 on kernel 2.6.11.7
On Fri, Apr 15, 2005 at 12:02:39PM +0100, Dr A V Le Blanc wrote: >... > rebooted the system and gone into afs and done a > > for i in `find . -type f -noleaf`;do cat $i >/dev/null;done > > The cache partition quickly gets overfull and input/output errors > appear. This looks like a real problem. I am now using the > Debian packakes for 1.3.81, but previously I compiled openafs > from source and had the same problem, so it does not appear to > be specific to the Debian packages. > > Is anyone else running 1.3.81 on 2.6.11 of some form, who is willing > to try the same test? > I have the same problem with my machine: linux-2.6.11.5 openafs-1.3.81 gcc-3.3.2 glibc-2.3.2 This is a self-made system using unmodified sources. The three commands "df", "du" and "fs getcache" are showing different results (unit=1k) df: 669132 of 1035692 fs: 432785 of 50 (I reduced the max size to avoid problems) du: 2520660 (!!!) And the cache partition is ext2. -- Hans-Werner Paulsen [EMAIL PROTECTED] MPI für Astrophysik Tel 089-3-2602 Karl-Schwarzschild-Str. 1 Fax 089-3-2235 D-85741 Garching ___ OpenAFS-info mailing list OpenAFS-info@openafs.org https://lists.openafs.org/mailman/listinfo/openafs-info
Re: [OpenAFS] Fileserver Problem with OpenAFS 1.3.73 on Debian Sarge
On Tue, Nov 09, 2004 at 11:44:48AM -0500, Derrick J Brashear wrote: > On Tue, 9 Nov 2004, Lars Schimmer wrote: > > >It seems to wirk correct, but now the problems appears: > >VLLog on the backup database server: > >Tue Nov 9 12:55:24 2004 ubik: A Remote Server has addresses: @(#) OpenAFS > >1.2.11 built 2004-01-11 > >~ [127.0.0.1 134.169.37.177] > >~ It will replace the following existing entry in the VLDB (same uuid): > >~ entry 4: [134.169.37.177] > > vos changeaddr -remove 127.0.0.1 and put 127.0.0.1 in NetRestrict; this is > a bug which will be fixed in 1.3.74 The problem is NOT fixed in version 1.3.74. I checked this on i386_linux24. I was not able to remove the localhost entry with this command: vos changeaddr -remove 127.0.0.1 Could not remove server 127.0.0.1 from the VLDB vlserver does not support the remove flag or VLDB: no such entry But creating the NetRestrict file, and restarting the fileserver/volserver removes this entry. -- Hans-Werner Paulsen [EMAIL PROTECTED] MPI für Astrophysik Tel 089-3-2602 Karl-Schwarzschild-Str. 1 Fax 089-3-2235 D-85741 Garching ___ OpenAFS-info mailing list [EMAIL PROTECTED] https://lists.openafs.org/mailman/listinfo/openafs-info
Re: [OpenAFS] Fileserver Problem with OpenAFS 1.3.73 on Debian Sarge
On Tue, Nov 09, 2004 at 01:18:44PM +0100, Lars Schimmer wrote: > I used the debian experimental packages of OpenAFS 1.3.73, > openafs-client - The AFS distributed filesystem- client support > openafs-fileserver - The AFS distributed filesystem- file server > openafs-modules-source - The AFS distributed filesystem- Module Sources > buitl by Maintainer: Sam Hartman <[EMAIL PROTECTED]> > > It seems to wirk correct, but now the problems appears: > VLLog on the backup database server: > Tue Nov 9 12:55:24 2004 ubik: A Remote Server has addresses: @(#) OpenAFS > 1.2.11 built 2004-01-11 > ~ [127.0.0.1 134.169.37.177] > ~ It will replace the following existing entry in the VLDB (same uuid): > ~ entry 4: [134.169.37.177] > > Hell, what went wrong? > Yes, the fileserver sees my server as 127.0.0.1, but why? Hello, on an amd64_linux24 system with OpenAFS 1.3.73 I have the same problem. 127.0.0.1 is added to "sysid" and send to the VLDB server. Creating a NetRestrict file with this IP address does solve this problem. Hans-Werner -- Hans-Werner Paulsen [EMAIL PROTECTED] MPI für Astrophysik Tel 089-3-2602 Karl-Schwarzschild-Str. 1 Fax 089-3-2235 D-85741 Garching ___ OpenAFS-info mailing list [EMAIL PROTECTED] https://lists.openafs.org/mailman/listinfo/openafs-info