Re: [OpenAFS] 1.3.77 Brokenness

[Jack Neely]

Derrick,

That patch works.  OpenAFS seems to work fairly well in memcache mode.
Not seeing file corruption of missing files.

The attached version of the patch (mainly for the list) applies to
1.3.77.  So, I'm using this + Matthew's patches for a working 1.3.77.

Jack
-- 
Jack Neely <[EMAIL PROTECTED]>
Realm Linux Administration and Development
PAMS Computer Operations at NC State University
GPG Fingerprint: 1917 5AC1 E828 9337 7AA4  EA6B 213B 765F 3B6A 5B89
--- src/afsd/afsd.c.old 2005-01-14 13:27:10.557882648 -0500
+++ src/afsd/afsd.c 2005-01-14 13:28:06.141432664 -0500
@@ -1932,7 +1932,11 @@
 if (afsd_debug)
printf("%s: Calling AFSOP_VOLUMEINFO: volume info file is '%s'\n", rn,
   fullpn_VolInfoFile);
-call_syscall(AFSOP_VOLUMEINFO, fullpn_VolInfoFile);
+
+/* once again, meaningless for a memory-based cache. */
+if (!(cacheFlags & AFSCALL_INIT_MEMCACHE))
+   call_syscall(AFSOP_VOLUMEINFO, fullpn_VolInfoFile);
+
 
 /*
  * Pass the kernel the name of the afs logging file holding the volume

Re: [OpenAFS] 1.3.77 Brokenness

[Jack Neely]

I was starting to wonder what that other patch actually did.  :-)

I also did s/printk/printf/ in your patch.

Attached is dmesg output.

Jack Neely

On Wed, Jan 12, 2005 at 04:45:32PM -0500, Derrick J Brashear wrote:
> correct diff.
> 
> and no, once you get an oops, you can't start afs again.

> ? .gdb_history
> ? diff
> ? ppc_darwin_70
> ? src/libafs/afs.ppc_darwin_70.plist
> ? src/packaging/MacOS/OpenAFS.pkg
> ? src/packaging/MacOS/OpenAFS.pkg.tar.gz
> ? src/venus/kdump.c.
> Index: src/afs/afs_memcache.c
> ===
> RCS file: /cvs/openafs/src/afs/afs_memcache.c,v
> retrieving revision 1.16
> diff -u -r1.16 afs_memcache.c
> --- src/afs/afs_memcache.c1 Dec 2004 23:38:56 -   1.16
> +++ src/afs/afs_memcache.c11 Jan 2005 22:13:19 -
> @@ -45,6 +45,7 @@
>   memCacheBlkSize = blkSize;
>  
>  memMaxBlkNumber = blkCount;
> +printk("memMaxBlkNumber %d\n", memMaxBlkNumber);
>  memCache = (struct memCacheEntry *)
>   afs_osi_Alloc(memMaxBlkNumber * sizeof(struct memCacheEntry));
>  if (flags & AFSCALL_INIT_MEMCACHE_SLEEP) {
> @@ -89,7 +90,7 @@
>  return 0;
>  }
>  
> -#if defined(AFS_SUN57_64BIT_ENV) || defined(AFS_SGI62_ENV)
> +#if defined(AFS_SUN57_64BIT_ENV) || defined(AFS_SGI62_ENV) || 
> defined(AFS_LINUX26_ENV)
>  void *
>  afs_MemCacheOpen(ino_t blkno)
>  #else
> @@ -100,6 +101,7 @@
>  struct memCacheEntry *mep;
>  
>  if (blkno < 0 || blkno > memMaxBlkNumber) {
> +  printk("blkno %d\n", blkno);
>   osi_Panic("afs_MemCacheOpen: invalid block #");
>  }
>  mep = (memCache + blkno);
> Index: src/afs/afs_prototypes.h
> ===
> RCS file: /cvs/openafs/src/afs/afs_prototypes.h,v
> retrieving revision 1.57
> diff -u -r1.57 afs_prototypes.h
> --- src/afs/afs_prototypes.h  1 Dec 2004 23:38:56 -   1.57
> +++ src/afs/afs_prototypes.h  11 Jan 2005 22:13:19 -
> @@ -427,7 +427,7 @@
>  /* afs_memcache.c */
>  extern int afs_InitMemCache(int blkCount, int blkSize, int flags);
>  extern int afs_MemCacheClose(struct osi_file *file);
> -#if defined(AFS_SUN57_64BIT_ENV) || defined(AFS_SGI62_ENV)
> +#if defined(AFS_SUN57_64BIT_ENV) || defined(AFS_SGI62_ENV) || 
> defined(AFS_LINUX26_ENV)
>  extern void *afs_MemCacheOpen(ino_t blkno);
>  #else
>  extern void *afs_MemCacheOpen(afs_int32 blkno);
> @@ -590,7 +590,7 @@
>  
>  /* ARCH/osi_file.c */
>  extern int afs_osicred_initialized;
> -#if defined(AFS_SUN57_64BIT_ENV) || defined(AFS_SGI62_ENV)
> +#if defined(AFS_SUN57_64BIT_ENV) || defined(AFS_SGI62_ENV) || 
> defined(AFS_LINUX26_ENV)
>  extern void *osi_UFSOpen(ino_t ainode);
>  #else
>  extern void *osi_UFSOpen(afs_int32 ainode);
> Index: src/afs/LINUX/osi_file.c
> ===
> RCS file: /cvs/openafs/src/afs/LINUX/osi_file.c,v
> retrieving revision 1.22
> diff -u -r1.22 osi_file.c
> --- src/afs/LINUX/osi_file.c  8 Dec 2004 17:21:04 -   1.22
> +++ src/afs/LINUX/osi_file.c  11 Jan 2005 22:13:19 -
> @@ -27,7 +27,11 @@
>  extern struct super_block *afs_cacheSBp;
>  
>  void *
> +#ifdef AFS_LINUX26_ENV
> +osi_UFSOpen(ino_t ainode)
> +#else
>  osi_UFSOpen(afs_int32 ainode)
> +#endif
>  {
>  register struct osi_file *afile = NULL;
>  extern int cacheDiskType;
> 


-- 
Jack Neely <[EMAIL PROTECTED]>
Realm Linux Administration and Development
PAMS Computer Operations at NC State University
GPG Fingerprint: 1917 5AC1 E828 9337 7AA4  EA6B 213B 765F 3B6A 5B89
ng delay loop... 3923.96 BogoMIPS (lpj=1961984)
Security Scaffold v1.0.0 initialized
SELinux:  Initializing.
SELinux:  Starting in permissive mode
There is already a security framework initialized, register_security failed.
selinux_register_security:  Registering secondary module capability
Capability LSM initialized as secondary
Mount-cache hash table entries: 512 (order: 0, 4096 bytes)
CPU: After generic identify, caps: 078bfbff e1d3fbff  
CPU: After vendor identify, caps:  078bfbff e1d3fbff  
CPU: L1 I Cache: 64K (64 bytes/line), D cache 64K (64 bytes/line)
CPU: L2 Cache: 1024K (64 bytes/line)
CPU: After all inits, caps:078bf3ff e1d3fbff  0010
Intel machine check architecture supported.
Intel machine check reporting enabled on CPU#0.
CPU: AMD Opteron(tm) Processor 146 stepping 08
Enabling fast FPU save and restore... done.
Enabling unmasked SIMD FPU exception support... done.
Checking 'hlt' instruction... OK.
ACPI: IRQ9 SCI: Edge set to Level Trigger.
checking if image is initramfs... it is
Freeing initrd memory: 476k freed
NET: Registered protocol family 16
PCI: PCI BIOS revision 2.10 entry at 0xf0031, last bus=1
PCI: Using configuration type 1
mtrr: v2.0 (20020519)
ACPI: Subsystem revision 20040816
ACPI: Interpreter enabled
ACPI: Using PIC for interrupt routing
ACPI: PCI Root Bridge [PCI0] (00:00)
PCI: Probing PCI hardware (bus 00)
ACPI: PCI Interrupt R

Re: [OpenAFS] 1.3.77 Brokenness

[Jack Neely]

Derrick,

No joy...no printk()s either...attached is dmesg where afs was "started"
on boot, and afterwards I tryied to start it again.

Jack

On Tue, Jan 11, 2005 at 05:15:09PM -0500, Derrick J Brashear wrote:
> No clue if it will help, but it's worth a try. if you get an oops, check 
> your messages file, you should have 2 kernel printfs (blkno, and 
> maxMemBlocks or somesuch)

> ? diff
> ? ppc_darwin_70
> ? src/libafs/afs.ppc_darwin_70.plist
> Index: src/afs/afs.h
> ===
> RCS file: /cvs/openafs/src/afs/afs.h,v
> retrieving revision 1.52
> diff -u -r1.52 afs.h
> --- src/afs/afs.h 13 Oct 2004 14:46:29 -  1.52
> +++ src/afs/afs.h 18 Oct 2004 10:24:47 -
> @@ -943,6 +943,7 @@
>  char mflags; /* Meta flags */
>  struct fcache f; /* disk image */
>  afs_int32 stamp; /* used with vtodc struct for hints */
> +afs_int32 bucket;   /* which bucket these dcache entries are in 
> */
>  
>  /*
>   * Locking rules:
> Index: src/afs/afs_dcache.c
> ===
> RCS file: /cvs/openafs/src/afs/afs_dcache.c,v
> retrieving revision 1.44
> diff -u -r1.44 afs_dcache.c
> --- src/afs/afs_dcache.c  13 Oct 2004 00:51:00 -  1.44
> +++ src/afs/afs_dcache.c  18 Oct 2004 10:24:52 -
> @@ -23,14 +23,24 @@
>  #include "afs/afs_osidnlc.h"
>  
>  /* Forward declarations. */
> -static void afs_GetDownD(int anumber, int *aneedSpace);
> +static void afs_GetDownD(int anumber, int *aneedSpace, afs_int32 buckethint);
>  static void afs_FreeDiscardedDCache(void);
>  static void afs_DiscardDCache(struct dcache *);
>  static void afs_FreeDCache(struct dcache *);
> +/* For split cache */
> +static afs_int32 afs_DCGetBucket(struct vcache *);
> +static void afs_DCAdjustSize(struct dcache *, afs_int32, afs_int32);
> +static void afs_DCSizeInit(void);
> +static afs_int32 afs_DCWhichBucket(afs_int32, afs_int32);
>  
>  /*
>   * - Exported definitions -
>   */
> +/* For split cache */
> +afs_int32 afs_blocksUsed_0;  /*1K blocks in cache - in theory is zero */
> +afs_int32 afs_blocksUsed_1;  /*1K blocks in cache */
> +afs_int32 afs_blocksUsed_2;  /*1K blocks in cache */
> +
>  afs_lock_t afs_xdcache;  /*Lock: alloc new disk cache entries */
>  afs_int32 afs_freeDCList;/*Free list for disk cache entries */
>  afs_int32 afs_freeDCCount;   /*Count of elts in freeDCList */
> @@ -120,6 +130,80 @@
>  struct afs_cacheOps *afs_cacheType;
>  
>  
> +static afs_int32
> +afs_DCGetBucket(struct vcache *avc) 
> +{
> +if (!avc->states & CStatd) 
> + printf("getbucket fid %d %d %d state %x\n", avc->fid.Fid.Volume, 
> +avc->fid.Fid.Vnode, avc->fid.Fid.Unique, avc->states);
> +
> +/* This should be replaced with some sort of user configurable function 
> */
> +if (avc->states & CRO) {
> + return 2;
> +} else if (avc->states & CBackup) {
> + return 1;
> +} else {
> + /* RW */
> +}
> +
> +/* main bucket */
> +return 1;
> +}
> +
> +static void 
> +afs_DCAdjustSize(struct dcache *adc, afs_int32 oldSize, afs_int32 newSize)
> +{
> +switch (adc->bucket) 
> +{
> +case 0:
> + afs_blocksUsed_0 += (newSize - oldSize);
> + break;
> +case 1:
> + afs_blocksUsed_1 += (newSize - oldSize);
> + break;
> +case 2:
> + afs_blocksUsed_2 += (newSize - oldSize);
> + break;
> +}
> +
> +return;
> +}
> +
> +static void 
> +afs_DCSizeInit(void) 
> +{
> +afs_blocksUsed_0 = afs_blocksUsed_1 = afs_blocksUsed_2 = 0;
> +}
> +
> +static afs_int32
> +afs_DCWhichBucket(afs_int32 phase, afs_int32 bucket) 
> +{
> +/* Short cut: if we don't know about it, try to kill it */
> +if (afs_blocksUsed_0) 
> + return 0;
> +
> +if (bucket == 1) {
> + /* RW/BK */
> +
> + /* 25% RW/BK */
> + if (((afs_blocksUsed_1/afs_cacheBlocks)*100) > 25) 
> + return 1;
> + /* 75% RO */
> + if (((afs_blocksUsed_2/afs_cacheBlocks)*100) > 75) 
> + return 2;
> +} else if (bucket == 2) {
> + /* RO */
> +
> + /* 25% RW/BK */
> + if (((afs_blocksUsed_1/afs_cacheBlocks)*100) > 25) 
> + return 1;
> + /* 75% RO */
> + if (((afs_blocksUsed_2/afs_cacheBlocks)*100) > 75) 
> + return 2;
> +}
> +
> +return 0;
> +}
>  
>  
>  /*
> @@ -245,7 +329,7 @@
>   afs_blocksUsed - afs_blocksDiscarded - cb_lowat;
>   slots_needed =
>   dc_hiwat - afs_freeDCCount - afs_discardDCCount;
> - afs_GetDownD(slots_needed, &space_needed);
> + afs_GetDownD(slots_needed, &space_needed, 0);
>   if ((space_needed <= 0) && (slots_needed <= 0)) {
>   break;
>   }
> @@ -340,6 +424,7 @@
>  if (!newSize)
>   adc->validPos = 0;
>  newSize = ((newSize + afs_fsfragsize) ^ 

Re: [OpenAFS] 1.3.77 Brokenness

[Lars Schimmer]

-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
Matthew Miller schrieb:
| On Mon, Jan 10, 2005 at 06:43:09PM -0500, Jack Neely wrote:
|
|>I keep RHEL/FC install trees in AFS which are served out via HTTP.  I
|>build the tree on a test machine and use rsync to move it out to AFS
|>land.  In this case my server is a RHEL3 linux box running OpenAFS
|>1.2.11.  After moving the tree with rsync (a little over 2G worth) I
|>see that the cache gets out of sync on large files (20MB, 70MB, etc.)
|>Rsync complains it cannot delete its dot-files (the file it creates
|>before moving it to the real filename).  These . files show in an 'ls'
|>and when I try to rm them I get "No file or directory" yet they still
|>appear in AFS.  Many of my RPM pacakges are corrupt as well.
|
|
| This is *exactly* the problem I've seen. It's reported at
| .
Just for information:
Debian Sarge Kernel 2.6.8 OpenAFS 1.3.76 on AFS Fileserver (with XFS for
/vicepa) and Debian Sarge Kernel 2.4.27 OpenAFS 1.3.76 as client seems not to
have any problems with large files.
Createt on the client a laaarge 8GB .tar file stored on the server and copied
back, no problem. Handle some more big files (dvd-images), no problems til yet.
And yes, diskcache, not memcache.
Cya
Lars
- --
- -
Technische Universität Braunschweig, Institut für Computergraphik
Tel.: +49 531 391-2109E-Mail: [EMAIL PROTECTED]
PGP-Key-ID: 0xB87A0E03
-BEGIN PGP SIGNATURE-
Version: GnuPG v1.2.4 (GNU/Linux)
Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org
iD8DBQFB5PonVguzrLh6DgMRAtyDAJ9y7W2dK0QVMLS6YgVKxPPlzQmWWwCgvIqa
2R783eggCPyC0fNTbyjRN7E=
=jM+U
-END PGP SIGNATURE-
___
OpenAFS-info mailing list
OpenAFS-info@openafs.org
https://lists.openafs.org/mailman/listinfo/openafs-info

Re: [OpenAFS] 1.3.77 Brokenness

[Jack Neely]

Yes.

-- 
Jack Neely <[EMAIL PROTECTED]>
Realm Linux Administration and Development
PAMS Computer Operations at NC State University
GPG Fingerprint: 1917 5AC1 E828 9337 7AA4  EA6B 213B 765F 3B6A 5B89
___
OpenAFS-info mailing list
OpenAFS-info@openafs.org
https://lists.openafs.org/mailman/listinfo/openafs-info

Re: [OpenAFS] 1.3.77 Brokenness

[Derrick J Brashear]

On Tue, 11 Jan 2005, Jack Neely wrote:
And this is what happens when I try to use memcache...
Neat.
#if defined(AFS_SUN57_64BIT_ENV) || defined(AFS_SGI62_ENV)
void *
afs_MemCacheOpen(ino_t blkno)
#else
void *
afs_MemCacheOpen(afs_int32 blkno)
#endif
{
struct memCacheEntry *mep;
if (blkno < 0 || blkno > memMaxBlkNumber) {
osi_Panic("afs_MemCacheOpen: invalid block #");
}
why do i suspect inodes are now a 64 bit type?
if i come up with a patch for this are you willing to try it?
___
OpenAFS-info mailing list
OpenAFS-info@openafs.org
https://lists.openafs.org/mailman/listinfo/openafs-info

Re: [OpenAFS] 1.3.77 Brokenness

[Jack Neely]

And this is what happens when I try to use memcache...

Jack
-- 
Jack Neely <[EMAIL PROTECTED]>
Realm Linux Administration and Development
PAMS Computer Operations at NC State University
GPG Fingerprint: 1917 5AC1 E828 9337 7AA4  EA6B 213B 765F 3B6A 5B89
15)
ACPI: PCI Interrupt Link [LNKB] (IRQs 3 4 5 7 *10 11 14 15)
ACPI: PCI Interrupt Link [LNKC] (IRQs 3 4 *5 7 10 11 14 15)
ACPI: PCI Interrupt Link [LNKD] (IRQs 3 4 5 7 10 11 14 15) *0, disabled.
ACPI: PCI Interrupt Link [LNKE] (IRQs 3 4 5 7 10 11 14 15) *0, disabled.
ACPI: PCI Interrupt Link [LNKF] (IRQs 3 4 5 7 10 11 14 15) *0, disabled.
ACPI: PCI Interrupt Link [LNKG] (IRQs 3 4 5 7 10 11 14 15) *0, disabled.
ACPI: PCI Interrupt Link [LNKH] (IRQs 3 4 5 7 10 11 14 15) *0, disabled.
Linux Plug and Play Support v0.97 (c) Adam Belay
usbcore: registered new driver usbfs
usbcore: registered new driver hub
PCI: Using ACPI for IRQ routing
ACPI: PCI Interrupt Link [LNKA] enabled at IRQ 11
ACPI: PCI interrupt :00:07.0[A] -> GSI 11 (level, low) -> IRQ 11
ACPI: PCI Interrupt Link [LNKC] enabled at IRQ 5
ACPI: PCI interrupt :00:08.0[A] -> GSI 5 (level, low) -> IRQ 5
ACPI: PCI Interrupt Link [LNKB] enabled at IRQ 10
ACPI: PCI interrupt :00:0a.0[A] -> GSI 10 (level, low) -> IRQ 10
ACPI: PCI interrupt :00:0b.0[A] -> GSI 11 (level, low) -> IRQ 11
ACPI: PCI interrupt :00:0f.0[B] -> GSI 10 (level, low) -> IRQ 10
ACPI: PCI interrupt :00:0f.1[A] -> GSI 11 (level, low) -> IRQ 11
ACPI: PCI interrupt :00:10.0[A] -> GSI 11 (level, low) -> IRQ 11
ACPI: PCI interrupt :00:10.1[A] -> GSI 11 (level, low) -> IRQ 11
ACPI: PCI interrupt :00:10.2[B] -> GSI 10 (level, low) -> IRQ 10
ACPI: PCI interrupt :00:10.3[B] -> GSI 10 (level, low) -> IRQ 10
ACPI: PCI interrupt :00:10.4[C] -> GSI 5 (level, low) -> IRQ 5
ACPI: PCI interrupt :00:11.5[C] -> GSI 5 (level, low) -> IRQ 5
ACPI: PCI interrupt :01:00.0[A] -> GSI 11 (level, low) -> IRQ 11
apm: BIOS version 1.2 Flags 0x03 (Driver version 1.16ac)
apm: overridden by ACPI.
audit: initializing netlink socket (disabled)
audit(1105455518.017:0): initialized
highmem bounce pool size: 64 pages
Total HugeTLB memory allocated, 0
VFS: Disk quotas dquot_6.5.1
Dquot-cache hash table entries: 1024 (order 0, 4096 bytes)
SELinux:  Registering netfilter hooks
Initializing Cryptographic API
ksign: Installing public key data
Loading keyring
- Added public key E07BC3E85BE30CFD
- User ID: Red Hat, Inc. (Kernel Module GPG key)
pci_hotplug: PCI Hot Plug PCI Core version: 0.5
vesafb: probe of vesafb0 failed with error -6
ACPI: Processor [CPU1] (supports C1)
Real Time Clock Driver v1.12
Linux agpgart interface v0.100 (c) Dave Jones
agpgart: Detected AGP bridge 0
agpgart: Maximum main memory to use for agp memory: 941M
agpgart: AGP aperture is 64M @ 0xf800
serio: i8042 AUX port at 0x60,0x64 irq 12
serio: i8042 KBD port at 0x60,0x64 irq 1
Serial: 8250/16550 driver $Revision: 1.90 $ 8 ports, IRQ sharing enabled
ttyS0 at I/O 0x3f8 (irq = 4) is a 16550A
ttyS1 at I/O 0x2f8 (irq = 3) is a 16550A
RAMDISK driver initialized: 16 RAM disks of 16384K size 1024 blocksize
divert: not allocating divert_blk for non-ethernet device lo
Uniform Multi-Platform E-IDE driver Revision: 7.00alpha2
ide: Assuming 33MHz system bus speed for PIO modes; override with idebus=xx
VP_IDE: IDE controller at PCI slot :00:0f.1
ACPI: PCI interrupt :00:0f.1[A] -> GSI 11 (level, low) -> IRQ 11
VP_IDE: chipset revision 6
VP_IDE: not 100% native mode: will probe irqs later
VP_IDE: VIA vt8237 (rev 00) IDE UDMA133 controller on pci:00:0f.1
ide0: BM-DMA at 0xfc00-0xfc07, BIOS settings: hda:DMA, hdb:pio
ide1: BM-DMA at 0xfc08-0xfc0f, BIOS settings: hdc:DMA, hdd:pio
Probing IDE interface ide0...
hda: WDC WD1200JB-00FUA0, ATA DISK drive
Using cfq io scheduler
ide0 at 0x1f0-0x1f7,0x3f6 on irq 14
Probing IDE interface ide1...
hdc: SONY CD-RW CRX320E, ATAPI CD/DVD-ROM drive
ide1 at 0x170-0x177,0x376 on irq 15
Probing IDE interface ide2...
ide2: Wait for ready failed before probe !
Probing IDE interface ide3...
ide3: Wait for ready failed before probe !
Probing IDE interface ide4...
ide4: Wait for ready failed before probe !
Probing IDE interface ide5...
ide5: Wait for ready failed before probe !
hda: max request size: 1024KiB
hda: 234441648 sectors (120034 MB) w/8192KiB Cache, CHS=16383/255/63, UDMA(100)
hda: cache flushes supported
 hda: hda1 hda2 hda3 hda4 < hda5 hda6 hda7 >
hdc: ATAPI 52X DVD-ROM CD-R/RW drive, 2048kB Cache, UDMA(33)
Uniform CD-ROM driver Revision: 3.20
ide-floppy driver 0.99.newide
usbcore: registered new driver hiddev
usbcore: registered new driver usbhid
drivers/usb/input/hid-core.c: v2.0:USB HID core driver
mice: PS/2 mouse device common for all mice
input: AT Translated Set 2 keyboard on isa0060/serio0
input: ImPS/2 Generic Wheel Mouse on isa0060/serio1
md: md driver 0.90.0 MAX_MD_DEVS=256, MD_SB_DISKS=27
NET: Registered protocol family 2
IP: routing cache hash table of 2048 buckets, 64Kbytes
TCP: Hash tables configur

Re: [OpenAFS] 1.3.77 Brokenness

[Jack Neely]

Any in reply to my own post...

Has anyone tried opening up FC3/RHEL4 beta's natilus window on a
directory in AFS, such as a home directory?  Really, really, slow.  The
window has been scaning my home directory for 15 minutes now and is
still not complete.  Load has gone to 6.

Lots of small files and some larger ones...only about 145MB worth...

Jack Neely
-- 
Jack Neely <[EMAIL PROTECTED]>
Realm Linux Administration and Development
PAMS Computer Operations at NC State University
GPG Fingerprint: 1917 5AC1 E828 9337 7AA4  EA6B 213B 765F 3B6A 5B89
___
OpenAFS-info mailing list
OpenAFS-info@openafs.org
https://lists.openafs.org/mailman/listinfo/openafs-info

Re: [OpenAFS] 1.3.77 Brokenness

[Derrick J Brashear]

On Tue, 11 Jan 2005, Rainer Toebbicke wrote:

One more data point:
I saw this yesterday on a 2.4.21-20.EL kernel with OpenAFS 1.3.73 client code 
(against a volume on an OpenAFS 1.2.13 server):
1.3.73 is old and stuff has been fixed since then. Can you try with 
something newer?
___
OpenAFS-info mailing list
OpenAFS-info@openafs.org
https://lists.openafs.org/mailman/listinfo/openafs-info

Re: [OpenAFS] 1.3.77 Brokenness

[Jack Neely]

On Tue, Jan 11, 2005 at 01:05:16AM -0500, Matthew Miller wrote:
> On Mon, Jan 10, 2005 at 06:43:09PM -0500, Jack Neely wrote:
> > I keep RHEL/FC install trees in AFS which are served out via HTTP.  I
> > build the tree on a test machine and use rsync to move it out to AFS
> > land.  In this case my server is a RHEL3 linux box running OpenAFS
> > 1.2.11.  After moving the tree with rsync (a little over 2G worth) I
> > see that the cache gets out of sync on large files (20MB, 70MB, etc.)
> > Rsync complains it cannot delete its dot-files (the file it creates
> > before moving it to the real filename).  These . files show in an 'ls'
> > and when I try to rm them I get "No file or directory" yet they still
> > appear in AFS.  Many of my RPM pacakges are corrupt as well.
> 
> This is *exactly* the problem I've seen. It's reported at
> .
> 

Indeed.

This is going to be a showstopper for us.

Jack

> 
> 
> -- 
> Matthew Miller   [EMAIL PROTECTED]
> Boston University Linux  -->
> 

-- 
Jack Neely <[EMAIL PROTECTED]>
Realm Linux Administration and Development
PAMS Computer Operations at NC State University
GPG Fingerprint: 1917 5AC1 E828 9337 7AA4  EA6B 213B 765F 3B6A 5B89
___
OpenAFS-info mailing list
OpenAFS-info@openafs.org
https://lists.openafs.org/mailman/listinfo/openafs-info

Re: [OpenAFS] 1.3.77 Brokenness

[Rainer Toebbicke]

Matthew Miller wrote:
On Mon, Jan 10, 2005 at 06:43:09PM -0500, Jack Neely wrote:
I keep RHEL/FC install trees in AFS which are served out via HTTP.  I
build the tree on a test machine and use rsync to move it out to AFS
land.  In this case my server is a RHEL3 linux box running OpenAFS
1.2.11.  After moving the tree with rsync (a little over 2G worth) I
see that the cache gets out of sync on large files (20MB, 70MB, etc.)
Rsync complains it cannot delete its dot-files (the file it creates
before moving it to the real filename).  These . files show in an 'ls'
and when I try to rm them I get "No file or directory" yet they still
appear in AFS.  Many of my RPM pacakges are corrupt as well.

This is *exactly* the problem I've seen. It's reported at
.

One more data point:
I saw this yesterday on a 2.4.21-20.EL kernel with OpenAFS 1.3.73 client 
code (against a volume on an OpenAFS 1.2.13 server):

'make clean' refused to work on the OpenAFS code on the grounds that it 
couldn't remove the include/afs and include/rx directories because the 
weren't empty. Looking in those directories with "find" showed nothing, 
but after 'fs flush' 1-2 files appeared in there.

Right now I'm in .../doc/html/AdminReference and 'ls' says:
ls: auarf026.htm: No such file or directory
ls: auarf027.htm: No such file or directory
ls: auarf028.htm: No such file or directory
ls: auarf029.htm: No such file or directory

Let's see:
[root AdminReference]# fs flush .
[root AdminReference]# ls -l
total 278
lrwxr-xr-x1 rtb  root   56 Jan 10 10:36 auarf000.htm -> 
../../../../openafs/doc/html/AdminReference/auarf000.htm
lrwxr-xr-x1 rtb  root   56 Jan 10 10:36 auarf002.htm -> 
../../../../openafs/doc/html/AdminReference/auarf002.htm
...

and so on, all symbolic links (I use lndir to maintain the source tree 
for multiple architectures).

Looks like directory chunks do not [always] get updated in the 1.3.7x 
client cache.

--
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
Rainer Toebbicke
European Laboratory for Particle Physics(CERN) - Geneva, Switzerland
Phone: +41 22 767 8985   Fax: +41 22 767 7155
___
OpenAFS-info mailing list
OpenAFS-info@openafs.org
https://lists.openafs.org/mailman/listinfo/openafs-info

Re: [OpenAFS] 1.3.77 Brokenness

[Stefaan]

Same problem here (openafs 1.3.77 on 2.6, server and client on same
machine), also produced by rsync (I got a similar problem when
trying find | cpio -p)

I find it strange that I can not repair the filesystem with
"bos salvage".  And when I try deleting everything, except for e.g. 5
files who are "No such file or directory", bos salvage still happily
reports more then 200 inodes...  (even though I tried the -orphan switch)

Stefaan
___
OpenAFS-info mailing list
OpenAFS-info@openafs.org
https://lists.openafs.org/mailman/listinfo/openafs-info

Re: [OpenAFS] 1.3.77 Brokenness

[Matthew Miller]

On Mon, Jan 10, 2005 at 06:43:09PM -0500, Jack Neely wrote:
> I keep RHEL/FC install trees in AFS which are served out via HTTP.  I
> build the tree on a test machine and use rsync to move it out to AFS
> land.  In this case my server is a RHEL3 linux box running OpenAFS
> 1.2.11.  After moving the tree with rsync (a little over 2G worth) I
> see that the cache gets out of sync on large files (20MB, 70MB, etc.)
> Rsync complains it cannot delete its dot-files (the file it creates
> before moving it to the real filename).  These . files show in an 'ls'
> and when I try to rm them I get "No file or directory" yet they still
> appear in AFS.  Many of my RPM pacakges are corrupt as well.

This is *exactly* the problem I've seen. It's reported at
.



-- 
Matthew Miller   [EMAIL PROTECTED]
Boston University Linux  -->
___
OpenAFS-info mailing list
OpenAFS-info@openafs.org
https://lists.openafs.org/mailman/listinfo/openafs-info

[OpenAFS] 1.3.77 Brokenness

[Jack Neely]

Folks,

I've been doing some testing with OpenAFS 1.3.77 on the RHEL4 betas.
Just been testing the client against Transarc/OpenAFS 1.2.x servers.

Mostly, things work pretty well.  However, I'm having problems with
large files and have a pretty easy test case.

I keep RHEL/FC install trees in AFS which are served out via HTTP.  I
build the tree on a test machine and use rsync to move it out to AFS
land.  In this case my server is a RHEL3 linux box running OpenAFS
1.2.11.  After moving the tree with rsync (a little over 2G worth) I
see that the cache gets out of sync on large files (20MB, 70MB, etc.)
Rsync complains it cannot delete its dot-files (the file it creates
before moving it to the real filename).  These . files show in an 'ls'
and when I try to rm them I get "No file or directory" yet they still
appear in AFS.  Many of my RPM pacakges are corrupt as well.

When I do the same operation with the 2.4 kernel runnign OpenAFS 1.2.11
things work as expected.

That should be easy enough to reproduce...if you need any more
information just let me know.

Jack Neely
-- 
Jack Neely <[EMAIL PROTECTED]>
Realm Linux Administration and Development
PAMS Computer Operations at NC State University
GPG Fingerprint: 1917 5AC1 E828 9337 7AA4  EA6B 213B 765F 3B6A 5B89
___
OpenAFS-info mailing list
OpenAFS-info@openafs.org
https://lists.openafs.org/mailman/listinfo/openafs-info