[Linux-cachefs] "CacheFiles: Assertion failed" kernel crashes

[Vitaly]

Thank you both, more details below:


   - example of mount
   filer:/homes/vitaly on /homes/vitaly type nfs
   (rw,nosuid,nodev,soft,timeo=600,retrans=2,rsize=65536,wsize=65536,tcp,nosharecache,fsc=vitaly,sloppy)
   - ay 12 14:50:13 server kernel: [623309.219951]
   May 12 14:50:13 server kernel: [623309.219963] CacheFiles: Assertion
   failed
   May 12 14:50:13 server kernel: [623309.219972] 1 == 0 is false
   May 12 14:50:13 server kernel: [623309.220004] ------------[ cut here
   ]------------
   May 12 14:50:13 server kernel: [623309.220016] kernel BUG at
   /build/buildd/linux-3.13.0/fs/cachefiles/interface.c:333!
   May 12 14:50:13 server kernel: [623309.220032] invalid opcode: 0000 [#1]
   SMP
   May 12 14:50:13 server kernel: [623309.220047] Modules linked in:
   nvidia_uvm(POF) nfsv3 rpcsec_gss_krb5 nfsv4 autofs4 cac
   hefiles bnep rfcomm bluetooth nfsd auth_rpcgss nfs_acl nfs lockd sunrpc
   fscache gpio_ich snd_hda_codec_hdmi joydev wmi x86
   _pkg_temp_thermal intel_powerclamp mac_hid coretemp kvm_intel kvm
   crct10dif_pclmul crc32_pclmul ghash_clmulni_intel aesni_
   intel aes_x86_64 lrw gf128mul glue_helper ablk_helper cryptd sb_edac
   snd_hda_intel snd_seq_midi edac_core snd_hda_codec sn
   d_seq_midi_event snd_rawmidi snd_hwdep ioatdma snd_pcm snd_seq
   nvidia(POF) mei_me mei snd_page_alloc snd_seq_device snd_ti
   mer snd drm lpc_ich soundcore ipmi_si parport_pc ppdev lp parport
   hid_generic isci igb ixgbe i2c_algo_bit usbhid libsas md
   io dca ahci ptp hid libahci pps_core scsi_transport_sas
   May 12 14:50:13 server kernel: [623309.220356] CPU: 6 PID: 9162 Comm:
   kworker/u50:1 Tainted: PF          O 3.13.0-24-gene
   ric #47-Ubuntu
   May 12 14:50:13 server kernel: [623309.220375] Hardware name: Supermicro
   X9DRG-O(T)F/X9DRG-O(T)F, BIOS 3.0  12/27/2013
   May 12 14:50:13 server kernel: [623309.220397] Workqueue: fscache_object
   fscache_object_work_func [fscache]
   May 12 14:50:13 server kernel: [623309.220428] task: ffff8839df36c7d0
   ti: ffff8820a3772000 task.ti: ffff8820a3772000
   May 12 14:50:13 server kernel: [623309.220472] RIP:
   0010:[<ffffffffa023a05b>]  [<ffffffffa023a05b>] cachefiles_put_object
   +0x22b/0x2a0 [cachefiles]
   - May 12 14:50:13 server kernel: [623309.220525] RSP:
   0018:ffff8820a3773dc8  EFLAGS: 00010282
   May 12 14:50:13 server kernel: [623309.220551] RAX: 000000000000000f
   RBX: ffff883c318c4f00 RCX: 0000000000000006
   May 12 14:50:13 server kernel: [623309.220595] RDX: 0000000000000007
   RSI: 0000000079887986 RDI: 0000000000000246
   May 12 14:50:13 server kernel: [623309.220638] RBP: ffff8820a3773dd8
   R08: 0000000000000092 R09: 0000000000000849
   May 12 14:50:13 server kernel: [623309.220681] R10: ffffffff8185a960
   R11: ffff8820a3773af6 R12: ffff883c318c4f00
   May 12 14:50:13 server kernel: [623309.220725] R13: ffff883c318c4f00
   R14: 0000000000000061 R15: ffff883c318c4f90
   May 12 14:50:13 server kernel: [623309.220768] FS:
   0000000000000000(0000) GS:ffff88409fc00000(0000) knlGS:00000000000000
   00
   May 12 14:50:13 server kernel: [623309.220813] CS:  0010 DS: 0000 ES:
   0000 CR0: 0000000080050033




On Tue, May 12, 2015 at 5:33 PM, Milosz Tanski <milosz at adfin.com> wrote:

> On Tue, May 12, 2015 at 10:08 AM, Vitaly <vitaly at karasik.org> wrote:
> > My server  crashes about once per week because cachefiles with
> "CacheFiles:
> > Assertion failed" .
> > It's Ubuntu 14.04 with "CacheFiles: Assertion failed"  kernel.
> >
> > Is there any fix or  workaround  for avoid these crashes?
> >
> > TIA,
> > Vitaly
>
> Vitaly,
>
> I believe there's been a lot of fixes for fscache since the kernel in
> 14.04 made by David and to some extent my self.
>
> http://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/log/?qt=grep&q=FS-Cache
> I believe David also has additional fixes that are queued up for NFS
> with fscache. If it's possible try using the vivid kernels on your
> machine.
>
> Can you please provide more details about the assertion? Like the log
> lines before it / any kernel oops. What filesystem you're backing with
> fscache. Without this information it's hard for us to help you out.
>
> Best,
> - Milosz
>
> --
> Milosz Tanski
> CTO
> 16 East 34th Street, 15th floor
> New York, NY 10016
>
> p: 646-253-9055
> e: milosz at adfin.com
>