Re: Hard lockup in ext4_finish_bio

On 10/08/2015 09:56 PM, John Stoffel wrote: >> "Nikolay" == Nikolay Borisov writes: > > Nikolay> On 10/08/2015 05:34 PM, John Stoffel wrote: >>> Great bug report, but you're missing the info on which kernel >>> you're > > Nikolay> This is on 3.12.47 (self compiled). It was evident on my >

Re: Hard lockup in ext4_finish_bio

On 10/08/2015 09:56 PM, John Stoffel wrote: >> "Nikolay" == Nikolay Borisov writes: > > Nikolay> On 10/08/2015 05:34 PM, John Stoffel wrote: >>> Great bug report, but you're missing the info on which kernel >>> you're > > Nikolay> This is on 3.12.47 (self compiled). It

Re: Hard lockup in ext4_finish_bio

> "Nikolay" == Nikolay Borisov writes: Nikolay> On 10/08/2015 05:34 PM, John Stoffel wrote: >> Great bug report, but you're missing the info on which kernel >> you're Nikolay> This is on 3.12.47 (self compiled). It was evident on my Nikolay> initial post, but I did forget to mention that in

Re: Hard lockup in ext4_finish_bio

Great bug report, but you're missing the info on which kernel you're running here... is this a vendor kernel or self-compiled? Nikolay> I've hit a rather strange hard lock up on one of my servers Nikolay> from the page writeback path, the actual backtrace is: Nikolay> [427149.717151]

Re: Hard lockup in ext4_finish_bio

On 10/08/2015 05:34 PM, John Stoffel wrote: > Great bug report, but you're missing the info on which kernel you're This is on 3.12.47 (self compiled). It was evident on my initial post, but I did forget to mention that in the reply. Also, I suspect even current kernel are susceptible to this

Re: Hard lockup in ext4_finish_bio

[Adding the dm people since I believe this might be in their field as well.] After further investigation I discovered that there was indeed a task, executing block_write_full_page_endio (this function is supposed to eventually call end_buffer_async_write which would then unblock the writeback

Hard lockup in ext4_finish_bio

Hello, I've hit a rather strange hard lock up on one of my servers from the page writeback path, the actual backtrace is: [427149.717151] [ cut here ] [427149.717553] WARNING: CPU: 23 PID: 4611 at kernel/watchdog.c:245 watchdog_overflow_callback+0x98/0xc0()

Hard lockup in ext4_finish_bio

Hello, I've hit a rather strange hard lock up on one of my servers from the page writeback path, the actual backtrace is: [427149.717151] [ cut here ] [427149.717553] WARNING: CPU: 23 PID: 4611 at kernel/watchdog.c:245 watchdog_overflow_callback+0x98/0xc0()

Re: Hard lockup in ext4_finish_bio

[Adding the dm people since I believe this might be in their field as well.] After further investigation I discovered that there was indeed a task, executing block_write_full_page_endio (this function is supposed to eventually call end_buffer_async_write which would then unblock the writeback

Re: Hard lockup in ext4_finish_bio

On 10/08/2015 05:34 PM, John Stoffel wrote: > Great bug report, but you're missing the info on which kernel you're This is on 3.12.47 (self compiled). It was evident on my initial post, but I did forget to mention that in the reply. Also, I suspect even current kernel are susceptible to this

Re: Hard lockup in ext4_finish_bio

Great bug report, but you're missing the info on which kernel you're running here... is this a vendor kernel or self-compiled? Nikolay> I've hit a rather strange hard lock up on one of my servers Nikolay> from the page writeback path, the actual backtrace is: Nikolay> [427149.717151]

Re: Hard lockup in ext4_finish_bio

> "Nikolay" == Nikolay Borisov writes: Nikolay> On 10/08/2015 05:34 PM, John Stoffel wrote: >> Great bug report, but you're missing the info on which kernel >> you're Nikolay> This is on 3.12.47 (self compiled). It was evident on my Nikolay> initial post, but I did forget