github-starred/ollama
2
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-05-07 08:30:05 -05:00
Code Issues 14.5k Packages Projects Releases 120 Wiki Activity

[GH-ISSUE #4342] echo 0 > /proc/sys/kernel/hung_task_timeout_secs, 然后显卡hung住了，最后服务器只能重启了 #28464

New Issue
Closed
opened 2026-04-22 06:39:24 -05:00 by GiteaMirror · 2 comments
GiteaMirror commented 2026-04-22 06:39:24 -05:00
Owner
Copy Link

Originally created by @YangFW on GitHub (May 11, 2024).
Original GitHub issue: https://github.com/ollama/ollama/issues/4342

What is the issue?

May 11 10:54:31 localhost kernel: [] ? system_call_fastpath+0x25/0x2a
May 11 10:54:31 localhost kernel: [] ? SyS_ioctl+0x81/0xa0
May 11 10:54:31 localhost kernel: [] ? SYSC_newstat+0x2e/0x60
May 11 10:54:31 localhost kernel: [] ? do_vfs_ioctl+0x3a8/0x5c0
May 11 10:54:31 localhost kernel: [] ? nvidia_unlocked_ioctl+0x20/0x30 [nvidia]
May 11 10:54:31 localhost kernel: [] ? nvidia_ioctl.isra.22+0x728/0x910 [nvidia]
May 11 10:54:31 localhost kernel: [] ? rm_ioctl+0x58/0xb0 [nvidia]
May 11 10:54:31 localhost kernel: [] ? _nv000731rm+0x16b/0xeb0 [nvidia]
May 11 10:54:31 localhost kernel: [] ? _nv000577rm+0x5d/0x70 [nvidia]
May 11 10:54:31 localhost kernel: [] ? _nv045215rm+0x5c/0x90 [nvidia]
May 11 10:54:31 localhost kernel: [] ? _nv045214rm+0x16f/0x320 [nvidia]
May 11 10:54:31 localhost kernel: [] ? _nv047020rm+0x175/0x3a0 [nvidia]
May 11 10:54:31 localhost kernel: [] ? _nv047087rm+0x54/0xd0 [nvidia]
May 11 10:54:31 localhost kernel: [] ? _nv045252rm+0x26c/0x2f0 [nvidia]
May 11 10:54:31 localhost kernel: [] _nv043989rm+0x10/0x40 [nvidia]
May 11 10:54:31 localhost kernel: [] os_acquire_rwlock_write+0x42/0x50 [nvidia]
May 11 10:54:31 localhost kernel: [] down_write+0x2d/0x41
May 11 10:54:31 localhost kernel: [] call_rwsem_down_write_failed+0x17/0x30
May 11 10:54:31 localhost kernel: [] ? __d_instantiate+0x2d/0xf0
May 11 10:54:31 localhost kernel: [] rwsem_down_write_failed+0x215/0x3c0
May 11 10:54:31 localhost kernel: [] schedule+0x29/0x70
May 11 10:54:31 localhost kernel: Call Trace:
May 11 10:54:31 localhost kernel: ollama_llama_se D ffff9ac9571205e0 0 1236 1183 0x00000080
May 11 10:54:31 localhost kernel: "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
May 11 10:54:31 localhost kernel: INFO: task ollama_llama_se:1236 blocked for more than 120 seconds.
May 11 10:52:39 localhost kernel: TCP: request_sock_TCP: Possible SYN flooding on port 42852. Sending cookies. Check SNMP counters.

服务128G，80线程，2*4090，nvidia driver：550，cuda版本：12.4
ollama 开启 codegemma，使用正常，但是过了一天就挂了，用不了了
日志倒叙如上，然后ollama服务、python服务全假死，kill不掉
/etc/sysctl.conf文件内容已经改成了下面的：
fs.inotify.max_user_watches=524288
vm.overcommit_memory = 1
net.core.somaxconn = 1024
vm.dirty_background_ratio=5
vm.dirty_ratio=10
net.ipv4.ip_forward=1
最后只能重启服务器解决这样的问题，请问有什么好的办法吗，是ollama的问题吗，服务器上该配的已经配好了

OS

Linux

GPU

Nvidia

CPU

No response

Ollama version

0.1.33

Originally created by @YangFW on GitHub (May 11, 2024). Original GitHub issue: https://github.com/ollama/ollama/issues/4342 ### What is the issue? May 11 10:54:31 localhost kernel: [<ffffffff85fc539a>] ? system_call_fastpath+0x25/0x2a May 11 10:54:31 localhost kernel: [<ffffffff85a71c21>] ? SyS_ioctl+0x81/0xa0 May 11 10:54:31 localhost kernel: [<ffffffff85a61d0e>] ? SYSC_newstat+0x2e/0x60 May 11 10:54:31 localhost kernel: [<ffffffff85a71988>] ? do_vfs_ioctl+0x3a8/0x5c0 May 11 10:54:31 localhost kernel: [<ffffffffc074c130>] ? nvidia_unlocked_ioctl+0x20/0x30 [nvidia] May 11 10:54:31 localhost kernel: [<ffffffffc074bf28>] ? nvidia_ioctl.isra.22+0x728/0x910 [nvidia] May 11 10:54:31 localhost kernel: [<ffffffffc1237e38>] ? rm_ioctl+0x58/0xb0 [nvidia] May 11 10:54:31 localhost kernel: [<ffffffffc1230abb>] ? _nv000731rm+0x16b/0xeb0 [nvidia] May 11 10:54:31 localhost kernel: [<ffffffffc07f8f9d>] ? _nv000577rm+0x5d/0x70 [nvidia] May 11 10:54:31 localhost kernel: [<ffffffffc07eb1bc>] ? _nv045215rm+0x5c/0x90 [nvidia] May 11 10:54:31 localhost kernel: [<ffffffffc07ead7f>] ? _nv045214rm+0x16f/0x320 [nvidia] May 11 10:54:31 localhost kernel: [<ffffffffc1062745>] ? _nv047020rm+0x175/0x3a0 [nvidia] May 11 10:54:31 localhost kernel: [<ffffffffc07e9614>] ? _nv047087rm+0x54/0xd0 [nvidia] May 11 10:54:31 localhost kernel: [<ffffffffc0801f2c>] ? _nv045252rm+0x26c/0x2f0 [nvidia] May 11 10:54:31 localhost kernel: [<ffffffffc10610c0>] _nv043989rm+0x10/0x40 [nvidia] May 11 10:54:31 localhost kernel: [<ffffffffc0758132>] os_acquire_rwlock_write+0x42/0x50 [nvidia] May 11 10:54:31 localhost kernel: [<ffffffff85fb716d>] down_write+0x2d/0x41 May 11 10:54:31 localhost kernel: [<ffffffff85bae557>] call_rwsem_down_write_failed+0x17/0x30 May 11 10:54:31 localhost kernel: [<ffffffff85a7519d>] ? __d_instantiate+0x2d/0xf0 May 11 10:54:31 localhost kernel: [<ffffffff85fb9455>] rwsem_down_write_failed+0x215/0x3c0 May 11 10:54:31 localhost kernel: [<ffffffff85fb7ca9>] schedule+0x29/0x70 May 11 10:54:31 localhost kernel: Call Trace: May 11 10:54:31 localhost kernel: ollama_llama_se D ffff9ac9571205e0 0 1236 1183 0x00000080 May 11 10:54:31 localhost kernel: "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message. May 11 10:54:31 localhost kernel: INFO: task ollama_llama_se:1236 blocked for more than 120 seconds. May 11 10:52:39 localhost kernel: TCP: request_sock_TCP: Possible SYN flooding on port 42852. Sending cookies. Check SNMP counters. 服务128G，80线程，2*4090，nvidia driver：550，cuda版本：12.4 ollama 开启 codegemma，使用正常，但是过了一天就挂了，用不了了 日志倒叙如上，然后ollama服务、python服务全假死，kill不掉 /etc/sysctl.conf文件内容已经改成了下面的： fs.inotify.max_user_watches=524288 vm.overcommit_memory = 1 net.core.somaxconn = 1024 vm.dirty_background_ratio=5 vm.dirty_ratio=10 net.ipv4.ip_forward=1 最后只能重启服务器解决这样的问题，请问有什么好的办法吗，是ollama的问题吗，服务器上该配的已经配好了 ### OS Linux ### GPU Nvidia ### CPU _No response_ ### Ollama version 0.1.33
GiteaMirror added the bug label 2026-04-22 06:39:24 -05:00
GiteaMirror commented 2026-04-22 06:39:26 -05:00
Author
Owner
Copy Link

@mchiang0610 commented on GitHub (May 11, 2024):

对不起，可以试一下 ollama 最新的 0.1.36 吗? 这个问题应该已经解决了

还有遇到这个问题的话，可以帮我们重开这个吗? 谢谢

<!-- gh-comment-id:2105623337 --> @mchiang0610 commented on GitHub (May 11, 2024): 对不起，可以试一下 ollama 最新的 0.1.36 吗? 这个问题应该已经解决了 还有遇到这个问题的话，可以帮我们重开这个吗? 谢谢
GiteaMirror commented 2026-04-22 06:39:27 -05:00
Author
Owner
Copy Link

@woxiangbo commented on GitHub (Aug 16, 2024):

@mchiang0610 竟然碰到一个懂中文的大佬，麻烦帮忙看下这个问题呢，搞不定了 https://github.com/ollama/ollama/issues/4131
感谢！

<!-- gh-comment-id:2293151548 --> @woxiangbo commented on GitHub (Aug 16, 2024): @mchiang0610 竟然碰到一个懂中文的大佬，麻烦帮忙看下这个问题呢，搞不定了 https://github.com/ollama/ollama/issues/4131 感谢！
Sign in to join this conversation.
No Branch/Tag Specified
Branches Tags
Labels
Clear labels
amd
api
app
bug
build
cli
cloud
compatibility
context-length
create
docker
documentation
embeddings
feature request
feedback wanted
good first issue
gpt-oss
gpu
harmony
help wanted
image
install
intel
js
launch
linux
macos
memory
mlx
model
needs more info
networking
nvidia
ollama.com
performance
pull-request
Mirrored from GitHub Pull Request
 python
question
registry
rendering
thinking
tools
top
vulkan
windows
wsl
No Label bug
Milestone
No items
No Milestone
Projects
Clear projects
No project
Assignees
Clear assignees
GiteaMirror ninjasurge
No Assignees
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: github-starred/ollama#28464
Delete Branch "%!s()"

Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?