| Bug #7902 | Same Error as Bug # 5656 but different Problem ! | ||
|---|---|---|---|
| Submitted: | 14 Jan 2005 10:39 | Modified: | 14 Feb 2005 17:39 |
| Reporter: | Anand K | Email Updates: | |
| Status: | No Feedback | Impact on me: | |
| Category: | MySQL Server: ISAM storage engine | Severity: | S3 (Non-critical) |
| Version: | 4.1.8-standard | OS: | Linux (RHEL ES rel 3.0 ;2.4.21-20.ELsmp) |
| Assigned to: | CPU Architecture: | Any | |
[14 Jan 2005 10:42]
Anand K
Show Varibales Command
Attachment: show_variables.txt (text/plain), 21.63 KiB.
[14 Jan 2005 14:02]
Martin Friebe
Hi Anand, available memory is not all. some op systems have individual limits about the code/data/stack size a process (or process group) can have. which you are likely to hit. Unfurtunatley I dont know any good documentation for any op system on this issue. Neither am I a linux expert. However the linuxthread library (under freebsd) allocates stack size to each thread. that might be similiar under RHEL. You might have it a limit configured in you Kernel. (the freebsd option is MAXSSIZ and with it set to 1GB I can have 800 connections) TO the msysql team: Even so thsi is not a problem of the amysql software, it would be nice to have a resource limitations and configurations section in the operating system pages, of the documentation.
[14 Jan 2005 17:39]
Hartmut Holzgraefe
The formula given in #5656 did not take the thread stacks into account, these take 256K each with static binaries released by MySQL that is linked againsta patched glibc and 2MB for binaries using regular glibc versions. 1000 connections * 2MB -> 2GB alone which might already cause trouble but it might also be that number of processes or threads is limited or that you are just running out of file descriptors (although that should give a different error message afair)
[14 Feb 2005 22:54]
Bugs System
No feedback was provided for this bug for over a month, so it is being suspended automatically. If you are able to provide the information that was originally requested, please do so and change the status of the bug back to "Open".
Description: [root@newdb root]# mysql ERROR 1135 (00000): Can't create a new thread (errno 11); if you are not out of available memory, you can consult the manual for a possible OS-dependent bug Free Memory at this time: total used free shared buffers cached Mem: 3599032 554856 3044176 0 39404 213856 -/+ buffers/cache: 301596 3297436 Swap: 4194224 0 4194224 load average: 0.34, 0.46, 0.37 Variable_name: max_connections Value: 1200 Variable_name: key_buffer_size Value: 8388600 Variable_name: sort_buffer_size Value: 1048568 Variable_name: read_rnd_buffer_size Value: 262144 Variable_name: read_buffer_size Value: 1044480 The problem occurs serveral times in a day .. as soon as the number of connections gets close to 1000. It will knock off a couple of connections and then everything will be alright. The calculation formula in Bug# 5656 does not work for me. I have also checked the file handlers and all that: [root@newdb root]# cat /proc/sys/fs/file-nr 1222 1044 366996 [root@newdb root]# How to repeat: On my box only .. the moment the number of connections gets close to 1000