Hot answers tagged

3

You could use pv: </mnt/nfs/image.img pv -L 5m >/dev/sda The -L flag limits the throughput to 5 megabytes per second. pv also writes to the stdout so you have to redirect to the target with >.


2

If you look at the kernel's inode source code, you can see that the ihash_entries is set at the kernel level only. There is no user or process level considerations at all. Adding those could drastically decrease performance which would be counter productive. It would also imply keeping track of all processes that used the cached entries, therefore ...


2

Spicificly for setrlimit Here are some of the more usefull command options that you may wish to look into; pulled'em from the man pages. RLIMIT_NOFILE Specifies a value one greater than the maximum file descriptor number that can be opened by this process. RLIMIT_NPROC The maximum number of processes (or, more precisely on Linux, threads) that can ...


2

Open files limit is per process. An user can have multiple processes that total FD count can be greater than open files limit. From setrlimit(3) man page: RLIMIT_NOFILE Specifies a value one greater than the maximum file descriptor number that can be opened by this process. Attempts (open(2), pipe(2), dup(2), etc.) to exceed this limit yield ...


2

While investigating another issue, I may have found something relevant. It wasn't possible to switch to another tty (Ctrl + Alt + F2): A start job is running for Login Service... Turns out this may be another systemd issue, which has its own limits. The following config file was created, which apparently fixed the issue: # mkdir /etc/systemd/logind.conf....


2

ulimit is a sh family (so bash, ksh etc) builtin. For csh family the command is limit. (zsh is complicated and allows both.) Normal users can not raise their hard limits. Only root can do that. On a typical Linux machine this is done via pam_limits (e.g. in /etc/security/limits.conf and files in /etc/security/limits.d). These settings will take effect ...


2

nproc was the problem: [root@localhost ~]# ps -eLf | grep pascal | wc -l 4068 [root@localhost ~]# cat /etc/security/limits.d/20-nproc.conf # Default limit for number of user's processes to prevent # accidental fork bombs. # See rhbz #432903 for reasoning. * soft nproc 4096 root soft nproc unlimited [root@localhost ~]# man ...


1

It turned out to be a small config mistake by myself. After 8 hours of brain-pain and fulle rebuilding the config file the fault turned out to be in the TransferLimit line where also the group was specified. (which is not needed because we use class definitian). TransferRate APPE,RETR,STOR,STOU 1000 group speedlimit should have been: TransferRate APPE,...


1

Since more than a decade, 32 bit Linux applications are able to access files larger than 2 TiB (2^31) thanks to the implementation of large file support. The current OS limitation is 8 EiB (2^63) which shouldn't hit the common of us before a while... You would need a file system that makes no lower limit on file size too.


1

File size is limited by filesystem type not by OS. Typically, OS supports several filesystems, so there is no such thing like "OS file size limit". There are limits for well-known filesystems: FAT32 - 4Gib NTFS - 16Eib ext2/3 - 16Gib - 2Tib (depends from block size) ext4 - 16Gib - 16Tib XFS - 9Eib ZFS - 16Eib


1

This kind of non-responsiveness, although the CPU is limited is often caused by swapping (i.e. your process pushes other tasks out to disk and getting them back in is going to take a lot of time). The best way to limit your memory usage is normally from within the program. If that is not possible and memory is consumed slowly (because it is not released) it ...


1

Resource limits are generally controlled through ulimit (user-based) or sysctl (system-based). For example, the kernel.shmmax parameter, set by sysctl defines the maximum size in bytes of a single shared memory segment that a Linux process can allocate in its virtual address space. ulimit is used to set the limits of normal user processes. These values ...


1

From man bashbuiltins: ulimit [-HSTabcdefilmnpqrstuvx [limit]] Provides control over the resources available to the shell and to processes started by it, on systems that allow such control. Use ulimit -a to show current limits.


1

The groups defined with the @group syntax in the limits.conf file can match groups defined in any group database back-end, i.e. files (/etc/group), nis, ldap, and whatever else nsswitch.conf might support. Assigning a group to an ldap user entry is not done by locating his/her entry somewhere in the hierarchy (like under ou=student in your question) but by ...


1

We went with cgroups in the end, since there really doesn't seem to be any other approach that would accomplish this. Cgroups allow CPU utilization limiting through the kernel scheduler, using cpu.cfs_period_us and cpu.cfs_quota_us. This avoids the explicit specification of CPU cores.



Only top voted, non community-wiki answers of a minimum length are eligible