Can you throttle the bandwidth to a tmpfs based ramdisk

bandwidthcgroupsiolimitramdisk

I created a ramdisk using tmpfs as follows:

sudo mount -t tmpfs -o size=1024m tmpfs /myramdisk

I was looking into cgroup blkio subsystem bandwidth limits "read_bps_device" and "write_bps_device" but they seem to require the major and minor number of your block device that you want to throttle access to.

Is there any way to achieve a similar bandwidth limit to the in-memory tmpfs ramdisk?

UPDATE: I found a roundabout hacky way of accomplishing this. I first nfs export the ramdisk mount. Then nfs mount it over loopback on the same machine. I then use linux network traffic shaping to set bandwidth limits. This seems to provide what I need. I'm in the process of performing some measurements to see the penalty this hideous layering entails. I'll update info on that in a few days.

Best Answer

stat or /proc/[PID]/mountinfo should still tell you what the device numbers are:

[root@XXXlin01 block]# stat --printf="%d" /tmp/mnt; echo
24
[root@XXXlin01 block]# stat --printf="%d" /tmp/mnt2; echo
25
[root@XXXlin01 block]# grep "/tmp/mnt" /proc/22195/mountinfo
40 20 0:24 / /tmp/mnt rw,relatime - tmpfs none rw
41 20 0:25 / /tmp/mnt2 rw,relatime - tmpfs none rw

Both outputs show 0:25 and 0:24 as the device numbers in question. The "0" device number is used for "unnamed" mounts (such as tmpfs, sysfs, nfs, procfs, etc). For example, here is an NFS mount:

[root@XXXlin01 block]# grep "/LinuxHome" /proc/22195/mountinfo
39 36 0:23 / /home/jad87 rw,relatime - nfs duhsnas-xxx:/ifs/Application\040File\040Shares/DUHS/LinuxHomeDir_fs/jad87 rw,vers=3,rsize=4096,wsize=4096,namlen=255,soft,proto=tcp,timeo=14,retrans=2,sec=sys,mountaddr=10.1x.92.106,mountvers=3,mountport=300,mountproto=udp,local_lock=none,addr=10.1x.92.106
42 36 0:26 / /home/tsa20 rw,relatime - nfs 10.1x.92.117:/ifs/Application\040File\040Shares/DUHS/LinuxHomeDir_fs/tsa20 rw,vers=3,rsize=4096,wsize=4096,namlen=255,soft,proto=tcp,timeo=14,retrans=2,sec=sys,mountaddr=10.1x.92.117,mountvers=3,mountport=300,mountproto=udp,local_lock=none,addr=10.1x.92.117

The NFS mounts above are 0:23 and 0:26 respectively.

Related Solutions

CentOS – How to Limit the Bandwidth Used by a Process

You can use iptables to mark a packet (--pid-owner ...), then use tc to shape the traffic. Also "--sid-owner" can be used to include threads and children of that process.

http://www.frozentux.net/iptables-tutorial/iptables-tutorial.html#OWNERMATCH

Match --pid-owner
Kernel 2.3, 2.4, 2.5 and 2.6
Example iptables -A OUTPUT -m owner --pid-owner 78
Explanation This match is used to match packets based on the Process ID (PID) that was responsible for them. This match is a bit harder to use, but one example would be only to allow PID 94 to send packets from the HTTP port (if the HTTP process is not threaded, of course). Alternatively we could write a small script that grabs the PID from a ps output for a specific daemon and then adds a rule for it. For an example, you could have a rule as shown in the Pid-owner.txt example

Can the “du” program be made less aggressive

Take a look at ionice. From man ionice:

This program sets or gets the io scheduling class and priority for a program. If no arguments or just -p is given, ionice will query the current io scheduling class and priority for that process.

To run du with the "idle" I/O class, which is the lowest priority available, you can do something like this:

ionice -c 3 du -s

This should stop du from interfering with other process' I/O. You might also want to consider renicing the program to lower its CPU priority, like so:

renice -n 19 "$duPid"

You can also do both at initialisation time:

nice -n 19 ionice -c 3 du

Best Answer

Related Solutions

CentOS – How to Limit the Bandwidth Used by a Process

Can the “du” program be made less aggressive

Related Question