Hi, I didnt realize development was still going on. Any reason why the port in FreeBSD has not been updated ? In terms of what has changed since 5.5, is there a list somewhere ? Having a quick look at the cvs web repo
I see libpdel has been removed as a dependency
some improvements to the CLI and web interface
netflow v9 support
new 'set iface group command' ? What does that do ?
new set iface name and descr
Anything else ?
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
You always can see, whats memory allocated for netgraph by use "vmstat -z" command
In it's output, search thees lines
NetGraph items:
NetGraph data items:
First - net.graph.maxalloc
Second - net.graph.maxdata
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
so you think that i shouldn't improve anything else, sysctl, tweak something in the source and so on?
what is the best way to get source from cvs and install it.
Thanks again,
Merry Christmass
Sami
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Mpd-5.6 was released in ports few days ago. So at this moment there is no reason to take CVS sources - they are the same now.
What's about tuning, there is no much recipes of "one size fits all" type now. Most of previous were integrated to work out of the box now. Fine tuning for specific workload and hardware still could be done, but requires in-place experimenting. The only thing I can recommend now in general is to disable devd or make it to not react on interface events.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
devd is not bad, but it runs number of shell scripts on interface events. That creates additional load on system, that may be significant in time of peak loads.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Hi,
I want to use MPD5.5 with freebsd to get l2tp/pptp & pppoe sessions of users.
I read somwhere its possible to run 6000+ sessions.
I would like your help to know what FBSD version and what ports, what sysctls should be set and so on to get best performance.
another question, i see on sourceforge patches fixed,are they applied on the FBSD port? or i should apply them manually?
Thanks in advance,
Sami
1. Increase net.graph.maxdata and net.graph.recvspace sysctl values
2. Get latest source from CVS manually
3. That's all :)
Hi, I didnt realize development was still going on. Any reason why the port in FreeBSD has not been updated ? In terms of what has changed since 5.5, is there a list somewhere ? Having a quick look at the cvs web repo
I see libpdel has been removed as a dependency
some improvements to the CLI and web interface
netflow v9 support
new 'set iface group command' ? What does that do ?
new set iface name and descr
Anything else ?
set iface froup add ability to group some interfeces as one group
This useful, if you use PF
Hi,
1. Increase net.graph.maxdata and net.graph.recvspace sysctl values
mpd2# sysctl net.graph | more
net.graph.msg_version: 8
net.graph.abi_version: 12
net.graph.maxdata: 128000
net.graph.maxalloc: 128000
net.graph.threads: 4
net.graph.control.proto: 2
net.graph.data.proto: 1
net.graph.family: 32
net.graph.recvspace: 40960
net.graph.maxdgram: 40960
do i need to increase?
2. Get latest source from CVS manually
how do I do that?
3. That's all :)
You always can see, whats memory allocated for netgraph by use "vmstat -z" command
In it's output, search thees lines
NetGraph items:
NetGraph data items:
First - net.graph.maxalloc
Second - net.graph.maxdata
Hi,
here is the output:
# vmstat -z
ITEM SIZE LIMIT USED FREE REQUESTS FAILURES
UMA Kegs: 208, 0, 95, 7, 95, 0
UMA Zones: 256, 0, 95, 10, 95, 0
UMA Slabs: 568, 0, 1826, 1, 6799, 0
UMA RCntSlabs: 568, 0, 771, 6, 771, 0
UMA Hash: 256, 0, 2, 13, 4, 0
16 Bucket: 152, 0, 75, 0, 75, 0
32 Bucket: 280, 0, 119, 7, 119, 0
64 Bucket: 536, 0, 80, 4, 80, 31
128 Bucket: 1048, 0, 184, 2, 184, 5
VM OBJECT: 216, 0, 43094, 322, 1800371, 0
MAP: 232, 0, 7, 25, 7, 0
KMAP ENTRY: 120, 55955, 52, 165, 114785, 0
MAP ENTRY: 120, 0, 922, 876, 3491969, 0
DP fakepg: 120, 0, 0, 0, 0, 0
SG fakepg: 120, 0, 0, 0, 0, 0
mt_zone: 2056, 0, 289, 33, 289, 0
16: 16, 0, 2149, 371, 9081505, 0
32: 32, 0, 2506, 423, 374615100, 0
64: 64, 0, 3923, 1173, 2216975, 0
128: 128, 0, 10519, 675, 1796549, 0
256: 256, 0, 3220, 2225, 812541, 0
512: 512, 0, 1514, 166, 216750, 0
1024: 1024, 0, 196, 56, 305451, 0
2048: 2048, 0, 525, 159, 5270465, 0
4096: 4096, 0, 675, 184, 185776, 0
Files: 80, 0, 330, 210, 826877, 0
TURNSTILE: 136, 0, 229, 31, 229, 0
umtx pi: 96, 0, 0, 0, 0, 0
MAC labels: 40, 0, 0, 0, 0, 0
PROC: 1120, 0, 50, 49, 124442, 0
THREAD: 1112, 0, 160, 68, 53449, 0
SLEEPQUEUE: 80, 0, 229, 32, 229, 0
VMSPACE: 392, 0, 25, 65, 124403, 0
cpuset: 72, 0, 2, 98, 2, 0
audit_record: 952, 0, 0, 0, 0, 0
mbuf_packet: 256, 0, 1024, 384, 712380446, 0
mbuf: 256, 0, 3, 389, 1429580636, 0
mbuf_cluster: 2048, 409600, 1408, 80, 1411, 0
mbuf_jumbo_page: 4096, 12800, 0, 27, 102, 0
mbuf_jumbo_9k: 9216, 6400, 0, 0, 0, 0
mbuf_jumbo_16k: 16384, 3200, 0, 0, 0, 0
mbuf_ext_refcnt: 4, 0, 0, 0, 0, 0
g_bio: 232, 0, 0, 3632, 1913357, 0
ttyinq: 160, 0, 135, 105, 465, 0
ttyoutq: 256, 0, 72, 63, 248, 0
ata_request: 320, 0, 0, 24, 18, 0
ata_composite: 336, 0, 0, 0, 0, 0
cryptop: 88, 0, 0, 0, 0, 0
cryptodesc: 72, 0, 0, 0, 0, 0
VNODE: 472, 0, 59526, 74, 264616, 0
VNODEPOLL: 112, 0, 0, 0, 0, 0
NAMEI: 1024, 0, 0, 36, 1643569, 0
S VFS Cache: 108, 0, 59750, 1498, 278933, 0
L VFS Cache: 328, 0, 125, 403, 2709, 0
DIRHASH: 1024, 0, 1031, 53, 1037, 0
NFSMOUNT: 624, 0, 0, 0, 0, 0
NFSNODE: 688, 0, 0, 0, 0, 0
pipe: 728, 0, 6, 49, 108846, 0
ksiginfo: 112, 0, 85, 971, 4287, 0
itimer: 344, 0, 0, 0, 0, 0
bridge_rtnode: 64, 0, 0, 0, 0, 0
KNOTE: 128, 0, 0, 116, 1408, 0
socket: 680, 204804, 363, 93, 84163, 0
ipq: 56, 12852, 0, 252, 44266, 0
udp_inpcb: 336, 204809, 56, 65, 69422, 0
udpcb: 16, 204960, 56, 448, 69422, 0
tcp_inpcb: 336, 204809, 104, 83, 1417, 0
tcpcb: 880, 204800, 104, 88, 1417, 0
tcptw: 72, 27800, 0, 200, 169, 0
syncache: 144, 15366, 0, 78, 1400, 0
hostcache: 136, 15372, 4, 108, 169, 0
tcpreass: 40, 25620, 0, 252, 70, 0
sackhole: 32, 0, 0, 303, 134, 0
sctp_ep: 1272, 25602, 0, 0, 0, 0
sctp_asoc: 2240, 40000, 0, 0, 0, 0
sctp_laddr: 48, 80064, 0, 216, 5, 0
sctp_raddr: 616, 80004, 0, 0, 0, 0
sctp_chunk: 136, 400008, 0, 0, 0, 0
sctp_readq: 104, 400032, 0, 0, 0, 0
sctp_stream_msg_out: 96, 400026, 0, 0, 0, 0
sctp_asconf: 40, 400008, 0, 0, 0, 0
sctp_asconf_ack: 48, 400032, 0, 0, 0, 0
ripcb: 336, 204809, 86, 68, 3240, 0
unpcb: 240, 204800, 10, 54, 4660, 0
rtentry: 200, 0, 153, 94, 1370, 0
IPFW dynamic rule: 120, 0, 0, 0, 0, 0
selfd: 56, 0, 355, 338, 1001010507, 0
SWAPMETA: 288, 116519, 0, 0, 0, 0
ip4flow: 56, 50211, 0, 0, 0, 0
ip6flow: 80, 50220, 0, 0, 0, 0
Mountpoints: 752, 0, 5, 10, 5, 0
FFS inode: 168, 0, 59494, 60, 264556, 0
FFS1 dinode: 128, 0, 0, 0, 0, 0
FFS2 dinode: 256, 0, 59494, 146, 264556, 0
NetGraph items: 72, 128006, 4, 112, 300107892, 0
NetGraph data items: 72, 128006, 0, 203, 679841047, 0
IpAcct: 2032, 0, 0, 0, 0, 0
any ideas what should i improve?
You are should pay attention on FAILURES column
Hi,
ITEM SIZE LIMIT USED FREE REQUESTS FAILURES
64 Bucket: 536, 0, 80, 4, 80, 31
128 Bucket: 1048, 0, 184, 2, 184, 5
on another box
ITEM SIZE LIMIT USED FREE REQUESTS FAILURES
128 Bucket: 1048, 0, 109, 83, 1791, 16143
what can i do with these bucket failures?
Sami
Nothing
:)
to what failures should i give attention?
Sami
No. This is the usual behavior. With this failures all work fine.
I am not found workaround for this.
okay, thank you.
so you think that i shouldn't improve anything else, sysctl, tweak something in the source and so on?
what is the best way to get source from cvs and install it.
Thanks again,
Merry Christmass
Sami
Mpd-5.6 was released in ports few days ago. So at this moment there is no reason to take CVS sources - they are the same now.
What's about tuning, there is no much recipes of "one size fits all" type now. Most of previous were integrated to work out of the box now. Fine tuning for specific workload and hardware still could be done, but requires in-place experimenting. The only thing I can recommend now in general is to disable devd or make it to not react on interface events.
amotin,
Thanks for your reply,
>The only thing I can recommend now in general is to disable devd or make it to not react on interface events.
how can i do that?
Sami
Hi,
I read a bit about devd, why is it bad for mpd?
Sami
In /etc/devd.conf comment sections with
match "system" "IFNET";
To disable it completely you may set devd_enable="NO" in /etc/rc.conf.
Note that devd used also for other devices, such as mouses, etc.
devd is not bad, but it runs number of shell scripts on interface events. That creates additional load on system, that may be significant in time of peak loads.
Hi,
here is what i have in /etc/devd.conf
notify 0 {
match "system" "IFNET";
match "type" "ATTACH";
action "/etc/pccard_ether $subsystem start";
};
notify 0 {
match "system" "IFNET";
match "type" "LINK_UP";
media-type "ethernet";
action "/etc/rc.d/dhclient quietstart $subsystem";
};
to disable both?
Sami
BTW: is mpd5.6 introduce new features? is that reflected in the documentation?
Sami
Yes. You can simply find it in /usr/local/share/doc/mpd5/mpd.html
I know where to find it :)
the question was if it was updated with the new features that added to 5.6 version….
Hmm.. Look "Changes since version 5.5" section
I see basicly no major changes of functionality of MPD, bugfixes are top priority, should i upgrade immediatly from 5.5 to 5.6?
has anyone else tested the version already in production?
Sami