High CPU Usage

[Serverbaboon]

Has anyone else seen high CPU usage since the the last update, a reboot fixes for a while?

The processes that eat the CPU are in bold below, just checking at the moment as work commitments have mean that this is the first time I have been able to investigate and I haven't looked at the other changes that have occurred after the upgrade yet so I don't want to raise a bug report (my last one disappeared as soon as I raised it).


I have seen a bug report for Samba, but that seems to have been fixed in Freenas before 9.3 came out.

HP NL54/16GB Ram/

last pid: 35542; load averages: 2.44, 2.61, 2.65 up 2+12:17:11 11:03:56
48 processes: 4 running, 44 sleeping

Mem: 330M Active, 127M Inact, 11G Wired, 2364K Cache, 4091M Free
ARC: 10G Total, 964M MFU, 9054M MRU, 4912K Anon, 43M Header, 33M Other
Swap: 12G Total, 12G Free


PID USERNAME THR PRI NICE SIZE RES STATE C TIME WCPU COMMAND
5871 root 3 36 0 71564K 11816K RUN 1 36.6H 69.14% syslog-ng
10699 root 1 91 0 16112K 8356K RUN 0 414:30 63.67% bzip2
35021 root 1 82 0 269M 22892K RUN 1 3:26 33.79% smbd
4697 root 12 20 0 142M 17760K uwait 1 11:15 0.00% collectd
4486 root 6 20 0 362M 170M usem 0 0:34 0.00% python2.7
3212 root 1 20 0 12040K 1572K select 1 0:18 0.00% powerd
4612 ladvd 1 20 0 22916K 2260K kqread 1 0:17 0.00% ladvd
2878 root 4 20 0 9904K 1560K rpcsvc 1 0:14 0.00% nfsd
3209 root 1 20 0 26368K 4056K select 0 0:08 0.00% ntpd
4610 root 1 20 0 22916K 2392K kqread 0 0:03 0.00% ladvd
5856 root 1 52 0 168M 61240K ttyin 1 0:02 0.00% python2.7
4912 root 4 52 0 173M 60572K select 1 0:02 0.00% python2.7
5661 root 1 43 0 20480K 2040K nanslp 1 0:01 0.00% cron
4353 root 1 32 10 18592K 3192K wait 1 0:01 0.00% sh
35198 root 1 20 0 309M 24628K select 0 0:00 0.00% smbd
4327 root 1 20 0 28092K 4272K nanslp 1 0:00 0.00% smartd
4605 nobody 1 20 0 16252K 2368K select 0 0:00 0.00% mdnsd
8511 www 1 20 0 30496K 5680K kqread 1 0:00 0.00% nginx

 

[Serverbaboon]

Will just add found this after noting a fall in CIFS performance and that the fans seemed to be louder on my microserver so wasn't sure whether this should be in the performance section.

The High CPU is a rest as well as transferring files.

As stated above I made a couple of changes around CIFS such as an auto mapping Debian client that scans my music share frequently for the UPNP server on it and I have changed the Charset a few times trying to fix some Euro Charset issues that were introduced by an Rsync process I used to run so am look at these areas at the moment.

 

[Serverbaboon]

Got the below email yesterday:


bzip2: I/O or other error, bailing out. Possible reason follows.
bzip2: No such file or directory
Input file = /var/log/messages.0, output file = /var/log/messages.0.bz2
newsyslog: `bzip2 -f /var/log/messages.0' terminated with a non-zero status (1)

This is in /var/log/messages

Jan 10 12:34:01 xxxx smbd[6215]: [2015/01/10 12:34:01.713986, 0] ../lib/util/charset/convert_string.c:438(convert_string_talloc_handle)
Jan 10 12:34:01 xxxx smbd[6215]: Conversion error: Illegal multibyte sequence(é - 11 - Mississippi-Mali Blues.flac)
Jan 10 12:34:01 xxxx smbd[6215]: [2015/01/10 12:34:01.714182, 0] ../lib/util/charset/convert_string.c:438(convert_string_talloc_handle)
Jan 10 12:34:01 xxxx smbd[6215]: Conversion error: Illegal multibyte sequence(é - 11 - Mississippi-Mali Blues.flac)
Jan 10 12:34:01 xxxx smbd[6215]: [2015/01/10 12:34:01.714693, 0] ../lib/util/charset/convert_string.c:438(convert_string_talloc_handle)

Which suggests its to do with my messing around with euro charsets causing excessive errors being logged when my Debian box scans my CIFS share, since this is something I have done is it worth filing a bug around this logging rollover(?) issue?