1 |
J. Roeleveld writes: |
2 |
|
3 |
> On Wednesday 25 August 2010 03:32:40 Alex Schuster wrote: |
4 |
|
5 |
> > I have an AMD Athlon(tm) Dual Core Processor 4850e CPU, on-board |
6 |
> > Radeon HD 3200 graphics, 4GB of memory, an 1.5 TB drive. Lots of LVM |
7 |
> > volumes, all encrypted, except for /usr/src and portage stuff. The |
8 |
> > system is ~amd64, and I have -march=k8-sse3 in my CFLAGS. Current |
9 |
> > kernel is 2.6.34-tuxonice, but I also tried others. I'm running KDE4 |
10 |
> > with desktop effects enabled, X itself takes about 30-40% of CPU |
11 |
> > time according to top. After system startup and login into KDE, 3.5G |
12 |
> > of RAM are occupied. This increases after a while, and I need swap |
13 |
> > space. Nothing to worry about I think. |
14 |
> |
15 |
> Encrypted filesystems can cause additional with activity, but I would |
16 |
> expect that to remain the same over a long period. |
17 |
|
18 |
And I just moved my PORTAGE_TMPDIR to an unencrypted partition. |
19 |
Can LVM create noticeable overhead? I also resized my logical volumes a |
20 |
couple of times, could this lead to some LVM fragmentation? |
21 |
|
22 |
> However, how is the write and read performance on those disks? |
23 |
|
24 |
Here's the output of hdparm -t for all drives, 4 times. |
25 |
|
26 |
/dev/sda: (SATA system drive) |
27 |
Timing buffered disk reads: 118 MB in 3.08 seconds = 38.37 MB/sec |
28 |
Timing buffered disk reads: 194 MB in 3.11 seconds = 62.47 MB/sec |
29 |
Timing buffered disk reads: 322 MB in 3.01 seconds = 106.82 MB/sec |
30 |
Timing buffered disk reads: 244 MB in 3.00 seconds = 81.21 MB/sec |
31 |
|
32 |
/dev/sdb: (PATA master) |
33 |
Timing buffered disk reads: 114 MB in 3.02 seconds = 37.70 MB/sec |
34 |
Timing buffered disk reads: 114 MB in 3.00 seconds = 37.97 MB/sec |
35 |
Timing buffered disk reads: 116 MB in 3.05 seconds = 38.06 MB/sec |
36 |
Timing buffered disk reads: 116 MB in 3.05 seconds = 38.07 MB/sec |
37 |
|
38 |
/dev/sdc: (PATA slave) |
39 |
Timing buffered disk reads: 164 MB in 3.03 seconds = 54.21 MB/sec |
40 |
Timing buffered disk reads: 166 MB in 3.02 seconds = 55.04 MB/sec |
41 |
Timing buffered disk reads: 166 MB in 3.01 seconds = 55.10 MB/sec |
42 |
Timing buffered disk reads: 158 MB in 3.01 seconds = 52.41 MB/sec |
43 |
|
44 |
/dev/sdd: (SATA backup drive) |
45 |
Timing buffered disk reads: 314 MB in 3.00 seconds = 104.55 MB/sec |
46 |
Timing buffered disk reads: 312 MB in 3.01 seconds = 103.67 MB/sec |
47 |
Timing buffered disk reads: 308 MB in 3.01 seconds = 102.34 MB/sec |
48 |
Timing buffered disk reads: 314 MB in 3.00 seconds = 104.60 MB/sec |
49 |
|
50 |
The system drive throughput varies a lot, depending on other I/O. |
51 |
|
52 |
I |
53 |
|
54 |
|
55 |
> You're running KDE4, guess you went for the default and use mysql for |
56 |
> "app- office/akonadi-server". |
57 |
> |
58 |
> I switched to using "sqlite" for this due to issues getting it to work |
59 |
> with mysql. I think this might help there? |
60 |
|
61 |
So I only have to set the sqlite use flag and remove the mysql use flag |
62 |
for akonadi-server? I'm doing this now. And this also gives an example for |
63 |
what is going on here: |
64 |
|
65 |
I have just logged into KDE. I did not log out since yesterday, and when |
66 |
started a VM with vmplayer, the system swapped like crazy, I could not use |
67 |
it for minutes. After this, the panel did not react any more, and the |
68 |
desktop background did not redraw, so I logged out and in again. The VM |
69 |
started fine now. Well, could be faster, but maybe it was okay. |
70 |
Then I started answering your mail, and tried to reemerge akonadi-server, |
71 |
but I had a type, so portage took a long search for akomadi-server. |
72 |
meanwhile the dektop became quite unresponsive, load went high, and I made |
73 |
a screenshot [*]. If you look at the top right, gkrellm shows this above |
74 |
'Proc'. The first increase at the left was after I started emerge, the 2nd |
75 |
at the right was after I pressed the PrtSc key. |
76 |
|
77 |
|
78 |
> > Performance does not feel too bad at first. But after a while, I |
79 |
> > cannot even play videos during emerges. The playback stutters, |
80 |
> > sometimes I have pauses for several seconds. As long as there is no |
81 |
> > swap space occpied, it's not so bad I think. Maybe I have a probelm |
82 |
> > with disk I/O, and things get much worse when swapping occurs. When |
83 |
> > I look at iotop, I see various programs like chromium and various |
84 |
> > KDE applications appear. I guess that's normal, but should not be |
85 |
> > noticeable. Hey, there were times when I created a 2G tmpfs for |
86 |
> > /var/tmp/portage, with only 3G on my 32bit system. BTW, I lowered my |
87 |
> > swappiness to 10. This helps a little I think, because the swapping |
88 |
> > occurs later, the system is more responsive. |
89 |
> |
90 |
> Do you also encrypt swap? |
91 |
|
92 |
Yes. |
93 |
|
94 |
> Disk I/O is, in my experience, a very common cause for "freeze-ups". |
95 |
> Can you test with unencrypted disks to see if the issue occurs then as |
96 |
> well? |
97 |
|
98 |
Yes, I can do this. It's some work, but I tried so much, why not this. I |
99 |
have some free space, and already have written a backup script that |
100 |
automatically creates LVM snapshots, decrypts them, and backs it up, so I |
101 |
can do this from the running system. |
102 |
|
103 |
|
104 |
> > And I have similar problems when copying data between some old PATA |
105 |
> > drives. When I copy stuff and do a mkfs on another partition, mplayer |
106 |
> > sometimes stutters and hangs for ten seconds. No joy. Working with |
107 |
> > KDE sucks, switching dektops sometimes takes ages, and even now I am |
108 |
> > typing faster than kmail can display the characters. That's with am |
109 |
> > emerge of chromium running, with PORTAGE_NICENESS=10 and using |
110 |
> > ionice -c 3. Load is around 8, but sometimes gets even higher. And |
111 |
> > then, load suddenly drops back to lower values, as if somthing was |
112 |
> > blocking. Some applications swapping, maybe. |
113 |
> |
114 |
> Very possibly, maybe an idea to check which applications are hogging |
115 |
> the memory. If it is the swapping of the system, then this will be |
116 |
> caused by the most memory-hungry processes. |
117 |
> |
118 |
> Can you post the result of: "ps axu"? |
119 |
> This will give an indication which processes are running and using a |
120 |
> lot of memory. |
121 |
|
122 |
First, here is free -m: |
123 |
total used free shared buffers cached |
124 |
Mem: 3452 3225 226 0 54 325 |
125 |
-/+ buffers/cache: 2844 607 |
126 |
Swap: 4094 935 3159 |
127 |
|
128 |
And here the output of top, sorted by memory. I think it is a similar |
129 |
output to ps axu, but more condensed and better readable via mail. The |
130 |
full ps aux output, sorted by memory, is in [2]. |
131 |
|
132 |
top - 21:23:03 up 1 day, 7 min, 11 users, load average: 0.04, 0.05, 0.02 |
133 |
Tasks: 418 total, 1 running, 417 sleeping, 0 stopped, 0 zombie |
134 |
Cpu(s): 17.8%us, 12.7%sy, 9.9%ni, 45.9%id, 13.5%wa, 0.1%hi, 0.1%si, 0.0%st |
135 |
Mem: 3534936k total, 3312312k used, 222624k free, 56264k buffers |
136 |
Swap: 4193272k total, 957908k used, 3235364k free, 334048k cached |
137 |
|
138 |
PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND |
139 |
25993 root 20 0 847m 475m 47m S 21 13.8 49:01.11 X |
140 |
26553 wonko 20 0 1312m 235m 10m S 0 6.8 0:28.32 java |
141 |
26925 wonko 20 0 698m 112m 18m S 0 3.2 3:30.75 kontact |
142 |
26961 wonko 20 0 526m 79m 13m S 0 2.3 2:16.01 chrome |
143 |
26566 wonko 20 0 1128m 72m 10m S 0 2.1 1:02.46 amarok |
144 |
27056 wonko 20 0 871m 57m 19m S 0 1.7 0:05.07 chrome |
145 |
30324 root 30 10 195m 55m 1136 S 0 1.6 0:09.46 emerge |
146 |
27051 wonko 20 0 874m 50m 8668 S 0 1.5 0:14.16 chrome |
147 |
26126 wonko 20 0 1028m 44m 11m S 2 1.3 9:18.94 plasma-desktop |
148 |
27046 wonko 20 0 878m 41m 7324 S 0 1.2 0:16.59 chrome |
149 |
27122 wonko 20 0 871m 36m 7660 S 0 1.0 0:40.03 chrome |
150 |
27036 wonko 20 0 865m 29m 7492 S 0 0.8 0:00.94 chrome |
151 |
27198 wonko 20 0 860m 28m 6240 S 0 0.8 0:04.76 chrome |
152 |
26101 wonko 20 0 427m 28m 8672 S 4 0.8 5:27.48 kwin |
153 |
26298 wonko 20 0 396m 26m 9.9m S 0 0.8 0:02.91 knotes |
154 |
27766 wonko 20 0 857m 26m 5924 S 0 0.8 0:03.03 chrome |
155 |
26903 root 20 0 64316 25m 288 S 0 0.7 0:00.33 screen |
156 |
27203 wonko 20 0 865m 24m 5960 S 0 0.7 0:04.06 chrome |
157 |
30226 wonko 20 0 367m 23m 9232 S 0 0.7 0:03.21 gwenview |
158 |
26221 wonko 20 0 609m 23m 3752 S 0 0.7 0:02.55 knotify4 |
159 |
|
160 |
X takes an awful lot, then comes java, which is running only for my |
161 |
tvbrowser. And many many chrome processes. |
162 |
|
163 |
> > Now I am out of ideas. I really hope someone here has one. I cannot |
164 |
> > work with this system any more when emerges are going on. |
165 |
> |
166 |
> Had similar issues with a desktop machine myself, managed to kill some |
167 |
> "features" that I wasn't using and this solved most of the problems. |
168 |
|
169 |
I hope I can say this soon, too. |
170 |
|
171 |
> Lets see where checking for IO-speeds and memory-usage of your apps |
172 |
> take us :) |
173 |
|
174 |
Thanks for your help! I appreciate this very much. |
175 |
|
176 |
Wonko |
177 |
|
178 |
[1] http://www.wonkology.org/comp/desktop/2010-08-25_emerge-akomadi.png |
179 |
[2] http://www.wonkology.org/gentoo/ |