Gentoo Archives: gentoo-performance

From: Joost Roeleveld <joost@××××××××.org>
To: gentoo-performance@l.g.o
Subject: Re: [gentoo-performance] Horrible performance
Date: Thu, 26 Aug 2010 10:04:17
Message-Id: 201008261107.29598.joost@antarean.org
In Reply to: Re: [gentoo-performance] Horrible performance by Alex Schuster
1 On Wednesday 25 August 2010 21:38:10 Alex Schuster wrote:
2 > J. Roeleveld writes:
3 > > On Wednesday 25 August 2010 03:32:40 Alex Schuster wrote:
4 > > > I have an AMD Athlon(tm) Dual Core Processor 4850e CPU, on-board
5 > > > Radeon HD 3200 graphics, 4GB of memory, an 1.5 TB drive. Lots of LVM
6 > > > volumes, all encrypted, except for /usr/src and portage stuff. The
7 > > > system is ~amd64, and I have -march=k8-sse3 in my CFLAGS. Current
8 > > > kernel is 2.6.34-tuxonice, but I also tried others. I'm running KDE4
9 > > > with desktop effects enabled, X itself takes about 30-40% of CPU
10 > > > time according to top. After system startup and login into KDE, 3.5G
11 > > > of RAM are occupied. This increases after a while, and I need swap
12 > > > space. Nothing to worry about I think.
13 > >
14 > > Encrypted filesystems can cause additional with activity, but I would
15 > > expect that to remain the same over a long period.
16 >
17 > And I just moved my PORTAGE_TMPDIR to an unencrypted partition.
18 > Can LVM create noticeable overhead? I also resized my logical volumes a
19 > couple of times, could this lead to some LVM fragmentation?
20
21 Theoretically, LVM will create an additional overhead.
22 But I am extensibly using LVM on all my machines and haven't noticed any
23 significant performance drops.
24
25 LVM-fragmentation is a definite possibility.
26 To defragment it, have a look at the following:
27
28 http://bisqwit.iki.fi/source/lvm2defrag.html
29 http://www.linuxquestions.org/questions/linux-software-2/how-do-i-lvm2-defrag-
30 or-move-based-on-logical-volumes-689335/
31
32 I played with the first one on an older machine once and it does work quite
33 nicely.
34
35 > > However, how is the write and read performance on those disks?
36 >
37 > Here's the output of hdparm -t for all drives, 4 times.
38 >
39 > /dev/sda: (SATA system drive)
40 > Timing buffered disk reads: 118 MB in 3.08 seconds = 38.37 MB/sec
41 > Timing buffered disk reads: 194 MB in 3.11 seconds = 62.47 MB/sec
42 > Timing buffered disk reads: 322 MB in 3.01 seconds = 106.82 MB/sec
43 > Timing buffered disk reads: 244 MB in 3.00 seconds = 81.21 MB/sec
44 >
45 > /dev/sdb: (PATA master)
46 > Timing buffered disk reads: 114 MB in 3.02 seconds = 37.70 MB/sec
47 > Timing buffered disk reads: 114 MB in 3.00 seconds = 37.97 MB/sec
48 > Timing buffered disk reads: 116 MB in 3.05 seconds = 38.06 MB/sec
49 > Timing buffered disk reads: 116 MB in 3.05 seconds = 38.07 MB/sec
50 >
51 > /dev/sdc: (PATA slave)
52 > Timing buffered disk reads: 164 MB in 3.03 seconds = 54.21 MB/sec
53 > Timing buffered disk reads: 166 MB in 3.02 seconds = 55.04 MB/sec
54 > Timing buffered disk reads: 166 MB in 3.01 seconds = 55.10 MB/sec
55 > Timing buffered disk reads: 158 MB in 3.01 seconds = 52.41 MB/sec
56 >
57 > /dev/sdd: (SATA backup drive)
58 > Timing buffered disk reads: 314 MB in 3.00 seconds = 104.55 MB/sec
59 > Timing buffered disk reads: 312 MB in 3.01 seconds = 103.67 MB/sec
60 > Timing buffered disk reads: 308 MB in 3.01 seconds = 102.34 MB/sec
61 > Timing buffered disk reads: 314 MB in 3.00 seconds = 104.60 MB/sec
62 >
63 > The system drive throughput varies a lot, depending on other I/O.
64
65 Those look ok to me, except that I would expect SATA-drives to be faster then
66 PATA drives.
67
68 > > You're running KDE4, guess you went for the default and use mysql for
69 > > "app- office/akonadi-server".
70 > >
71 > > I switched to using "sqlite" for this due to issues getting it to work
72 > > with mysql. I think this might help there?
73 >
74 > So I only have to set the sqlite use flag and remove the mysql use flag
75 > for akonadi-server? I'm doing this now. And this also gives an example for
76 > what is going on here:
77
78 And unset mysql.
79 There is one issue that needs to be resolved manually with getting it to work
80 with sqlite.
81 See:
82 http://forums.gentoo.org/viewtopic-t-834883-view-
83 next.html?sid=3ae77f5bfba5e006e8745eedb4b6cfc4
84
85 Here is the bit that will solve the problem:
86 --
87 $ cd ~/.local/share/akonadi
88 $ sqlite3 akonadi.db
89 sqlite> INSERT INTO ResourceTable (name, isVirtual) VALUES
90 ('akonadi_search_resource', 1);
91 sqlite> .exit
92 --
93
94 > I have just logged into KDE. I did not log out since yesterday, and when
95 > started a VM with vmplayer, the system swapped like crazy, I could not use
96 > it for minutes. After this, the panel did not react any more, and the
97 > desktop background did not redraw, so I logged out and in again. The VM
98 > started fine now. Well, could be faster, but maybe it was okay.
99 > Then I started answering your mail, and tried to reemerge akonadi-server,
100 > but I had a type, so portage took a long search for akomadi-server.
101 > meanwhile the dektop became quite unresponsive, load went high, and I made
102 > a screenshot [*]. If you look at the top right, gkrellm shows this above
103 > 'Proc'. The first increase at the left was after I started emerge, the 2nd
104 > at the right was after I pressed the PrtSc key.
105
106 VMWare allows virtual machines to use more memory then is actually available.
107 Also, there are settings in VMWare (possibly enabled by default) that cause
108 the memory to be duplicated onto disk.
109 This can cause issues like you are seeing.
110
111 > > > Performance does not feel too bad at first. But after a while, I
112 > > > cannot even play videos during emerges. The playback stutters,
113 > > > sometimes I have pauses for several seconds. As long as there is no
114 > > > swap space occpied, it's not so bad I think. Maybe I have a probelm
115 > > > with disk I/O, and things get much worse when swapping occurs. When
116 > > > I look at iotop, I see various programs like chromium and various
117 > > > KDE applications appear. I guess that's normal, but should not be
118 > > > noticeable. Hey, there were times when I created a 2G tmpfs for
119 > > > /var/tmp/portage, with only 3G on my 32bit system. BTW, I lowered my
120 > > > swappiness to 10. This helps a little I think, because the swapping
121 > > > occurs later, the system is more responsive.
122 > >
123 > > Do you also encrypt swap?
124 >
125 > Yes.
126 >
127 > > Disk I/O is, in my experience, a very common cause for "freeze-ups".
128 > > Can you test with unencrypted disks to see if the issue occurs then as
129 > > well?
130 >
131 > Yes, I can do this. It's some work, but I tried so much, why not this. I
132 > have some free space, and already have written a backup script that
133 > automatically creates LVM snapshots, decrypts them, and backs it up, so I
134 > can do this from the running system.
135
136 Ok, am interested to see if running unencrypted actually has benefits here.
137
138 > > > And I have similar problems when copying data between some old PATA
139 > > > drives. When I copy stuff and do a mkfs on another partition, mplayer
140 > > > sometimes stutters and hangs for ten seconds. No joy. Working with
141 > > > KDE sucks, switching dektops sometimes takes ages, and even now I am
142 > > > typing faster than kmail can display the characters. That's with am
143 > > > emerge of chromium running, with PORTAGE_NICENESS=10 and using
144 > > > ionice -c 3. Load is around 8, but sometimes gets even higher. And
145 > > > then, load suddenly drops back to lower values, as if somthing was
146 > > > blocking. Some applications swapping, maybe.
147 > >
148 > > Very possibly, maybe an idea to check which applications are hogging
149 > > the memory. If it is the swapping of the system, then this will be
150 > > caused by the most memory-hungry processes.
151 > >
152 > > Can you post the result of: "ps axu"?
153 > > This will give an indication which processes are running and using a
154 > > lot of memory.
155 >
156 > First, here is free -m:
157 > total used free shared buffers cached
158 > Mem: 3452 3225 226 0 54 325
159 > -/+ buffers/cache: 2844 607
160 > Swap: 4094 935 3159
161 >
162 > And here the output of top, sorted by memory. I think it is a similar
163 > output to ps axu, but more condensed and better readable via mail. The
164 > full ps aux output, sorted by memory, is in [2].
165 >
166 <snipped top>
167 >
168 > X takes an awful lot, then comes java, which is running only for my
169 > tvbrowser. And many many chrome processes.
170
171 How many web browser windows do you have open? :)
172
173 Also, do you have file indexing enabled?
174
175 > > > Now I am out of ideas. I really hope someone here has one. I cannot
176 > > > work with this system any more when emerges are going on.
177 > >
178 > > Had similar issues with a desktop machine myself, managed to kill some
179 > > "features" that I wasn't using and this solved most of the problems.
180 >
181 > I hope I can say this soon, too.
182
183 In my experience, X uses more memory when a lot of windows are open.
184 And yours uses about 4 times as much as mine.
185 But then again, I don't have much running at the moment.
186
187 > > Lets see where checking for IO-speeds and memory-usage of your apps
188 > > take us :)
189 >
190 > Thanks for your help! I appreciate this very much.
191
192 I'll do my best :)
193
194 --
195 Joost

Replies

Subject Author
Re: [gentoo-performance] Horrible performance Alex Schuster <wonko@×××××××××.org>