I may soon become a fan of static files if this kind of thing keeps up.
Here is the output that you are referring to.
Dynamic Files:
Slot # Inode Device Ref Count Htype Split Merge Curmod Basemod
Largerec Filesp Selects Nextsplit
0 1900545 59425 1 20 80 50 2 2
1628 2344 0 1
1 1835009 59425 110 20 80 50 1943 1024
1628 3223064 0 920
2 4177921 59425 108 20 80 50 28333 16384
809 23500000 0 11950
3 4063233 59425 51 20 80 50 33579 32768
809 27851124 0 812
4 4472833 59425 5 20 80 50 11563 8192
809 9590804 0 3372
5 4194305 59425 52 20 80 50 5286 4096
809 4384076 0 1191
6 3325953 59425 7 20 80 50 2 2
809 1248 0 1
7 884737 59425 125 20 80 50 138 128
1617 227492 0 11
8 950273 59425 51 20 80 50 45 32
1628 74316 0 14
9 4603905 59425 7 20 80 50 1 1
809 332 0 1
10 3440641 59425 21 20 80 50 285 256
809 235552 0 30
11 3522561 59425 5 20 80 50 469 256
1628 776848 0 214
12 4587521 59425 7 20 80 50 31 16
809 20540 0 16
13 1032193 59425 5 20 80 50 2 2
1012 0 0 1
14 5423105 59425 5 20 80 50 95 64
809 78012 0 32
15 5505025 59425 107 20 80 50 5277 4096
1628 5443692 0 1182
16 1802241 59425 6 20 80 50 61141 32768
1628 93494588 0 28374
17 2097163 59425 1 20 80 50 1 1
1628 116 0 1
18 3915785 59425 1 20 80 50 1 1
1628 440 0 1
19 1146881 59425 5 20 80 50 12076 8192
809 10015980 0 3885
20 1015809 59425 117 20 80 50 2566 2048
1628 2645064 0 519
21 1736710 59425 6 20 80 50 2018 1024
1628 3346044 0 995
22 409601 59425 5 20 80 50 1 1
1628 0 0 1
23 4145153 59425 11 20 80 50 6592 4096
1628 10933940 0 2497
24 5406723 59425 5 20 80 50 17 16
1628 18748 0 2
25 2785285 59425 5 20 80 50 12 8
1628 16128 0 5
26 1654785 59425 8 20 80 50 959 512
809 794616 0 448
27 1916937 59425 5 20 80 50 1 1
1628 0 0 1
28 2572297 59425 5 20 80 50 70 64
1628 114548 0 7
29 2244617 59425 5 20 80 50 749 512
1628 1241084 0 238
30 2998273 59425 1 20 80 50 732 512
809 606688 0 221
31 4800513 59425 57 20 80 50 511 256
809 421456 0 256
32 5160961 59425 83 20 80 50 8 8
798 5740 0 1
33 5685249 59425 22 20 80 50 158 128
798 124812 0 31
34 5668865 59425 22 20 80 50 7504 4096
798 5827804 0 3409
35 3784707 59425 19 20 80 50 7050 4096
1628 11693600 0 2955
36 2392073 59425 1 20 80 50 130 128
1628 215180 0 3
37 6062081 59425 84 20 80 50 165 128
809 134864 0 38
38 4521985 59425 44 20 80 50 3 2
809 1736 0 2
39 540673 59425 34 20 80 50 1178 1024
1628 1952808 0 155
40 851969 59425 35 20 80 50 160 128
1628 264580 0 33
41 999425 59425 34 20 80 50 6038 4096
809 5007736 0 1943
42 3293185 59425 34 20 80 50 338 256
1628 560272 0 83
43 1409029 59425 34 20 80 50 293 256
1628 484500 0 38
44 1671177 59425 34 20 80 50 6123 4096
1628 10156060 0 2028
45 1638409 59425 34 20 80 50 6094 4096
1628 10108308 0 1999
46 1933321 59425 34 20 80 50 4849 4096
1628 8042972 0 754
47 2686981 59425 34 20 80 50 757 512
1628 1255440 0 246
48 1540101 59425 34 20 80 50 70 64
1628 115976 0 7
49 245761 59425 34 20 80 50 166730 131072
1628 76583768 0 35659
50 606209 59425 38 20 80 50 167819 131072
1628 78390732 0 36748
51 5439489 59425 58 20 80 50 33 32
809 26760 0 2
52 5914625 59425 34 20 80 50 2712 2048
1628 4495256 0 665
53 5537797 59425 37 20 80 50 8227 8192
1628 13647504 0 36
54 3129349 59425 34 20 80 50 7 4
1628 10656 0 4
55 4112385 59425 22 20 80 50 34 32
809 28084 0 3
56 4947969 59425 53 20 80 50 11290 8192
809 9363600 0 3099
57 4653057 59425 53 20 80 50 84 64
1617 138036 0 21
58 4784129 59425 42 20 80 50 12800 8192
809 10616204 0 4609
59 1392641 59425 103 20 80 50 49572 32768
1628 82233368 0 16805
60 2801665 59425 53 20 80 50 19154 16384
1628 31772772 0 2771
61 770049 59425 1 20 80 50 1 1
1617 632 0 1
62 3358721 59425 1 20 80 50 29 16
798 23828 0 14
63 2179073 59425 16 20 80 50 1210 1024
1628 2006556 0 187
64 3391489 59425 16 20 80 50 2340 2048
809 1940700 0 293
65 3457025 59425 1 20 80 50 4 4
798 2720 0 1
66 3719169 59425 1 20 80 50 14 8
1012 48 0 7
67 4964353 59425 5 20 80 50 114 64
809 92060 0 51
68 2260993 59425 2 20 80 50 2928 2048
1628 4856788 0 881
69 5570565 59425 2 20 80 50 1389 1024
1628 2302952 0 366
70 688129 59425 16 20 80 50 67779 65536
1628 12435760 0 2244
71 4374529 59425 24 20 80 50 1629 1024
809 1350312 0 606
72 5029889 59425 23 20 80 50 8 8
1012 152 0 1
73 5947397 59425 1 20 80 50 2443 2048
1628 4051380 0 396
74 5095425 59425 23 20 80 50 1 1
798 452 0 1
75 5013505 59425 16 20 80 50 1235 1024
809 1022696 0 212
76 5144577 59425 16 20 80 50 948 512
809 784832 0 437
78 5128193 59425 2 20 80 50 66 64
809 54244 0 3
79 5079041 59425 2 20 80 50 8724 8192
809 7235280 0 533
80 4128769 59425 2 20 80 50 3317 2048
1628 5501132 0 1270
81 3883013 59425 1 20 80 50 3067 2048
1628 5087236 0 1020
83 3932161 59425 1 20 80 50 1770 1024
1628 2935060 0 747
84 983041 59425 8 20 80 50 30176 16384
1628 50056712 0 13793
85 5652481 59425 8 20 80 50 9197 8192
1012 7627696 0 1006
86 4161538 59425 8 20 80 50 19960 16384
1628 33109872 0 3577
87 344065 59425 7 20 80 50 18 16
1012 14180 0 3
88 376833 59425 7 20 80 50 2 2
1012 556 0 1
89 393217 59425 7 20 80 50 5 4
1012 3312 0 2
90 425985 59425 7 20 80 50 2 2
1012 0 0 1
91 458753 59425 7 20 80 50 719 512
809 595388 0 208
92 475137 59425 7 20 80 50 2 2
1012 328 0 1
93 491521 59425 7 20 80 50 6 4
1012 4248 0 3
94 507905 59425 7 20 80 50 2 2
1628 272 0 1
95 524289 59425 7 20 80 50 2 2
1012 348 0 1
96 360449 59425 7 20 80 50 724 512
1012 600092 0 213
I had no reason for messing with MFILES except is seemed low and
reasonable based in the desc in uvconfig. I will probably put it back
to 12 since it didn't do anything.
T30FILE is set to the default of 200.
Based on the above it looks like a fair amoutn of file sizing is in
order.
Anthony
-----Original Message-----
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] On Behalf Of Dan Fitzgerald
Sent: Monday, February 14, 2005 2:03 PM
To: [email protected]
Subject: RE: [U2] Universe/Linux Performance Issue
Just to see... try an analyze.shm -d, and see how many dynamic files you
have (<aside/> My personal opinion is that anything above "0" is
probably
too many, with some special exceptions. My saying this usually sparks a
debate</aside>). Then look at the T30FILES parameter in uvconfig.
Also, in looking at the analyze.shm -d, if current mod exceeds base mod,
you
should resize that file. As it is listed by device & inode number, do a
find
. -inum <inode #> to identify the file.
"Our greatest duty in this life is to help others. And please, if you
can't
help them, could you at least not hurt them?" - H.H. the Dalai Lama
"When buying & selling are controlled by legislation, the first thing to
be
bought & sold are the legislators" - P.J. O'Rourke
Dan Fitzgerald
>From: "Anthony Dzikiewicz" <[EMAIL PROTECTED]>
>Reply-To: [email protected]
>To: <[email protected]>
>Subject: RE: [U2] Universe/Linux Performance Issue
>Date: Mon, 14 Feb 2005 11:20:39 -0500
>
>-----Original Message-----
>From: [EMAIL PROTECTED]
>[mailto:[EMAIL PROTECTED] On Behalf Of Scott
>Richardson
>Sent: Monday, February 14, 2005 9:51 AM
>To: [email protected]
>Subject: Re: [U2] Universe/Linux Performance Issue
>
>
>Moving/Editing a parameter in uvconfig is only a small step in process
>of making that change active in your systems actively running uv.
>Simply editing the uvconfig file itself does not dynamically change the
>currently running uv.
>
>You'll have to re-generate, or build, UV after making those changes,
>and that, of course will require uv being stopped, and restarted after
>the new parameter changes being made in the uvconfig file. Please
>consult the appropriate official UV documentation for details on how to
>do this, or ask for assistance from a consultant if you're not
>comfortable with these types of things.
>
>---
>This much I do know an have done. Somethings read are understandable
>in theory, but what they really do is sometimes another thing in the
>real world. Experience is everything.
>----
>
>If you make any Linux kernel parameter changes, same concept applies:
>you'll need to rebuild the Linux kernel and reboot for those to take
>affect. Again, if not comfortable here, well worth having a consultant
>come in for a few hours and help you with both tasks, and you could
>take notes and learn, in case you need to tackle such tasks again in
>the future.
>
>---
>No Linux changes are required. Linux works and doesn't seem to be the
>issue
>---
>
>
>As for file sizing - just becasue those files have always been about
>that size, and always been Dynamic, doesn't mean that they're correct.
>Again, if not comfortable, ...
>
>--
>Remember that I also said that we were on DG/UX and it always handled
>this with no problem. My feeling is that the particular version of
>Universe for DG/UX might have been tweaked a little. One time when I
>was up at the DG benchmark lab in Marlboro, MA, I was told that they
>had source code for Universe and had made some changes that had to do
>with posix compliance. So, Im guessing that maybe the DG version
>multi-threaded this a little better.
>---
>
>
>I'd suggest the FAST File Sizing tool/application from Fitzgerald &
>Long.
>
>It could be that the overhead of Dynamic Files in your situation may
>not be worth the benefit, especially if those files are not growing
>over time. Actual Mileages can and do vary here, of course.
>
>HTH!
>
>----- Original Message -----
>From: "Anthony Dzikiewicz" <[EMAIL PROTECTED]>
>To: <[email protected]>
>Sent: Monday, February 14, 2005 9:31 AM
>Subject: RE: [U2] Universe/Linux Performance Issue
>
>
> > -----Original Message-----
> > From: [EMAIL PROTECTED]
> > [mailto:[EMAIL PROTECTED] On Behalf Of Scott
> > Richardson
> > Sent: Monday, February 14, 2005 8:31 AM
> > To: [email protected]
> > Subject: Re: [U2] Universe/Linux Performance Issue
> >
> >
> > In addition to what Trevor mentions, you may want to have a look at
> > the uvconfig parameters, as well as the underlying OS kernel
> > parameters, including file system settings, etc.
> >
> > I don't really have experience with messing with these params. I
> > did increase MFILES to 300 from 12
> >
> >
> > Also, file system fragmentation may be an issue, depending on how
> > long
>
> > it's been since it was looked at and addressed, especially if there
> > is
>
> > a lot of reading / writing going on with this file system.
> >
> > How is your /tmp space, and swap space?
> >
> > No issues there. /tmp is a plenty and swap is never more than 1/2
> > used
> >
> > What flavor and versions of Linux, and UV?
> >
> > Red Hat 3
> >
> > This is a process that your run when other users are not on the
> > system, and possibly have these files open?
> >
> > Tried both ways. Last night I got in when no one was on. I shut
> > down
>
> > UV and brought it back up fresh - same results.
> >
> > You are correct to first suspect file sizing. Depending on how
> > properly, or imporperly tuned they are, Dynamic files can introduce
> > some overhead.
> >
> > Just a few ideas off top of the head...
> >
> > Appreciated - Thanks
> >
> > Anthony
> >
> >
> >
> > ----- Original Message -----
> > From: "Anthony Dzikiewicz" <[EMAIL PROTECTED]>
> > To: <[email protected]>
> > Sent: Sunday, February 13, 2005 7:11 PM
> > Subject: [U2] Universe/Linux Performance Issue
> >
> >
> > > I am experiencing some strange behavior on our Linux machine. We
> > > have
> > been
> > > running on this machine since October. I am just getting into
> > > some file cleanup, year end rolls, etc.. What is happening is
> > > that the machine
> > comes
> > > down to a crawl.
> > >
> > > We were on DG/UX and this kind of thing never happened on that
> > > platform.
> > I
> > > could 'roll' files, etc and I never experienced anything like
> > > this.
> > >
> > > Linux itself seems to be moving along fine. Anything in Universe
> > > takes almost 1/2 minute to get any response.
> > >
> > > Without the purge/cleanup programs running, everything is just
> > > peachy.
> > >
> > > The files are all Dynamic. The programs are doing readu/delete,
> > > write. That is read a record and write it to a history file and
> > > delete
> >
> > > from
> > current
> > > file. Or read a record, move this years data to last years fields
> > > and
> > write
> > > the record back out.
> > >
> > > Anybody experience this sort of thing ? Anyway to tweak things up
> > > ?
>
> > > I
> > have
> > > an idea that the files might not be 'sized' correctly. I also
> > > remember
> > that
> > > Universe on DG/UX was tweaked specfically for that OS. So, maybe
> > > it
>
> > > was just more efficient at doing this kind of thing.
> > >
> > > Any suggestions would be appreciated.
> > >
> > > Thanks
> > > Anthony
> > >
> > > --
> > > No virus found in this outgoing message.
> > > Checked by AVG Anti-Virus.
> > > Version: 7.0.300 / Virus Database: 265.8.7 - Release Date:
> > > 2/10/2005
> > > -------
> > > u2-users mailing list
> > > [email protected]
> > > To unsubscribe please visit http://listserver.u2ug.org/
> > -------
> > u2-users mailing list
> > [email protected]
> > To unsubscribe please visit http://listserver.u2ug.org/
> > -------
> > u2-users mailing list
> > [email protected]
> > To unsubscribe please visit http://listserver.u2ug.org/
>-------
>u2-users mailing list
>[email protected]
>To unsubscribe please visit http://listserver.u2ug.org/
>-------
>u2-users mailing list
>[email protected]
>To unsubscribe please visit http://listserver.u2ug.org/
-------
u2-users mailing list
[email protected]
To unsubscribe please visit http://listserver.u2ug.org/
-------
u2-users mailing list
[email protected]
To unsubscribe please visit http://listserver.u2ug.org/
