Menu
▾
▴
genex-dev — GeneX developers
You can subscribe to this list here.
| 2001 |
Jan
(135) |
Feb
(57) |
Mar
(84) |
Apr
(43) |
May
(77) |
Jun
(51) |
Jul
(21) |
Aug
(55) |
Sep
(37) |
Oct
(56) |
Nov
(75) |
Dec
(23) |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 2002 |
Jan
(32) |
Feb
(174) |
Mar
(121) |
Apr
(70) |
May
(55) |
Jun
(20) |
Jul
(23) |
Aug
(15) |
Sep
(12) |
Oct
(58) |
Nov
(203) |
Dec
(90) |
| 2003 |
Jan
(37) |
Feb
(15) |
Mar
(14) |
Apr
(57) |
May
(7) |
Jun
(40) |
Jul
(36) |
Aug
(1) |
Sep
(56) |
Oct
(38) |
Nov
(105) |
Dec
(2) |
| 2004 |
Jan
|
Feb
(117) |
Mar
(69) |
Apr
(160) |
May
(165) |
Jun
(35) |
Jul
(7) |
Aug
(80) |
Sep
(47) |
Oct
(23) |
Nov
(8) |
Dec
(42) |
| 2005 |
Jan
(19) |
Feb
(2) |
Mar
|
Apr
|
May
|
Jun
|
Jul
|
Aug
|
Sep
|
Oct
|
Nov
|
Dec
|
| 2007 |
Jan
|
Feb
|
Mar
|
Apr
|
May
|
Jun
(2) |
Jul
|
Aug
|
Sep
|
Oct
|
Nov
|
Dec
|
| 2010 |
Jan
|
Feb
(2) |
Mar
|
Apr
|
May
|
Jun
|
Jul
|
Aug
|
Sep
|
Oct
|
Nov
|
Dec
|
|
From: <ja...@op...> - 2001-06-15 23:39:30
|
"Joe Roden" <ro...@ai...> writes: > > Can you reconfigure GeneX to *not* use tcp/ip to connect with the DB, > > and give that a try. > > Will do, and that may well do the trick. That sounds like the sort of > tweak that could cause the problem. I'm attempting to recreate your problem at my end, I'll let you know if that does the trick. jas. |
|
From: Harry M. <man...@ho...> - 2001-06-15 23:35:16
|
Hmm - Jason's probably right - I misread your pg_hba.conf to mean that you had allowed /only/ domain sockets (altho it's pretty obvious in retrospect that you had allowed other hosts to connect via tcp). Ugh - hope you got his answer before you tried re-installing postgres from scratch. On the other hand, re-installing the new version of the DB and GeneX-Server are still a good idea... hjm "Jason E. Stewart" wrote: > > This documentation describes how connect behaves given different > > combinations of arguments, and reminded me that PostgreSQL > > authenticates users and allows connections according to the file: > > > > /var/lib/pgsql/data/pg_hba.conf > > > > For our installation, this file looks like: > > > > # By default, allow anything over UNIX domain sockets and localhost. > > local all trust > > host all 127.0.0.1 255.255.255.255 trust > > host all 131.215.34.21 255.255.255.255 crypt > > host all 131.215.34.24 255.255.255.255 crypt > > host all 131.215.34.25 255.255.255.255 crypt > > host all 131.215.34.38 255.255.255.255 crypt > > host all 131.215.34.70 255.255.255.255 crypt > > This is most likely the culprit. Even though this *ought* to work, I'm > sure that every installation we've run uses Unix sockets and *not* > tcp/ip sockets to communicate with the DB. So I bet there is some > authentication issue that we had not considered. > > Can you reconfigure GeneX to *not* use tcp/ip to connect with the DB, > and give that a try. -- Cheers, Harry Harry J Mangalam -- (949) 856 2847 (v&f) -- hj...@nc... || man...@ho... |
|
From: Joe R. <ro...@ai...> - 2001-06-15 23:24:11
|
Thanks Harry & Jason for your quick responses. > Which verson of postgres are you using? We're running PostgreSQL 7.0.3. Certainly it will be worth moving GeneX to 1.0.4 as soon as time permits. > Can you reconfigure GeneX to *not* use tcp/ip to connect with the DB, > and give that a try. Will do, and that may well do the trick. That sounds like the sort of tweak that could cause the problem. Harry wrote... > Incidentally, what technologies are you using? We're using mostly custom 2-color microarrays, though we have collaborators with Affy setups as well. We've done some in-house work on an object database & web front-end work tracking lab materials through cluster analysis results, but we're lacking the support and commitment needed to develop it further. We do lots of cluster analyses, and are pushing towards gene regulatory network models. Jason wrote... > Do you think there would be any interest in having me come > and give a GeneX talk? That sounds quite doable. We'll contact you about this- Thanks! Joe |
|
From: <ja...@op...> - 2001-06-15 22:56:11
|
> Hi, we're interested in giving GeneX a try, so we've installed the > software on our linux cluster. Hi, we're interested in making sure that GeneX works ;-) > But we've run into a small snag that I've decided to ask for your > help on. It's likely a trivial problem (probably not even a bug!) > but it's holding things up a bit. In case other users might hit > this problem I thought I'd pass it by you. Thanks! To reach a wider audience, you can use the 'genex-dev' list (to which I've Cc:ed this reply). That way more people are both likely to see and reply. > This documentation describes how connect behaves given different > combinations of arguments, and reminded me that PostgreSQL > authenticates users and allows connections according to the file: > > /var/lib/pgsql/data/pg_hba.conf > > For our installation, this file looks like: > > # By default, allow anything over UNIX domain sockets and localhost. > local all trust > host all 127.0.0.1 255.255.255.255 trust > host all 131.215.34.21 255.255.255.255 crypt > host all 131.215.34.24 255.255.255.255 crypt > host all 131.215.34.25 255.255.255.255 crypt > host all 131.215.34.38 255.255.255.255 crypt > host all 131.215.34.70 255.255.255.255 crypt This is most likely the culprit. Even though this *ought* to work, I'm sure that every installation we've run uses Unix sockets and *not* tcp/ip sockets to communicate with the DB. So I bet there is some authentication issue that we had not considered. Can you reconfigure GeneX to *not* use tcp/ip to connect with the DB, and give that a try. Re-run install-all.pl and it should automatically detect the existence of your previous configuration file and ask you whether you want to run in 'automatic' mode or 'prompting' mode. You want 'prompting' (so that you can change some of the config params). The one you want to change is prefixed by '*** GeneX DB Host Configuration ***'. Instead of entering the true hostname, use 'localhost'. The rest should be the same. NOTE: this will only work if your http server is running on the same machine as your postmaster daemon. > P.S. I'm working with Barbara Wold's group at Caltech, assisting > with data analysis & management; (some of us are at JPL, but most of > the group is at Caltech, which runs JPL). A few other groups at > Caltech and USC looking to us for microarray database solutions as > well, which is why we're looking at community projects > vs. home-grown solutions. Probably lots of the analyses we do could > be connected up or added to GeneX, though Caltech likes to keep > rights. I see how you've handled connections to other licensed > software, which could probably work for Caltech as well. So if we > get GeneX going and generate sufficient interest, we'll have more to > discuss. Interesting. I was an undergrad at Caltech in the '80s. Barbara started there at about the same time. I remember her as an ambitious and somewhat stressed-out researcher going for tenure. She was a fun instructor, and I was glad to hear she had gotten tenure. I'm looking for a good excuse to get to SoCal again to visit some Tech friends. Do you think there would be any interest in having me come and give a GeneX talk? It would be related to my MGED talk -- although I promise not to go so badly overtime :-( I could even work together with your group and help get things set up if need be. Cheers, jas. |
|
From: Harry M. <man...@ho...> - 2001-06-15 22:17:57
|
Hi Joe, We'll help out as much as possible... Joe Roden wrote: > > Hi, we're interested in giving GeneX a try, so we've installed the > software on our linux cluster. But we've run into a small snag that > I've decided to ask for your help on. It's likely a trivial problem > (probably not even a bug!) but it's holding things up a bit. In case > other users might hit this problem I thought I'd pass it by you. > > [ Just noticed I've been working with GeneX 1.0.0, which is now > getting a bit old. So maybe this whole message should be taken with > a grain of salt... please take a look anyway since I wrote up this > whole damn message! If one of the newer releases addresses this just > tell me and I'll be apologetic. ] ..or apoplectic.. :( You HAVE been digging deep. It wouldn't hurt to update to the latest version so we're onteh same page. there have been some reasonably large improvements lately and it should be a trivial update (just download the recent tarball, unpack it, run the installer again and select 'Update' instead of 'New' when prompted. It updates the whole package in-place. You also might want to re-install the DB (drop the genex database and recreate it automatically via the included 'init-dump.pl' script that's part of the installation. That way you know that you're up-to-date. Which verson of postgres are you using? We recently switched to requirng version >=7.0. And I couldn't read this off your status page - not related to your installation - we've seen this happen to other installations and are trying to debug it. Can you do a 'normal' query on the DB? Hmmm.. no you can't - I just tried. This looks like either a Postgres updating issue or a GeneX db versioning incompatibility. My suggestions (in order of difficulty for you) would be to: 1) update the GeneX distro as above. The new one is at: http://genex.ncgr.org/genex/download/genex-server/GeneX-Server-1.0.4.tar.gz 2) drop the genex database and recreate it by means noted above? The script will get the new dump for you, but if not, it's located at: http://genex.ncgr.org/genex/download/DB/genex-initdb-latest.gz 3) if neither of these work and you're using a version 6 postgres DB, definitely upgrade it. This will be a bit of a pain if you've installed it from source, but pretty straightforward if via rpm or deb. [enormous amount of debugging deleted] > Thanks! > > Joe Roden > JPL/Caltech > (818) 393-5348 > > P.S. I'm working with Barbara Wold's group at Caltech, assisting with > data analysis & management; (some of us are at JPL, but most of the > group is at Caltech, which runs JPL). A few other groups at Caltech > and USC looking to us for microarray database solutions as well, which > is why we're looking at community projects vs. home-grown solutions. > Probably lots of the analyses we do could be connected up or added to > GeneX, though Caltech likes to keep rights. I see how you've handled > connections to other licensed software, which could probably work for > Caltech as well. So if we get GeneX going and generate sufficient > interest, we'll have more to discuss. This sounds like it would be valuable for both of us. We'd be very interested in keeping in touch and collaborating on needed features (assuming we can solve this bug). As noted in some recent traffic on this topic, we're probably going to spit the analyses off into a separate module so that they can be used with different DB backends, so even if you do end up building your own, you can use the GeneX analyses. Incidentally, what technologies are you using? -- Cheers, Harry Harry J Mangalam -- (949) 856 2847 (v&f) -- hj...@nc... || man...@ho... |
|
From: Harry M. <man...@ho...> - 2001-06-15 16:47:49
|
Jamie Kincaid wrote: > > Harry, > > I'm answering you directly here, but I'd like your permission to broadcast this to our GeneX discussion group as well. > > Thats fine thanks - I sent it off. > > So if I understand you, you want to interface the GeneX tools to a maxdSQL backend? > > Yes - the geneX software is one of the options i'm looking at in order to > provide users with tools to query microarray data. > > Maybe I could be of some help in development, I'm fluent in perl mostly > through CGI scripts but I have coded in 'object perl' as well. > I have no experience with XML. > Anyway - lets keep in touch. Certainly. I'll have a detailed look at the installation script adn see if it can be modularized a bit more.. ie: /usr/bin/perl install-all.pl --just_analysis_tools and your DB scripts would have to be slightly modified to pass the location of the emitted files as a HIDDEN tag, but that's about it I think. > All the best > > Jamie > > FYI - the backend I'm using is MySQL this shouldn't matter very much as far as the analysis tools go. It's all file-based from that point on. -- Cheers, Harry Harry J Mangalam -- (949) 856 2847 (v&f) -- hj...@nc... || man...@ho... |
|
From: Harry M. <man...@ho...> - 2001-06-15 16:41:04
|
with his permission... hjm -------- Original Message -------- Subject: Re: GeneX analysis tools Date: Thu, 14 Jun 2001 08:57:50 -0700 From: Harry Mangalam <man...@ho...> To: Jamie Kincaid <Jam...@no...> References: <Pine.SOL.3.96.1010614161830.1446B-100000@thale> Hi Jamie, I'm answering you directly here, but I'd like your permission to broadcast this to our GeneX discussion group as well. Jamie Kincaid wrote: > > Hi Harry, > I was wondering if i could ask a couple of questions regarding the > analysis tools provided by geneX. > Firstly an introduction! - I'm working on providing an arabidopsis > microarray database using the maxdSQL software. I will be the sole curator > of the database ie inputing the data (as we are running a microarray > service for the community) and the goal is to allow the public to > view / retrieve / analyse the results on line > > Anyway, I notice that all the geneX analytical tools use a web interface > and was wondering if geneX can be run successfully over the internet or > whether the intention is to run it as a local installation. Has it ever > been tried? It has certainly been tried over the web - I'm in california, and help to admin GeneX sites at UCI (just down the hill), and in New Mexico. Most other installations are also 'internet' installations in that their users are distributed around the server by 1-20 router hops (the UCI site has regular users in Australia). It CAN be a local installation - both I and other developers carry it around with us on our laptops. > My intention is to provide on-line analysis tools much in the way that SMD > has provided, however the SMD software requires an Oracle back-end and so > I can't directly use it. I've also played with maxdView but its not > web-enabled. Yes, the Oracle requirement for SMD and other associated admin makes it unsuitable for small installations (tho it's a pretty neat piece of work). > I don't want to re-invent the wheel and develop one-off software to allow > users to interact with our database so I'm trying to guage the possibility > of using the software you guys have developed. I know you've both wrapped > existing software and created your own and therefore the range of tools > would be great! > > Is it possible do you think ? - even with a maxdSQL backend ?? So if I understand you, you want to interface the GeneX tools to a maxdSQL backend? If so, there's no reason why it couldn't be done and in fact I was just thinking over the last 2 days that a better way to modularize the server and make it more useful would be to separate the database part from the analysis part. They already are almost entirely separate (the analysis tools work fine with both the version 1 database and the version 2 (in development) database as the interface hasn't changed. the only thing really keeping them together is the installation script. Changing that is just a bit of grunt work. The interface to the tools is exceptionally stupid/simple. You ask the database for some data, it dumps it to disk and hands off the location of the file to the analysis tool. There's the ineveitable file i/o overhead for doing so, but it makes for an absolutely clean separation between them and generally the analysis time is so much greater than the i/o time that it hardly makes a difference. I've already started work on describing the API and I'd like to provide a skeleton for actually adding analyses as well. Your situation would make a good test case for this approach and would certainly encourage more people to contribute analyses if those analyses could be shared more easily with other database backends. thinking out loud here, formally, what it would need to be really slick is an XML-like description of the analysis, the parameters and format it needs from the DB and a script should be able to parse it to generate an HTML stanza that shows all the available analyses and the parameters they need. > I've spoken to Dave (Hancock) beforehand to give him feedback on maxdSQL > but have yet to ask his opinion on the subject. David already has a pretty good working relationship with NCGR (altho I stood him up at the last MGED - pass on my apologies ;) and we certainly would like to have an even better interaction between maxd and the GeneX suite. one of the ISYS tools uses some Maxd analyses; the least we could do is make ours available to maxd. Only problem thing is time constraints - I might not be able to do the whole thing. > Thanks for your time, > > Jamie > > _______________________________________________________________________________ > > Jamie Kincaid > Developer > http://nasc.nott.ac.uk/ > Nottingham Arabidopsis Stock Centre (NASC) > Plant Science Division > School of Biosciences > University of Nottingham, Nottingham, NG7 2RD > > Phone: 0115 951 3091 Fax: 0115 951 3297 -- Cheers, Harry Harry J Mangalam -- (949) 856 2847 (v&f) -- hj...@nc... || man...@ho... |
|
From: Harry M. <man...@ho...> - 2001-06-15 16:30:20
|
Hi H, I apologize - I'm an idiot - I cut the new release yesterday, installed it, and then forgot to make it readable / downloadable. It can now be downloaded from: http://genex.ncgr.org/genex/download/genex-server/GeneX-Server-1.0.4.tar.gz harry H Stier wrote: > > Below is the result of your feedback form. It was submitted by: > H Stier (dh...@ho...) > on Friday, June 15, 2001 at 02:12:08 > --------------------------------------------------------------------------- > > subject: GeneX Problem Report > > realname: H Stier > > email: dh...@ho... > > Organization: University > > Location: Frankfurt > > Country: Germany > > Operatingsys: Windows 98 > > Problem: I want to download GeneX but when I click the "download" button I always get to the same page and cannot download. > > > > > > > --------------------------------------------------------------------------- -- Cheers, Harry Harry J Mangalam -- (949) 856 2847 (v&f) -- hj...@nc... || man...@ho... |
|
From: <ja...@op...> - 2001-06-15 14:40:45
|
"Ron Ophir" <ls...@wi...> writes: > I am one step further: I have a JAR file at the server side which > includes the XML file and the rest. Good, glad to hear it. What was the problem? User Error, or a problem in the GeneX system? > I have stopped where the documentation stops, but no other script > extracts the data and inserts them into GeneX. In other words what > is next after save-expset.pl? Harry already mentioned that it > relate to xml2db.pl but what is the proper way to runit. Is it > manually or save-expset.pl sould cal it. Right. xml2db.pl is the script you want to use. We didn't want to automate this for every user, so we left it manual. The idea being the DB curator get's an email, and it is up to the curator to run xml2db and handle insertion into the DB. This concept was OK, as long as you were running a National Resource DB (like ArrayExpress, or GEO), but since most people want to run a lab DB, they want automated insertion of data. Then new data loader will do that for you, the old Curation Tool => xml2db combination is not set up that way. It should be trivial to add it to save-experiment, though. I'm running some test files with xml2db right now, and would be happy to provide some pointers if you need them. HTH, jas. PS. Please do reply to the list and not directly to me, this is very useful information to all GeneX users, Thanks! |
|
From: <ja...@op...> - 2001-06-15 02:39:38
|
"Ron Ophir" <Ls...@wi...> writes: > I am trying to upload data using curation tool, however I had no > success. Although at the client all my data exist in xml file there > is some kbs in the jar file at the server side I get jar file with > size zero. I enclose here the control-file: [snip] > My client and genex server are at two sides of firewall. Do you that > is matter? Sorry, I can't say. This seems pretty strange. Are you using an SSL server? jas. |
|
From: <ja...@op...> - 2001-06-15 02:37:25
|
"Harry Mangalam" <man...@ho...> writes:
> Your point is well taken, and in fact we were at one time going to do the
> whole thing as an 'application server'. That is, spawn the analytical
> process on another machine with a custom perl interface for doing so (the
> remains of that code can still be detected in older version fo the
> server), but it started to get to be more trouble than it was worth.
This was something like our original design:
------------
| WWW Server |
------------
|
|
------------ TCP/IP --------------------- TCP/IP -----------------
| DB server |--------| Application server |---------| Analysis Server |
------------ --------------------- -----------------
The idea being the WWW server would just handle HTTP requests, and
return data to the user. The 'Application Server' would farm out any
jobs to the DB server or the Analysis server, compile the data and
send it back to the WWW server. That way we could easily implement
queuing on the App server if we wanted. Also, you could make the
Analysis server be a big beefy SMP box, likewise the DB server. You
could also have the WWW on one side of a firewall and the others
behind the firewall.
jas.
|
|
From: Harry M. <man...@ho...> - 2001-06-14 22:54:51
|
Hi Ron, Your point is well taken, and in fact we were at one time going to do the= whole thing as an 'application server'. That is, spawn the analytical = process on another machine with a custom perl interface for doing so (the= remains of that code can still be detected in older version fo the server), but it started to get to be more trouble than it was worth. Michael's point is good - PostgreSQL (or any decent modern RDBMS) can communicate via sockets so it can be definition run on another machine and communicate via tcp/ip. I think we do support that by allowing you to choose either tcp/ip or unix domain sockets for communicating with the= DB. you can also use other software such as LSF (commercial) or GNU queue (free) to spawn jobs on other machines. I'm planning to play with gnu queue as I just got another PC I can use as a compute client. It DOES require the use of NFS (as you've shown below) to communicate job scheduling and updates back and forth between master and slaves. NFS does complicate things a bit, both in setup and security, so we haven't implemnted this yet. Best wishes Harry Ron Ophir wrote: # Genex integrates database and microarray analysis tools. Whereas # database transaction are relatively fast, microarray analysis are # cpu and memory consuming. Therefore I thought of implement them on # to different servers. The scheme I see in mind is the following: # = # |-----------------| |------------------| # | | | | # | | | | # | HTTPd | | HTTPd | # | | | | # |-----------------| |------------------| # | | # | | # =AF---------------- NFS ---------------------- # | DB server |--------------------| Application server | # =AF---------------- ---------------------- # = # all I have to do then is to redirect the links in the html documents. # Moreover, I may install-all twice on each server and miror the # database in that way each server can back up the other by restoring # the html documents for single mode (server) using. What do you think # wil it work? # = # In addition, that may be another solution for R on MacOSx, if use # MacOSX as a databse server and PC as application server. # = # Ron # = # = -- = Cheers, Harry Harry J Mangalam -- (949) 856 2847 (v&f) -- hj...@nc... || mangalam@home= =2Ecom |
|
From: Michael P. <mic...@ho...> - 2001-06-14 18:35:59
|
Ron,
You shouldn't need to duplicate the GeneX "application" server. Instead,
just run postgresql on a second machine, and load the genex database on
that. There is a parameter "GENEX_HOST" that gets set during the =
installation
process to specify the hostname of the system where the database is =
located.
Unless there are problems at the postgres level, then there shouldn't be =
any problem
with mixing platforms for the database server and genex server.
There are a couple of caveats with this:
1) I don't know if this configuration has been tested through the =
install-all.pl script
I notice at least one place where construction of a connect string =
involving a
remote host is commented. You may need to manually load the database =
dump on
the database server.
2) you will need to make sure that postgresql allows network =
connections. That may
require modifying a file "pg_hba.conf" on the database server. You can =
test for
a working connection from your genex host with "pgsql -h <dbhostname> =
genex genex"
I've used the remote connections from other machines to the genex db in =
postgresql,
but I've not set this up for our genex server.
Regarding the mirroring, that may be more difficult and would need to be =
resolved purely
at the postgresql level.
Regards,
Michael Pear
----- Original Message -----=20
From: Ron Ophir=20
To: gen...@li...=20
Sent: Thursday, June 14, 2001 4:49 AM
Subject: [GeneX-dev] What do you think of that?
Hi All,
Genex integrates database and microarray analysis tools. Whereas =
database transaction are relatively fast, microarray analysis are cpu =
and memory consuming. Therefore I thought of implement them on to =
different servers. The scheme I see in mind is the following:
|-----------------| =
|------------------|
| | =
| |
| | =
| |
| HTTPd | =
| HTTPd |
| | =
| |
|-----------------| =
|------------------|
| =
|
| =
|
=AF---------------- NFS =
-------------------------
| DB server |----------------------------------------| =
Application server |
=AF---------------- =
=AF-----------------------
all I have to do then is to redirect the links in the html documents. =
Moreover, I may install-all twice on each server and miror the database =
in that way each server can back up the other by restoring the html =
documents for single mode (server) using.
What do you think wil it work?
In addition, that may be another solution for R on MacOSx, if use =
MacOSX as a databse server and PC as application server.
Ron
|
|
From: Harry M. <man...@ho...> - 2001-06-14 17:52:17
|
Hi there, (I'm broadcasting this to our dev mailing list as it may be of broader interest.) You can get both programs with the genex-server package which can be gotten at: http://genebox.ncgr.org/genex/download/genex-server/GeneX-Server-1.0.3.tar.gz You can install the whole GeneX system in which case you'll get the database system as well as the analysis tools or if you JUST want xcluster, you can go into the extracted directory: GeneX-Server-1.0.3 xcluster/ |-- xcluster.form.pl.in |-- xcluster.help.html |-- xcluster.notes |-- xcluster.test.dataset `-- xcluster_handler.pl.in and edit & install them yourself. the files ending in '.in' are input files that contain variables (%%REPLACE_ME%%) that need to be replaced with variables that define various things about your system. They should be pretty easy to figure out: xcluster.form.pl.in: %%START_PERL%% %%HTML_ROOT_URL%% # bare HTML root (often 'http://your.machine.com') %%XCLUSTER_DIR%% # combines the genex dir w/ the xcluster dir (often 'genex/xcluster' # but if installing only xcluster, it could be 'xcluster') %%CGI_ROOT_URL%% # bare CGI-BIN root as above (often # 'http://your.machine.com/cgi-bin') xcluster_handler.pl.in: %%START_PERL%% [as above] %%WWWHOST%% # the Fully Qualified Domain Name of the host machine %%CGITMPDIR%% # where you want it to stash temp files (often '/tmp/genex') %%GENEX_HTM_URL%% # where you would install genex (often # 'http://your.machine.com/genex') %%XCLUSTER_DIR%% # [as above] %%HTTP_ERR_LOG%% # where your httpd error log is %%GNUTAIL%% # the full path to your GNU 'tail' %%GNUTAR%% # the full path to your GNU 'tar' %%XCLUSTER%% # the full path to the xcluster binary remove the '.in' suffix from the files, move them to the places you've indicated, set the permissions appropriately, and try them out. We are considering making all the analyses a separate module so you can install them without the database as well as with it. Smita Mitra wrote: > > Dear Harry, > > We have obtained a lisence from Stanford for using > xcluster. We would like to use your > xcluster_handler.pl and xcluster_form.pl for data > analysis. > > Would you be able to provide me with these 2 programs > and some notes on the directory structure you have in > place with xcluster and the programs xcluster calls. > This would make the installation of xcluster and its > integration with a database much easier for us. > > Thanks a lot! > Regards, Smita. > > __________________________________________________ > Do You Yahoo!? > Spot the hottest trends in music, movies, and more. > http://buzz.yahoo.com/ -- Cheers, Harry Harry J Mangalam -- (949) 856 2847 (v&f) -- hj...@nc... || man...@ho... |
|
From: Ron O. <Ls...@wi...> - 2001-06-14 11:49:34
|
Hi All,
Genex integrates database and microarray analysis tools. Whereas database =
transaction are relatively fast, microarray analysis are cpu and memory =
consuming. Therefore I thought of implement them on to different servers. =
The scheme I see in mind is the following:
|-----------------| =
|------------------|
| | =
| |
| | =
| |
| HTTPd | =
| HTTPd |
| | =
| |
|-----------------| =
|------------------|
| =
|
| =
|
=AF---------------- NFS =
-------------------------
| DB server |----------------------------------------| Application =
server |
=AF---------------- =
=AF-----------------------
all I have to do then is to redirect the links in the html documents. =
Moreover, I may install-all twice on each server and miror the database in =
that way each server can back up the other by restoring the html documents =
for single mode (server) using.
What do you think wil it work?
In addition, that may be another solution for R on MacOSx, if use MacOSX =
as a databse server and PC as application server.
Ron
|
|
From: Harry M. <man...@ho...> - 2001-06-12 23:24:39
|
Mine (on harwin and hagrid) and actually looks the same (GeneX Database Information is blank) but it is filled out on genex and genomics..Huh!?? - it doesn't look like a OSX issue per se... I'll check into it to see what's the diffs are hjm "Jason E. Stewart" wrote: > > "Harry Mangalam" <man...@ho...> writes: > > > So add another OS to the compatibility list (well, just barely..) > > We'll have to get such a beast in-house to run the installer before > > we can try this. > > > > Win2k next. and then Atari.. > > OSX, huh. Not bad. > > I had a poke at the system, and I was surprised to see the status page > not working properly: > > http://sabath.labmed.washington.edu/cgi-bin/genex/samples/status.pl > > the 'GeneX Database Information' table isn't being filled out > properly. Harry, does your status page on Harwin or Genebox look OK? > Or did something in install-all.pl get broken? > > > The only way curretnly to get data into the system is with the > > java-based Curation Tool which can work, but which is > > ... difficult... to use. Nevertheless it does (painfully) work: > > > > http://genex.ncgr.org/genex/download/curation_tool/ > > > > but I'd suggest waiting for large scale data loading until we get > > its Perl-based successor the DataLoader up and running - it's late > > but it should be available in alpha/beta in the next coupla weeks. > > We're also loading some more data from the NCI/Ross data set to let > > people have a bit more data to play with and that should be ready by > > weeks end. > > Well, yes. It is an issue. The data-loader will only work on the > GeneX-1.1 development DB. So when it comes out, there may be some > distinctly rough edges to deal with, but it will be a great deal more > fun to work with than the curation tool currently is. > > > Daniel E Sabath wrote: > > > > > > Here's the answer: R was fine! However, install-all.pl didn't like > > > my versions of mkdir and install. It kept giving me the "what > > > gibberish" error and asking me to tell it where to look for mkdir > > > and install (i.e., /bin/mkdir), but I was missing an important > > > message that kept going by--install-all.pl thought these didn't > > > work. > > Yes, this is important. Harry, this seems like another example of why > we *need* to classify the errors into: > > * fatal_error: A critical component is missing, can't proceed with > installation. > * error: An important piece is missing, but you can proceed at your > own risk > * warning: an optional piece is missing, installation will continue. > > jas. > > _______________________________________________ > Genex-dev mailing list > Gen...@li... > http://lists.sourceforge.net/lists/listinfo/genex-dev -- Cheers, Harry Harry J Mangalam -- (949) 856 2847 (v&f) -- hj...@nc... || man...@ho... |
|
From: <ja...@op...> - 2001-06-12 22:56:50
|
"Harry Mangalam" <man...@ho...> writes: > So add another OS to the compatibility list (well, just barely..) > We'll have to get such a beast in-house to run the installer before > we can try this. > > Win2k next. and then Atari.. OSX, huh. Not bad. I had a poke at the system, and I was surprised to see the status page not working properly: http://sabath.labmed.washington.edu/cgi-bin/genex/samples/status.pl the 'GeneX Database Information' table isn't being filled out properly. Harry, does your status page on Harwin or Genebox look OK? Or did something in install-all.pl get broken? > The only way curretnly to get data into the system is with the > java-based Curation Tool which can work, but which is > ... difficult... to use. Nevertheless it does (painfully) work: > > http://genex.ncgr.org/genex/download/curation_tool/ > > but I'd suggest waiting for large scale data loading until we get > its Perl-based successor the DataLoader up and running - it's late > but it should be available in alpha/beta in the next coupla weeks. > We're also loading some more data from the NCI/Ross data set to let > people have a bit more data to play with and that should be ready by > weeks end. Well, yes. It is an issue. The data-loader will only work on the GeneX-1.1 development DB. So when it comes out, there may be some distinctly rough edges to deal with, but it will be a great deal more fun to work with than the curation tool currently is. > Daniel E Sabath wrote: > > > > Here's the answer: R was fine! However, install-all.pl didn't like > > my versions of mkdir and install. It kept giving me the "what > > gibberish" error and asking me to tell it where to look for mkdir > > and install (i.e., /bin/mkdir), but I was missing an important > > message that kept going by--install-all.pl thought these didn't > > work. Yes, this is important. Harry, this seems like another example of why we *need* to classify the errors into: * fatal_error: A critical component is missing, can't proceed with installation. * error: An important piece is missing, but you can proceed at your own risk * warning: an optional piece is missing, installation will continue. jas. |
|
From: Harry M. <man...@ho...> - 2001-06-12 05:23:39
|
Yeehah! That's wonderful (but sorry I wasn't able to grok it better in the beginning - I'll add your comments to the INSTALL file. I cruised your database and ran thru it end to end with the DB to CyberT adn it does work, including the xgobi routines! Congratulations! So add another OS to the compatibility list (well, just barely..) We'll have to get such a beast in-house to run the installer before we can try this. Win2k next. and then Atari.. I wonder why Apple decided not to fit OSX with the std Gnu file utils..? Doesn't it also come with gcc and other Gnu utils? Or is this some kind of Mach-required weirdness? The only way curretnly to get data into the system is with the java-based Curation Tool which can work, but which is ... difficult... to use. Nevertheless it does (painfully) work: http://genex.ncgr.org/genex/download/curation_tool/ but I'd suggest waiting for large scale data loading until we get its Perl-based successor the DataLoader up and running - it's late but it should be available in alpha/beta in the next coupla weeks. We're also loading some more data from the NCI/Ross data set to let people have a bit more data to play with and that should be ready by weeks end. Daniel E Sabath wrote: > > > Here's the answer: R was fine! However, install-all.pl didn't like > my versions of mkdir and install. It kept giving me the "what > gibberish" error and asking me to tell it where to look for mkdir > and install (i.e., /bin/mkdir), but I was missing an important > message that kept going by--install-all.pl thought these didn't > work. > > I downloaded, compiled, and installed the fileutils source code > form gnu.org, and the GeneX install-all.pl script was finally happy > and finished without errors. Check it out: > <http://sabath.labmed.washington.edu/genex/index.shtml>! > > Now I need to figure out how to get our upcoming array data into > our database. Where would you recommend looking for the > documentation on how to do this? > > Thanks for all your help. Sorry for sending you on wild goose > chases. If anyone else is having trouble with Mac OS X, let them > know it can be done (I assume I wasn't the first, yes?)! > > Dan > -- > Daniel E. Sabath, M.D., Ph.D. > University of Washington Department of Laboratory Medicine > Box 357110 > Seattle, WA 98195-7110 USA > Phone: (206) 598-6833 Fax: (206) 598-6189 > http://www.labmed.washington.edu -- Cheers, Harry Harry J Mangalam -- (949) 856 2847 (v&f) -- hj...@nc... || man...@ho... |
|
From: Ron O. <Ls...@wi...> - 2001-06-07 13:17:43
|
Hi everybody, I am trying to upload data using curation tool, however I had no success. = Although at the client all my data exist in xml file there is some kbs in = the jar file at the server side I get jar file with size zero. I enclose = here the control-file: !XY=3DSTART-OF-SUBMISSION********** !XY=3DSTART-TIME|991915393 !XY=3DSUBMIT-ID|HopefulyMyFirst-698522918.1482 !XY=3DSUBMIT-DIR|/usr/local/genex/submission/HopefulyMyFirst-698522918.1482= !XY=3DEXPSET-NAME|HopefulyMyFirst !XY=3DEXPSET-ID|HopefulyMyFirst-698522918.1482 !XY=3DEXPSET-FILE|E:\Downloads\Curation Tools\ct101_usrwin\data\expsets\sub= mitted\HopefulyMyFirst\Hope fulyMyFirst.jar !XY=3DACTUAL-FILE| !XY=3DUSER-COMMENTS| !XY=3DUSER-EMAIL|ron...@we... !XY=3DUSER-ID|67 !XY=3DUSER-NAME|genex My client and genex server are at two sides of firewall. Do you that is = matter? thanks in advance. Ron=20 |
|
From: Harry M. <man...@ho...> - 2001-06-05 21:54:59
|
Hi Dan, &^%^##$@^*^)(*^$&^&@^ Well, I'm starting to run out of explanations except for the one that's the least helpful - that you appear to have a damaged R binary and it may just not work. Brian Ripley did mention that this was a bug in the current version that will be corrected in the next version. You mentioned that the reason that you couldn't compile R from source was that you couldn't find a fortran compiler. Did you try building either g77 or f2c? Otherwise we're at a dead end until one of us gets Mac OSX running. On the other hand, it might be that even with these you might end up with an R binary that wouldn't work. You can install xgobi by hand, tho I'm not sure how to do it offhand. Mybe the thing to do is to wait until the next release of R for Mac OSX comes out and try it then..? GeneX will have moved up a bit in stability as well. Sorry this isn't more helpful, but without access to MacOSX, you're somewhat on your own.. harry Daniel E Sabath wrote: > > On Tuesday, June 5, 2001, at 12:07 PM, Harry Mangalam wrote: > > > R --silent INSTALL --library=/usr/local/genex/lib/R/library \ > > /Users/dan/sw/src/GeneX-Server-1.0.3/CyberT-dist/xgobi_1.2-2.tar.gz > > > > if that's where the files still are. At lest that way, we'll know > > if R can install the thing correctly. > > Harry, > > The above command hung in the exact same way as the full install. > FWIW, the xgobi tarball is in the correct directory. > > > If it fails to do this, manually unpack the xgobi tarball and then > > try it again with the > > command: > > After manually unpacking the xgobi tarball I got a directory called > xgobi, which overwrote a file previously called xgobi. I renamed > the new directory xgobi_1.2-2 and re-ran install-all.pl. This > created a new xgobi file. Then I did: > > > R --silent INSTALL --library=/usr/local/genex/lib/R/library \ > > /Users/dan/sw/src/GeneX-Server-1.0.3/CyberT-dist/xgobi > > which returned: > > gzip: /Users/dan/sw/src/GeneX-Server-1.0.3/CyberT-dist/xgobi: not > in gzip format > ERROR: cannot extract package from /Users/dan/sw/src/GeneX- > Server-1.0.3/CyberT-dist/xgobi > > It appears it's looking for a tarball and not finding it. > > Any other suggestions? > > Dan > > -- > Daniel E. Sabath, M.D., Ph.D. > University of Washington Department of Laboratory Medicine > Box 357110 > Seattle, WA 98195-7110 USA > Phone: (206) 598-6833 Fax: (206) 598-6189 > http://www.labmed.washington.edu -- Cheers, Harry Harry J Mangalam -- (949) 856 2847 (v&f) -- hj...@nc... || man...@ho... |
|
From: Harry M. <man...@ho...> - 2001-06-05 20:08:50
|
Hi Dan, Can you try installing the xgobi lib manually? By that I mean go to the top level installation dir (where install-all.pl is) and type: R --silent INSTALL --library=/usr/local/genex/lib/R/library \ /Users/dan/sw/src/GeneX-Server-1.0.3/CyberT-dist/xgobi_1.2-2.tar.gz if that's where the files still are. At lest that way, we'll know if R can install the thing correctly. If it fails to do this, manually unpack the xgobi tarball and then try it again with the command: R --silent INSTALL --library=/usr/local/genex/lib/R/library \ /Users/dan/sw/src/GeneX-Server-1.0.3/CyberT-dist/xgobi (you might have to rename the newly created directory xgobi_1.2-2 so as not to conflict with the xgobi library that is also created. This has never been a problem area for other platforms tho..) hjm Daniel E Sabath wrote: > > On Tuesday, June 5, 2001, at 11:15 AM, Harry Mangalam wrote: > > > Hmmm... sounds like bug to me - but it doesn't fail on x86 with > > the same R version - IS your tar working oK? I suspect it is or > > you wouldn't be able to get into OUR installer. Nevertheless, > > unpack > > it in place and try running the installer again.. > > > > If it STILL doesn't work, lemme know. > > Harry, > > I got my precompiled R from <http://www.stat.ucla.edu/~deleeuw/R>. > There was a newer version there since the last time I looked, but I > still get the same error. Perhaps that version of R is missing > something I need. > > Dan > -- > Daniel E. Sabath, M.D., Ph.D. > University of Washington Department of Laboratory Medicine > Box 357110 > Seattle, WA 98195-7110 USA > Phone: (206) 598-6833 Fax: (206) 598-6189 > http://www.labmed.washington.edu -- Cheers, Harry Harry J Mangalam -- (949) 856 2847 (v&f) -- hj...@nc... || man...@ho... |
|
From: Prof B. D R. <ri...@st...> - 2001-06-05 18:55:43
|
On Tue, 5 Jun 2001, Harry Mangalam wrote: > Hi All, > > We're having some problems with the Mac OSX port of R V1.2.3 binary that came from CRAN in that a (Perl) scripted installation of the xgobi library halts unexpectedly. The commandline that fails is: > > R --silent INSTALL --library=/usr/local/genex/lib/R/library \ > /Users/dan/sw/src/GeneX-Server-1.0.3/CyberT-dist/xgobi_1.2-2.tar.gz > > The comment from the affected user is: > > It still hangs just after "Installing source package > 'xgobi'", then " R" appears on the next line (including the space). *That* happens if there are no Rq... files in the R directory. Is tar working correctly? I would try untaring the package, checking that the R/* files are there and then installing the unpacked version. BTW, the hanging won't happen in the next release. > > I saw the thread on " Difficulty with R CMD INSTALL " and was wondering if others have had a similar problem on OSX. > > Using R 1.2.3 on x86 linux (RH & Debian), I've never seen this problem, so I'm wondering if it's a porting artefact. > > -- > Cheers, > Harry > > Harry J Mangalam -- (949) 856 2847 (v&f) -- hj...@nc... || man...@ho... > -.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.- > r-help mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html > Send "info", "help", or "[un]subscribe" > (in the "body", not the subject !) To: r-h...@st... > _._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._ > -- Brian D. Ripley, ri...@st... Professor of Applied Statistics, http://www.stats.ox.ac.uk/~ripley/ University of Oxford, Tel: +44 1865 272861 (self) 1 South Parks Road, +44 1865 272860 (secr) Oxford OX1 3TG, UK Fax: +44 1865 272595 |
|
From: Harry M. <man...@ho...> - 2001-06-05 18:08:33
|
Hi All, We're having some problems with the Mac OSX port of R V1.2.3 binary that came from CRAN in that a (Perl) scripted installation of the xgobi library halts unexpectedly. The commandline that fails is: R --silent INSTALL --library=/usr/local/genex/lib/R/library \ /Users/dan/sw/src/GeneX-Server-1.0.3/CyberT-dist/xgobi_1.2-2.tar.gz The comment from the affected user is: It still hangs just after "Installing source package 'xgobi'", then " R" appears on the next line (including the space). I saw the thread on " Difficulty with R CMD INSTALL " and was wondering if others have had a similar problem on OSX. Using R 1.2.3 on x86 linux (RH & Debian), I've never seen this problem, so I'm wondering if it's a porting artefact. -- Cheers, Harry Harry J Mangalam -- (949) 856 2847 (v&f) -- hj...@nc... || man...@ho... |
|
From: Harry M. <man...@ho...> - 2001-06-05 16:44:37
|
Well, the easy answer is that yes, it /could/ be the case - I don't know how your sudoers file is set up, but if it doesn't allow you to use all the commands that the install-all.pl script uses, it would probably fail. Try the install by 'su - root' if you can. hjm "Daniel E.Sabath" wrote: > > On Monday, June 4, 2001, at 04:05 PM, Harry Mangalam wrote: > > > I've been trying to make the installation fail like yours did and the > > closest I can come is if you execute the script OUTSIDE of the > > installation directory. ie from the directory ABOVE the > > GeneX-Server-1.0.3 or from somewhere else by just giving the path: > > > > /usr/bin/perl path/to/Genex-Server-1.0.3/install-all.pl > > > > Is that a possibility? As you might guess, you do need to execute it > > from the same dir as the install-all.pl script. (and we should add a > > check to make sure).. > > Harry, > > Yes, I was running it from the GeneX directory. Specifically, the > command line I'm using is: > > % sudo /usr/bin/perl install-all.pl > > It just occurred to me that I'm using sudo rather than su'ing to root. > Could this be the problem? I may answer my own question before I hear > back from you. I'll let you know what happens. > > Thanks, > > Dan > -- > Daniel E. Sabath, M.D., Ph.D. > Department of Laboratory Medicine, University of Washington > Box 357110 > Seattle, WA 98195-7110 > Office: 206-598-6833 Lab: 206-543-9032 Fax: 206-598-6189 > http://www.labmed.washington.edu -- Cheers, Harry Harry J Mangalam -- (949) 856 2847 (v&f) -- hj...@nc... || man...@ho... |
|
From: Harry M. <man...@ho...> - 2001-06-04 21:00:37
|
What IS it with you Mac OSX people. We've lately had more OSX questions than all other OSs combined... :) The coming platform apparently.. Daniel E Sabath wrote: > > I'm having a bit of trouble installing GeneX 1.0.3 on a MacOS > X/Darwin system. I successfully (I think) installed everything as > described in the INSTALL document, but the install-all.pl script > hung when xgobi was supposed to be getting updated. Do you mean that the xgobi/xgvis paths were getting set or the xgobi R lib was getting installed? The errors indicate that: /Users/dan/sw/src/GeneX-Server-1.0.3/CyberT-dist/xgobi_1.2-2.tar.gz wasn't in place or wasn't readable. and the succeeding errors indicate a similar problem. Did you allow the installation script to create the /usr/local/genex directory? (altho I'd expect a lot more erros if that was the case...), but you don't seem to have installed the Genex.pm bits properly either as I can't make a query to the DB either. That goes along with the line 2498 error. Could you have accidentally changed the privs on some of the installation directories? Also, were you able to verify that the DB was installed correctly before you went further (using the init-dump.pl script)? > I've attached > my Option.reminders and genex-errors.txt files for your perusal. Thanks for doing this - it makes a diff! I tried to connect to your server to see if I could figure out anything wrong by inspection and you also need to allow server side includes as well. Obviously, most of the install went OK as I can see much of the installation laid out as it should be, but it isn't functional yet either. > Thanks, > > Dan > > -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- > Name: genex-errors.txt > genex-errors.txt Type: Plain Text (text/plain) > Encoding: quoted-printable > > Name: Option.Reminders > Option.Reminders Type: unspecified type (application/octet-stream) > Encoding: quoted-printable > > -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- > > -- > Daniel E. Sabath, M.D., Ph.D. > University of Washington Department of Laboratory Medicine > Box 357110 > Seattle, WA 98195-7110 USA > Phone: (206) 598-6833 Fax: (206) 598-6189 > http://www.labmed.washington.edu -- Cheers, Harry Harry J Mangalam -- (949) 856 2847 (v&f) -- hj...@nc... || man...@ho... |
54 messages has been excluded from this view by a project administrator.
