Re: [NTG-context] calling ConTeXt as a background process (SOLVED)
On 26-10-2010 10:56, Taco Hoekwater wrote: On 10/26/10 22:09, Hans Hagen wrote: On 26-10-2010 9:28, Taco Hoekwater wrote: Hi Henning, On 10/08/10 21:44, Henning Hraban Ramm wrote: Ok, the failures were the same with fcgi and gunicorn app servers. But it _works_ with django's dev server (which you must not use in production due to memory/security holes etc.). We just had a somewhat similar problem (texlua hung at 100% cpu when run via apache) and it turns out, after lots of debugging, that mtxrun cannot do anything without $HOME. Perhaps yours was a different symptom of the same problem ? try this if you want to test CPU heat generation: $ env HOME= context knuth so what should HOME default to then? "."? Not a default is what is needed, what is needed is that it should work without a HOME at all (server processes typically dont have one). sure, but I guess that some path specs have ~ and that we get a loop in the resolver due to this and for these patsh with ~ we need a proper default then Let's sort this out offlist. okay Hans - Hans Hagen | PRAGMA ADE Ridderstraat 27 | 8061 GH Hasselt | The Netherlands tel: 038 477 53 69 | voip: 087 875 68 74 | www.pragma-ade.com | www.pragma-pod.nl - ___ If your question is of interest to others as well, please add an entry to the Wiki! maillist : ntg-context@ntg.nl / http://www.ntg.nl/mailman/listinfo/ntg-context webpage : http://www.pragma-ade.nl / http://tex.aanhet.net archive : http://foundry.supelec.fr/projects/contextrev/ wiki : http://contextgarden.net ___
Re: [NTG-context] calling ConTeXt as a background process (SOLVED)
On 10/26/10 22:09, Hans Hagen wrote: > On 26-10-2010 9:28, Taco Hoekwater wrote: >> >> Hi Henning, >> >> On 10/08/10 21:44, Henning Hraban Ramm wrote: >>> Ok, the failures were the same with fcgi and gunicorn app servers. >>> >>> But it _works_ with django's dev server (which you must not use in >>> production due to memory/security holes etc.). >> >> We just had a somewhat similar problem (texlua hung at 100% cpu >> when run via apache) and it turns out, after lots of debugging, >> that mtxrun cannot do anything without $HOME. Perhaps yours was >> a different symptom of the same problem ? >> >> try this if you want to test CPU heat generation: >> >>$ env HOME= context knuth > > so what should HOME default to then? "."? Not a default is what is needed, what is needed is that it should work without a HOME at all (server processes typically dont have one). Let's sort this out offlist. Best wishes, Taco ___ If your question is of interest to others as well, please add an entry to the Wiki! maillist : ntg-context@ntg.nl / http://www.ntg.nl/mailman/listinfo/ntg-context webpage : http://www.pragma-ade.nl / http://tex.aanhet.net archive : http://foundry.supelec.fr/projects/contextrev/ wiki : http://contextgarden.net ___
Re: [NTG-context] calling ConTeXt as a background process (SOLVED)
Am 2010-10-26 um 21:28 schrieb Taco Hoekwater: We just had a somewhat similar problem (texlua hung at 100% cpu when run via apache) and it turns out, after lots of debugging, that mtxrun cannot do anything without $HOME. Perhaps yours was a different symptom of the same problem ? Ah, I didn't check HOME; I'll add that to my environment, just in case. (But ATM it runs well.) try this if you want to test CPU heat generation: $ env HOME= context knuth (use one process for each available cpu thread, of course) ;-) I'm using a rented virtual server and won't feel the heat. (And I guess the hoster will just throttle or disable it if I try to melt it.) Thank you! Greetlings from Lake Constance! Hraban --- http://www.fiee.net/texnique/ http://wiki.contextgarden.net https://www.cacert.org (I'm an assurer) ___ If your question is of interest to others as well, please add an entry to the Wiki! maillist : ntg-context@ntg.nl / http://www.ntg.nl/mailman/listinfo/ntg-context webpage : http://www.pragma-ade.nl / http://tex.aanhet.net archive : http://foundry.supelec.fr/projects/contextrev/ wiki : http://contextgarden.net ___
Re: [NTG-context] calling ConTeXt as a background process (SOLVED)
On 26-10-2010 9:28, Taco Hoekwater wrote: Hi Henning, On 10/08/10 21:44, Henning Hraban Ramm wrote: Ok, the failures were the same with fcgi and gunicorn app servers. But it _works_ with django's dev server (which you must not use in production due to memory/security holes etc.). We just had a somewhat similar problem (texlua hung at 100% cpu when run via apache) and it turns out, after lots of debugging, that mtxrun cannot do anything without $HOME. Perhaps yours was a different symptom of the same problem ? try this if you want to test CPU heat generation: $ env HOME= context knuth so what should HOME default to then? "."? - Hans Hagen | PRAGMA ADE Ridderstraat 27 | 8061 GH Hasselt | The Netherlands tel: 038 477 53 69 | voip: 087 875 68 74 | www.pragma-ade.com | www.pragma-pod.nl - ___ If your question is of interest to others as well, please add an entry to the Wiki! maillist : ntg-context@ntg.nl / http://www.ntg.nl/mailman/listinfo/ntg-context webpage : http://www.pragma-ade.nl / http://tex.aanhet.net archive : http://foundry.supelec.fr/projects/contextrev/ wiki : http://contextgarden.net ___
Re: [NTG-context] calling ConTeXt as a background process (SOLVED)
Hi Henning, On 10/08/10 21:44, Henning Hraban Ramm wrote: > Ok, the failures were the same with fcgi and gunicorn app servers. > > But it _works_ with django's dev server (which you must not use in > production due to memory/security holes etc.). We just had a somewhat similar problem (texlua hung at 100% cpu when run via apache) and it turns out, after lots of debugging, that mtxrun cannot do anything without $HOME. Perhaps yours was a different symptom of the same problem ? try this if you want to test CPU heat generation: $ env HOME= context knuth (use one process for each available cpu thread, of course) Best wishes, Taco ___ If your question is of interest to others as well, please add an entry to the Wiki! maillist : ntg-context@ntg.nl / http://www.ntg.nl/mailman/listinfo/ntg-context webpage : http://www.pragma-ade.nl / http://tex.aanhet.net archive : http://foundry.supelec.fr/projects/contextrev/ wiki : http://contextgarden.net ___
Re: [NTG-context] calling ConTeXt as a background process (SOLVED)
Am 2010-10-10 um 10:47 schrieb Patrick Gundlach: Thank you all for your patience, ideas and help! Sorry, that I didn't quite understand the problem/solution, could you describe in a few sentences what problem was and how you solved it? I actually couldn't find out what the real problem was. ;-) I tried to start ConTeXt from a Django web application server. (Running ConTeXt from an active shell was no problem.) My Django FCGI processes were running under daemontools (http://cr.yp.to/daemontools.html ) supervision, i.e. started by its supervise daemon. In this setup, texlua hung at least at uname (and maybe other) calls. I found no error messages on that, I guess the environment was restricted in some way. Since I thought it might have to do with the FCGI server, I replaced that by Green Unicorn (http://gunicorn.org/), but that made no difference WRT ConTeXt. When I changed process supervision from daemontools to supervisor (http://supervisord.org/ ), I got no further problems. Greetlings from Lake Constance! Hraban --- http://www.fiee.net/texnique/ http://wiki.contextgarden.net https://www.cacert.org (I'm an assurer) ___ If your question is of interest to others as well, please add an entry to the Wiki! maillist : ntg-context@ntg.nl / http://www.ntg.nl/mailman/listinfo/ntg-context webpage : http://www.pragma-ade.nl / http://tex.aanhet.net archive : http://foundry.supelec.fr/projects/contextrev/ wiki : http://contextgarden.net ___
Re: [NTG-context] calling ConTeXt as a background process (SOLVED)
> > Thank you all for your patience, ideas and help! Sorry, that I didn't quite understand the problem/solution, could you describe in a few sentences what problem was and how you solved it? Thanks Patrick ___ If your question is of interest to others as well, please add an entry to the Wiki! maillist : ntg-context@ntg.nl / http://www.ntg.nl/mailman/listinfo/ntg-context webpage : http://www.pragma-ade.nl / http://tex.aanhet.net archive : http://foundry.supelec.fr/projects/contextrev/ wiki : http://contextgarden.net ___
Re: [NTG-context] calling ConTeXt as a background process (SOLVED)
Ok, the failures were the same with fcgi and gunicorn app servers. But it _works_ with django's dev server (which you must not use in production due to memory/security holes etc.). While ConTeXt runs it looks like: 5852 25767 TS 21 18:04 pts/000:00:00 \_ python manage.py runserver 8000 1477 5852 TS 23 18:07 pts/000:00:03 \_ /var/www/ cerebrale/bin/python manage.py runserver 8000 3395 1477 TS 22 18:08 pts/000:00:00 \_ /bin/sh -c date "+%Y-%m-%d %H:%M:%S" && cd /var/www/.../_textemp && /var/opt/ context/tex/texmf-linux-64/bin/context --purgeall --batchmode -- result="/var/www/... 3397 3395 TS 22 18:08 pts/000:00:00 \_ /bin/ sh /var/opt/context/tex/texmf-linux-64/bin/context --purgeall -- batchmode --result=/var/www/... 3398 3397 TS 15 18:08 pts/000:00:00 \_ texlua /var/opt/context/tex/texmf-linux-64/bin/mtxrun --script context --purgeall --batchmode --result=/var/www/... 3400 3398 TS 21 18:08 pts/000:00:00 \_ luatex --interaction=batchmode --fmt=/var/opt/context/tex/texmf- cache/.../cont-en --lua=/var/opt/context/tex... TeX completes fast, and everything seems well afterwards. But I started that app server from an active shell. Maybe that's the difference? Right, if I start gunicorn manually, it also works. So, the culprit seems to be the automatic background start of the app server via daemontools. I'll have to test other means of starting and supervising the processes - e.g. runit or supervisor. Since I like Python and don't want to replace "init" yet, I try the latter first. Some hours later: YES IT WORKS with supervisor so far! Thank you all for your patience, ideas and help! Greetlings from Lake Constance! Hraban --- http://www.fiee.net/texnique/ http://wiki.contextgarden.net https://www.cacert.org (I'm an assurer) ___ If your question is of interest to others as well, please add an entry to the Wiki! maillist : ntg-context@ntg.nl / http://www.ntg.nl/mailman/listinfo/ntg-context webpage : http://www.pragma-ade.nl / http://tex.aanhet.net archive : http://foundry.supelec.fr/projects/contextrev/ wiki : http://contextgarden.net ___
Re: [NTG-context] calling ConTeXt as a background process
On 8-10-2010 4:57, Taco Hoekwater wrote: Hi, On 10/08/2010 04:30 PM, Henning Hraban Ramm wrote: I just don't understand why any lua script would call uname externally if it's available as os.uname? My guess is that this is historical. Hans? yes, and also related to some issues on osx where the built in uname was not yet ok (needs checking, but not today by me) - Hans Hagen | PRAGMA ADE Ridderstraat 27 | 8061 GH Hasselt | The Netherlands tel: 038 477 53 69 | voip: 087 875 68 74 | www.pragma-ade.com | www.pragma-pod.nl - ___ If your question is of interest to others as well, please add an entry to the Wiki! maillist : ntg-context@ntg.nl / http://www.ntg.nl/mailman/listinfo/ntg-context webpage : http://www.pragma-ade.nl / http://tex.aanhet.net archive : http://foundry.supelec.fr/projects/contextrev/ wiki : http://contextgarden.net ___
Re: [NTG-context] calling ConTeXt as a background process
On Oct 8, 2010, at 16:30 , Henning Hraban Ramm wrote: > Am 2010-10-08 um 15:54 schrieb Taco Hoekwater: > >> On 10/08/2010 03:47 PM, Henning Hraban Ramm wrote: >>> >>> 16002 15978 TS 21 15:41 ? 00:00:01 \_ /var/www/xxx/bin/python >>> /var/www/.../manage.py run_gunicorn -c /var/www/.../gunicorn-settings.py >>> 16210 16002 TS 14 15:42 ? 00:00:08 \_ luatex --interaction=batchmode >>> --fmt=/var/opt/context/tex/texmf-cache/.../formats/cont-en >>> --lua=/var/opt/context/tex/texmf-cache/... >>> 16212 16210 TS 17 15:42 ? 00:00:00 \_ [uname] >> >> Hm, defunct, eh? >> >> The luatex binary itself definitely does not call uname as a >> process, therefore this has to be the os.resultof() function >> that Hans implements in l-os.lua. I am suspecting issues with >> redirection now, as that function is defined as: >> >> function os.resultof(command) >> local handle = io.popen(command,"r") >> return handle and handle:read("*all") or "" >> end The strange thing is that running luatex --interaction=batchmode --fmt=".../formats/cont-en" --lua=".../formats/cont-en.lui" --backend=pdf hello causes luatex (or say l-os.lua) to call 'uname -m' via os.resultof However, when doing mtxrun --script context --batchmode --once hello *only* texlua calls 'uname -m' and the following luatex child does *not* anymore. Why? >> of course this will fail/block rather horribly if uname -m >> does not write to STDOUT (and that is not just uname, thre >> are a few more uses of os.resultof()). if that is the case you should not see any reads after the child forks: clone(child_stack=0, flags=CLONE_CHILD_CLEARTID|CLONE_CHILD_SETTID|SIGCHLD, child_tidptr=0x7feb7a0369d0) = 14998 [...] read(3, "x86_64\n", 4096) = 7 read(3, "", 4096) = 0 --- SIGCHLD (Child exited) @ 0 (0) --- > My script now calls first os.uname and then the external one. > From the shell it works like expected. > And even from the server process! That alone doesn't seem to be the problem. > > Here's the trace: > > strace -ff texlua call_uname.lua > > [...] I'm running out of ideas. From your script os.resultof("uname -m") works and from luatex it doesn't. What does this trace distinguish from the one of luatex? Florian ___ If your question is of interest to others as well, please add an entry to the Wiki! maillist : ntg-context@ntg.nl / http://www.ntg.nl/mailman/listinfo/ntg-context webpage : http://www.pragma-ade.nl / http://tex.aanhet.net archive : http://foundry.supelec.fr/projects/contextrev/ wiki : http://contextgarden.net ___
Re: [NTG-context] calling ConTeXt as a background process
Just installed Ruby, fetched the latest beta and tried texexec: 26382 26367 TS 20 16:38 ?00:00:01 \_ /var/www/xxx/bin/ python /var/www/xxx/manage.py run_gunicorn -c /var/www/.../gunicorn- settings.py 26509 26382 TS 20 16:38 ?00:00:00 \_ /bin/sh -c texexec --batchmode --once hello >> _context.log 26510 26509 TS 20 16:38 ?00:00:00 \_ /bin/sh /var/ opt/context/tex/texmf-linux-64/bin/texexec --batchmode --once hello 26511 26510 TS 14 16:38 ?00:00:21 \_ texlua / var/opt/context/tex/texmf-linux-64/bin/mtxrun --usekpse --execute texexec --batchmode --once hello 26512 26511 TS 14 16:38 ?00:00:00 \_ [uname] Argl, I didn't knew texexec also depends on mtxrun/lua... Ok, patching mtxrun again, now I even replace os.uname, but I get the same as above. Hraban ___ If your question is of interest to others as well, please add an entry to the Wiki! maillist : ntg-context@ntg.nl / http://www.ntg.nl/mailman/listinfo/ntg-context webpage : http://www.pragma-ade.nl / http://tex.aanhet.net archive : http://foundry.supelec.fr/projects/contextrev/ wiki : http://contextgarden.net ___
Re: [NTG-context] calling ConTeXt as a background process
Hi, On 10/08/2010 04:30 PM, Henning Hraban Ramm wrote: I just don't understand why any lua script would call uname externally if it's available as os.uname? My guess is that this is historical. Hans? Best wishes, Taco Sorry, no smart ideas about the rest of the problems any more ___ If your question is of interest to others as well, please add an entry to the Wiki! maillist : ntg-context@ntg.nl / http://www.ntg.nl/mailman/listinfo/ntg-context webpage : http://www.pragma-ade.nl / http://tex.aanhet.net archive : http://foundry.supelec.fr/projects/contextrev/ wiki : http://contextgarden.net ___
Re: [NTG-context] calling ConTeXt as a background process
Am 2010-10-08 um 15:54 schrieb Taco Hoekwater: On 10/08/2010 03:47 PM, Henning Hraban Ramm wrote: 16002 15978 TS 21 15:41 ? 00:00:01 \_ /var/www/xxx/bin/python /var/www/.../manage.py run_gunicorn -c /var/www/.../gunicorn- settings.py 16210 16002 TS 14 15:42 ? 00:00:08 \_ luatex --interaction=batchmode --fmt=/var/opt/context/tex/texmf-cache/.../formats/cont-en --lua=/var/opt/context/tex/texmf-cache/... 16212 16210 TS 17 15:42 ? 00:00:00 \_ [uname] Hm, defunct, eh? The luatex binary itself definitely does not call uname as a process, therefore this has to be the os.resultof() function that Hans implements in l-os.lua. I am suspecting issues with redirection now, as that function is defined as: function os.resultof(command) local handle = io.popen(command,"r") return handle and handle:read("*all") or "" end The version in the garden has: function os.resultof(command) ioflush() -- else messed up logging local handle = io.popen(command,"r") if not handle then -- print("unknown command '".. command .. "' in os.resultof") return "" else return handle:read("*all") or "" end end and it is called like so: os.resultof('uname -m') of course this will fail/block rather horribly if uname -m does not write to STDOUT (and that is not just uname, thre are a few more uses of os.resultof()). I tried to redirect stdout and stderr to my log file, maybe that's a problem then? If I call using os.system, I can't redirect, but the result is the same. I can redirect uname's output like so: uname -m >> context.log So I guess, it's using stdout at least normally - perhaps not via os.resultof, no clue. I just don't understand why any lua script would call uname externally if it's available as os.uname? Or does os.uname call os.resultof('uname')? (Probably not, otherwise Florian's suggestion wouldn't have worked.) print(os.resultof('uname -m')) gives attempt to call field 'resultof' (a nil value) Can I import that function anyhow? I just copied it to my script and replaced ioflush with io.flush. My script now calls first os.uname and then the external one. From the shell it works like expected. And even from the server process! That alone doesn't seem to be the problem. Here's the trace: strace -ff texlua call_uname.lua execve("/var/opt/context/tex/texmf-linux-64/bin/texlua", ["texlua", "call_uname.lua"], [/* 10 vars */]) = 0 brk(0) = 0x6a38000 mmap(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0x2b8eb789f000 uname({sys="Linux", node="aine.fiee.net", ...}) = 0 access("/etc/ld.so.nohwcap", F_OK) = -1 ENOENT (No such file or directory) mmap(NULL, 8192, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0x2b8eb78a access("/etc/ld.so.preload", R_OK) = -1 ENOENT (No such file or directory) open("/etc/ld.so.cache", O_RDONLY) = 4 fstat(4, {st_mode=S_IFREG|0644, st_size=20573, ...}) = 0 mmap(NULL, 20573, PROT_READ, MAP_PRIVATE, 4, 0) = 0x2b8eb78a2000 close(4)= 0 access("/etc/ld.so.nohwcap", F_OK) = -1 ENOENT (No such file or directory) open("/lib/libdl.so.2", O_RDONLY) = 4 read(4, "\177ELF\2\1\1\0\0\0\0\0\0\0\0\0\3\0> \0\1\0\0\0\20\16\0\0\0\0\0\0@"..., 832) = 832 fstat(4, {st_mode=S_IFREG|0644, st_size=14616, ...}) = 0 mmap(NULL, 2109728, PROT_READ|PROT_EXEC, MAP_PRIVATE|MAP_DENYWRITE, 4, 0) = 0x2b8eb7aa mprotect(0x2b8eb7aa2000, 2097152, PROT_NONE) = 0 mmap(0x2b8eb7ca2000, 8192, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_FIXED| MAP_DENYWRITE, 4, 0x2000) = 0x2b8eb7ca2000 close(4)= 0 access("/etc/ld.so.nohwcap", F_OK) = -1 ENOENT (No such file or directory) open("/lib/libm.so.6", O_RDONLY)= 4 read(4, "\177ELF\2\1\1\0\0\0\0\0\0\0\0\0\3\0>\0\1\0\0\0P> \0\0\0\0\0\0@"..., 832) = 832 fstat(4, {st_mode=S_IFREG|0644, st_size=534736, ...}) = 0 mmap(NULL, 2629848, PROT_READ|PROT_EXEC, MAP_PRIVATE|MAP_DENYWRITE, 4, 0) = 0x2b8eb7ca4000 mprotect(0x2b8eb7d26000, 2093056, PROT_NONE) = 0 mmap(0x2b8eb7f25000, 8192, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_FIXED| MAP_DENYWRITE, 4, 0x81000) = 0x2b8eb7f25000 close(4)= 0 access("/etc/ld.so.nohwcap", F_OK) = -1 ENOENT (No such file or directory) open("/lib/libc.so.6", O_RDONLY)= 4 read(4, "\177ELF\2\1\1\0\0\0\0\0\0\0\0\0\3\0> \0\1\0\0\0\300\342\1\0\0\0\0\0@"..., 832) = 832 fstat(4, {st_mode=S_IFREG|0755, st_size=1375536, ...}) = 0 mmap(NULL, 3482232, PROT_READ|PROT_EXEC, MAP_PRIVATE|MAP_DENYWRITE, 4, 0) = 0x2b8eb7f27000 mprotect(0x2b8eb8071000, 2093056, PROT_NONE) = 0 mmap(0x2b8eb827, 20480, PROT_READ|PROT_WRITE, MAP_PRIVATE| MAP_FIXED|MAP_DENYWRITE, 4, 0x149000) = 0x2b8eb827 mmap(0x2b8eb8275000, 17016, PROT_READ|PROT_WRITE, MAP_PRIVATE| MAP_FIXED|MAP_ANONYMOUS, -1, 0) = 0x2b8eb8275000 close(4)= 0 mmap(NULL, 4096, PROT_READ|PROT_WR
Re: [NTG-context] calling ConTeXt as a background process
> > function os.resultof(command) >local handle = io.popen(command,"r") >return handle and handle:read("*all") or "" > end > shouldn't the handle be close()d? Patrick ___ If your question is of interest to others as well, please add an entry to the Wiki! maillist : ntg-context@ntg.nl / http://www.ntg.nl/mailman/listinfo/ntg-context webpage : http://www.pragma-ade.nl / http://tex.aanhet.net archive : http://foundry.supelec.fr/projects/contextrev/ wiki : http://contextgarden.net ___
Re: [NTG-context] calling ConTeXt as a background process
On Fri, Oct 8, 2010 at 3:47 PM, Henning Hraban Ramm wrote: > Am 2010-10-08 um 13:46 schrieb Florian Wobbe: >> >> Right, so you are not even getting to the point where luatex is called. >> This is still texlua executing instructions from mtxrun. > >>> strace -p 14256 >>> attach: ptrace(PTRACE_ATTACH, ...): Operation not permitted >>> (as the same user as well as as root) >> >> Something to do with SElinux? > > As far as I can tell, SElinux isn't active. (libselinux is installed, but no > other selinux package) > >> Here, uname -m should be called. But I guess you still have all unames >> replaced by their return value I guess? > > Yes, I didn't reinstate the original mtxrun script. > >>> brk(0x1c3e1000) = 0x1c3e1000 >>> uname({sys="Linux", node="aine.fiee.net", ...}) = 0 >>> --- SIGTERM (Terminated) @ 0 (0) --- >>> +++ killed by SIGTERM +++ >> >> Here it should continue with reading texmfcnf.lua. >> >> The last >> >> uname({sys="Linux", node="aine.fiee.net", ... >> >> you get is from the os.uname call I would guess. > > Probably. > >> Try putting the lines >> >> for k, v in next, os.uname() do >> print(k, v) >> end >> >> in a file and run it with texlua from the script to be sure. > > That works: > > texlua call_uname.lua > > version #1 SMP Tue Jul 6 14:34:09 MSD 2010 > release 2.6.18-028stab070.2 > machine x86_64 > nodename aine.fiee.net > sysname Linux > >> Then you might also try to invoke luatex directly: >> >> luatex --interaction=batchmode --fmt="/formats/cont-en" >> --lua="/formats/cont-en.lui" --backend=pdf hello > > Now I get the defunct uname again: > > 16002 15978 TS 21 15:41 ? 00:00:01 \_ /var/www/xxx/bin/python > /var/www/.../manage.py run_gunicorn -c /var/www/.../gunicorn-settings.py > 16210 16002 TS 14 15:42 ? 00:00:08 \_ luatex > --interaction=batchmode > --fmt=/var/opt/context/tex/texmf-cache/.../formats/cont-en > --lua=/var/opt/context/tex/texmf-cache/... > 16212 16210 TS 17 15:42 ? 00:00:00 \_ [uname] > > (Meanwhile I switched from fcgi to gunicorn app server, but that didn't > change anything WRT ConTeXt.) > > I get no log output. valgrind ? -- luigi ___ If your question is of interest to others as well, please add an entry to the Wiki! maillist : ntg-context@ntg.nl / http://www.ntg.nl/mailman/listinfo/ntg-context webpage : http://www.pragma-ade.nl / http://tex.aanhet.net archive : http://foundry.supelec.fr/projects/contextrev/ wiki : http://contextgarden.net ___
Re: [NTG-context] calling ConTeXt as a background process
On 10/08/2010 03:47 PM, Henning Hraban Ramm wrote: 16002 15978 TS 21 15:41 ? 00:00:01 \_ /var/www/xxx/bin/python /var/www/.../manage.py run_gunicorn -c /var/www/.../gunicorn-settings.py 16210 16002 TS 14 15:42 ? 00:00:08 \_ luatex --interaction=batchmode --fmt=/var/opt/context/tex/texmf-cache/.../formats/cont-en --lua=/var/opt/context/tex/texmf-cache/... 16212 16210 TS 17 15:42 ? 00:00:00 \_ [uname] Hm, defunct, eh? The luatex binary itself definitely does not call uname as a process, therefore this has to be the os.resultof() function that Hans implements in l-os.lua. I am suspecting issues with redirection now, as that function is defined as: function os.resultof(command) local handle = io.popen(command,"r") return handle and handle:read("*all") or "" end and it is called like so: os.resultof('uname -m') of course this will fail/block rather horribly if uname -m does not write to STDOUT (and that is not just uname, thre are a few more uses of os.resultof()). Best wishes, Taco ___ If your question is of interest to others as well, please add an entry to the Wiki! maillist : ntg-context@ntg.nl / http://www.ntg.nl/mailman/listinfo/ntg-context webpage : http://www.pragma-ade.nl / http://tex.aanhet.net archive : http://foundry.supelec.fr/projects/contextrev/ wiki : http://contextgarden.net ___
Re: [NTG-context] calling ConTeXt as a background process
Am 2010-10-08 um 13:46 schrieb Florian Wobbe: Right, so you are not even getting to the point where luatex is called. This is still texlua executing instructions from mtxrun. strace -p 14256 attach: ptrace(PTRACE_ATTACH, ...): Operation not permitted (as the same user as well as as root) Something to do with SElinux? As far as I can tell, SElinux isn't active. (libselinux is installed, but no other selinux package) Here, uname -m should be called. But I guess you still have all unames replaced by their return value I guess? Yes, I didn't reinstate the original mtxrun script. brk(0x1c3e1000) = 0x1c3e1000 uname({sys="Linux", node="aine.fiee.net", ...}) = 0 --- SIGTERM (Terminated) @ 0 (0) --- +++ killed by SIGTERM +++ Here it should continue with reading texmfcnf.lua. The last uname({sys="Linux", node="aine.fiee.net", ... you get is from the os.uname call I would guess. Probably. Try putting the lines for k, v in next, os.uname() do print(k, v) end in a file and run it with texlua from the script to be sure. That works: texlua call_uname.lua version #1 SMP Tue Jul 6 14:34:09 MSD 2010 release 2.6.18-028stab070.2 machine x86_64 nodenameaine.fiee.net sysname Linux Then you might also try to invoke luatex directly: luatex --interaction=batchmode --fmt="/formats/cont-en" -- lua="/formats/cont-en.lui" --backend=pdf hello Now I get the defunct uname again: 16002 15978 TS 21 15:41 ?00:00:01 \_ /var/www/xxx/bin/ python /var/www/.../manage.py run_gunicorn -c /var/www/.../gunicorn- settings.py 16210 16002 TS 14 15:42 ?00:00:08 \_ luatex -- interaction=batchmode --fmt=/var/opt/context/tex/texmf-cache/.../ formats/cont-en --lua=/var/opt/context/tex/texmf-cache/... 16212 16210 TS 17 15:42 ?00:00:00 \_ [uname] (Meanwhile I switched from fcgi to gunicorn app server, but that didn't change anything WRT ConTeXt.) I get no log output. BTW: Thank you very much for your patience! I owe you! (And you, Taco, of course, and the others!) Greetlings from Lake Constance! Hraban --- http://www.fiee.net/texnique/ http://wiki.contextgarden.net https://www.cacert.org (I'm an assurer) ___ If your question is of interest to others as well, please add an entry to the Wiki! maillist : ntg-context@ntg.nl / http://www.ntg.nl/mailman/listinfo/ntg-context webpage : http://www.pragma-ade.nl / http://tex.aanhet.net archive : http://foundry.supelec.fr/projects/contextrev/ wiki : http://contextgarden.net ___
Re: [NTG-context] calling ConTeXt as a background process
On Oct 8, 2010, at 12:00 , Henning Hraban Ramm wrote: > Am 2010-10-08 um 11:25 schrieb Florian Wobbe: > >> This output looks quite messy because it is composed of the mixed output for >> five processes. I don't suspect the fault in luatex executable itself, >> however, mtxrun calls uname -m in a subshell and my guess is that it hangs >> there. Have you tried calling uname alone from your script? > > As I told Taco and Luigi off-list: >> Simply calling uname works >> as well as calling it in a shell script >> as well as calling another Python script that calls uname. > > > >> Next you could do >> strace -ff -o ctxtrace mtxrun --script context --batchmode --once hello > > Ok, here we are: > 14203 ?Sl 0:01 | \_ /var/www/xxx/bin/python > /var/www/.../manage.py runfcgi method=threaded ... > 14255 ?S 0:00 | \_ strace -ff -o ctxtrace mtxrun > --script context --batchmode --once hello > 14256 ?R 0:06 | \_ texlua > /var/opt/context/tex/texmf-linux-64/bin/mtxrun --script context --batchmode > --once hello > >> This gives you hopefully four ctxtrace. files. > > I got only one for 14256 Right, so you are not even getting to the point where luatex is called. This is still texlua executing instructions from mtxrun. >> Then do (for each pid) >> strace -p >> before killing luatex and see where it hangs. Interrupt and kill your luatex >> and have a look at the ctxtrace.* files. > > strace -p 14256 > attach: ptrace(PTRACE_ATTACH, ...): Operation not permitted > (as the same user as well as as root) Something to do with SElinux? > [...] > read(3, "his script\n-- ... later we will u"..., 4096) = 4096 > read(3, "xecute\") then\n\n-- execute scr"..., 4096) = 4096 > read(3, "guments_after)\nresolvers.dowi"..., 4096) = 2858 > read(3, ""..., 4096)= 0 > close(3)= 0 > munmap(0x2acfe2776000, 4096)= 0 > brk(0x1c37a000) = 0x1c37a000 > brk(0x1c39d000) = 0x1c39d000 > brk(0x1c3c) = 0x1c3c Here, uname -m should be called. But I guess you still have all unames replaced by their return value I guess? > brk(0x1c3e1000) = 0x1c3e1000 > uname({sys="Linux", node="aine.fiee.net", ...}) = 0 > --- SIGTERM (Terminated) @ 0 (0) --- > +++ killed by SIGTERM +++ Here it should continue with reading texmfcnf.lua. The last uname({sys="Linux", node="aine.fiee.net", ... you get is from the os.uname call I would guess. Try putting the lines for k, v in next, os.uname() do print(k, v) end in a file and run it with texlua from the script to be sure. Then you might also try to invoke luatex directly: luatex --interaction=batchmode --fmt="/formats/cont-en" --lua="/formats/cont-en.lui" --backend=pdf hello Florian ___ If your question is of interest to others as well, please add an entry to the Wiki! maillist : ntg-context@ntg.nl / http://www.ntg.nl/mailman/listinfo/ntg-context webpage : http://www.pragma-ade.nl / http://tex.aanhet.net archive : http://foundry.supelec.fr/projects/contextrev/ wiki : http://contextgarden.net ___
Re: [NTG-context] calling ConTeXt as a background process
On 8-10-2010 10:00, Henning Hraban Ramm wrote: Sorry for the OT thread, I suspected the problem being more in the LuaTeX area. I never had problems running luatex (texlua) on linux boxes as part of a web service so I cannot image luatex being the problem. Can you see what pdftex does? Hans - Hans Hagen | PRAGMA ADE Ridderstraat 27 | 8061 GH Hasselt | The Netherlands tel: 038 477 53 69 | voip: 087 875 68 74 | www.pragma-ade.com | www.pragma-pod.nl - ___ If your question is of interest to others as well, please add an entry to the Wiki! maillist : ntg-context@ntg.nl / http://www.ntg.nl/mailman/listinfo/ntg-context webpage : http://www.pragma-ade.nl / http://tex.aanhet.net archive : http://foundry.supelec.fr/projects/contextrev/ wiki : http://contextgarden.net ___
Re: [NTG-context] calling ConTeXt as a background process
Am 2010-10-08 um 11:25 schrieb Florian Wobbe: This output looks quite messy because it is composed of the mixed output for five processes. I don't suspect the fault in luatex executable itself, however, mtxrun calls uname -m in a subshell and my guess is that it hangs there. Have you tried calling uname alone from your script? As I told Taco and Luigi off-list: Simply calling uname works as well as calling it in a shell script as well as calling another Python script that calls uname. Next you could do strace -ff -o ctxtrace mtxrun --script context --batchmode --once hello Ok, here we are: 14203 ?Sl 0:01 | \_ /var/www/xxx/bin/python /var/ www/.../manage.py runfcgi method=threaded ... 14255 ?S 0:00 | \_ strace -ff -o ctxtrace mtxrun --script context --batchmode --once hello 14256 ?R 0:06 | \_ texlua /var/opt/context/ tex/texmf-linux-64/bin/mtxrun --script context --batchmode --once hello This gives you hopefully four ctxtrace. files. I got only one for 14256 Then do (for each pid) strace -p before killing luatex and see where it hangs. Interrupt and kill your luatex and have a look at the ctxtrace.* files. strace -p 14256 attach: ptrace(PTRACE_ATTACH, ...): Operation not permitted (as the same user as well as as root) After killing my log file is empty. The ctxtrace file says: execve("/var/opt/context/tex/texmf-linux-64/bin/mtxrun", ["mtxrun", "-- script", "context", "--batchmode", "--once", "hello"], [/* 10 vars */]) = 0 brk(0) = 0xe092000 mmap(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0x2b782d964000 uname({sys="Linux", node="aine.fiee.net", ...}) = 0 access("/etc/ld.so.nohwcap", F_OK) = -1 ENOENT (No such file or directory) mmap(NULL, 8192, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0x2b782d965000 access("/etc/ld.so.preload", R_OK) = -1 ENOENT (No such file or directory) open("/etc/ld.so.cache", O_RDONLY) = 3 fstat(3, {st_mode=S_IFREG|0644, st_size=20573, ...}) = 0 mmap(NULL, 20573, PROT_READ, MAP_PRIVATE, 3, 0) = 0x2b782d967000 close(3)= 0 access("/etc/ld.so.nohwcap", F_OK) = -1 ENOENT (No such file or directory) open("/lib/libc.so.6", O_RDONLY)= 3 read(3, "\177ELF\2\1\1\0\0\0\0\0\0\0\0\0\3\0> \0\1\0\0\0\300\342\1\0\0\0\0\0@"..., 832) = 832 fstat(3, {st_mode=S_IFREG|0755, st_size=1375536, ...}) = 0 mmap(NULL, 3482232, PROT_READ|PROT_EXEC, MAP_PRIVATE|MAP_DENYWRITE, 3, 0) = 0x2b782db65000 mprotect(0x2b782dcaf000, 2093056, PROT_NONE) = 0 mmap(0x2b782deae000, 20480, PROT_READ|PROT_WRITE, MAP_PRIVATE| MAP_FIXED|MAP_DENYWRITE, 3, 0x149000) = 0x2b782deae000 mmap(0x2b782deb3000, 17016, PROT_READ|PROT_WRITE, MAP_PRIVATE| MAP_FIXED|MAP_ANONYMOUS, -1, 0) = 0x2b782deb3000 close(3)= 0 mmap(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0x2b782deb8000 arch_prctl(ARCH_SET_FS, 0x2b782deb86e0) = 0 mprotect(0x2b782deae000, 12288, PROT_READ) = 0 munmap(0x2b782d967000, 20573) = 0 brk(0) = 0xe092000 brk(0xe0b3000) = 0xe0b3000 execve("/var/opt/context/tex/texmf-linux-64/bin/texlua", ["texlua", "/ var/opt/context/tex/texmf-linux"..., "--script", "context", "-- batchmode", "--once", "hello"], [/* 10 vars */]) = 0 brk(0) = 0x1c0fd000 mmap(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0x2acfe2773000 uname({sys="Linux", node="aine.fiee.net", ...}) = 0 access("/etc/ld.so.nohwcap", F_OK) = -1 ENOENT (No such file or directory) mmap(NULL, 8192, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0x2acfe2774000 access("/etc/ld.so.preload", R_OK) = -1 ENOENT (No such file or directory) open("/etc/ld.so.cache", O_RDONLY) = 3 fstat(3, {st_mode=S_IFREG|0644, st_size=20573, ...}) = 0 mmap(NULL, 20573, PROT_READ, MAP_PRIVATE, 3, 0) = 0x2acfe2776000 close(3)= 0 access("/etc/ld.so.nohwcap", F_OK) = -1 ENOENT (No such file or directory) open("/lib/libdl.so.2", O_RDONLY) = 3 read(3, "\177ELF\2\1\1\0\0\0\0\0\0\0\0\0\3\0> \0\1\0\0\0\20\16\0\0\0\0\0\0@"..., 832) = 832 fstat(3, {st_mode=S_IFREG|0644, st_size=14616, ...}) = 0 mmap(NULL, 2109728, PROT_READ|PROT_EXEC, MAP_PRIVATE|MAP_DENYWRITE, 3, 0) = 0x2acfe2974000 mprotect(0x2acfe2976000, 2097152, PROT_NONE) = 0 mmap(0x2acfe2b76000, 8192, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_FIXED| MAP_DENYWRITE, 3, 0x2000) = 0x2acfe2b76000 close(3)= 0 access("/etc/ld.so.nohwcap", F_OK) = -1 ENOENT (No such file or directory) open("/lib/libm.so.6", O_RDONLY)= 3 read(3, "\177ELF\2\1\1\0\0\0\0\0\0\0\0\0\3\0>\0\1\0\0\0P> \0\0\0\0\0\0@"..., 832) = 832 fstat(3, {st_mode=S_IFREG|0644, st_size=534736, ...}) = 0 mmap(NULL, 2629848, PROT_R
Re: [NTG-context] calling ConTeXt as a background process
On Fri, Oct 8, 2010 at 11:21 AM, Henning Hraban Ramm wrote: > Am 2010-10-08 um 10:12 schrieb Patrick Gundlach: > >> I don't think it's in the LuaTeX area, for LuaTeX is just a regular >> process. > > Hm, I suspect it might be some system call inside LuaTeX that is blocked or > the like. > >> I have to admit that I have run into similar problems with different >> software I tried to run from a "server process" (whatever this means) and it >> always took me some time to get it right. At the end it was aways clear to >> me why it wouldn't run (but I can't remember anymore, since it is many years >> ago). So I suggest to replace the LuaTeX call with different other software >> and see where it breaks. Use software that reads STDIN, that writes to >> STDOUT/STDERR and alike and see if they are ok. > > A simple case like > python -c "print 'Hello'" > works ok. > I tried shell commands and simple tools (echo, set, date, touch) before, and > strace works, too. > > Even "texlua --help" works, so it's not texlua generally. > > texlua --script context --batchmode --once hello > gives: "Script file context not found" > (at least a usable error message, nothing hangs) > The same with > /var/opt/context/tex/setuptex && texlua --script context --batchmode --once > hello I use something like this os.system("""bash -c 'cd /opt/luatex/minimals-new/tex ; source setuptex ; echo $PATH;cd "%(jobpath)s" && context --batch "%(filetex)s" 2>/dev/null 1>/dev/null'""" %kv) -- luigi ___ If your question is of interest to others as well, please add an entry to the Wiki! maillist : ntg-context@ntg.nl / http://www.ntg.nl/mailman/listinfo/ntg-context webpage : http://www.pragma-ade.nl / http://tex.aanhet.net archive : http://foundry.supelec.fr/projects/contextrev/ wiki : http://contextgarden.net ___
Re: [NTG-context] calling ConTeXt as a background process
> Am 2010-10-08 um 09:59 schrieb Florian Wobbe: > >>> Thank you, but the problem is most probably something in my server setup >>> that doesn't fit something in LuaTeX. >>> As I mentioned: I got no problems calling TeX directly, just from the >>> server process. >> >> All right. I have one more suggestion. If you don't want to dive into >> debugging via gdb you should try another attempt at strace. Previously you >> posted a tracedump of context. But it only shows context (the shell script) >> waiting and nothing related to luatex. You may want to make strace follow >> forks (-f). > > > Sorry, I've no experience using strace. Some things are explained here: http://blog.ksplice.com/tag/unix/ > Here we are: > > This is how my process tree looked: > 14025 ?Sl 0:01 | \_ /var/www/xxx/bin/python > /var/www/xxx/.../manage.py runfcgi ... > 14116 ?S 0:00 | \_ /bin/sh -c set >> _context.log && > strace -f context --batchmode --once hello >> _context.log > 14117 ?S 0:00 | \_ strace -f context --batchmode > --once hello > 14118 ?T 0:00 | \_ /bin/sh > /var/opt/context/tex/texmf-linux-64/bin/context --batchmode --once hello > 14119 ?R 0:11 | \_ texlua > /var/opt/context/tex/texmf-linux-64/bin/mtxrun --script context --batchmode > --once hello > > The logfile was empty, texlua seemed to hang. I killed 14119 and got: > [...] > [pid 14119] brk(0xaa3c000) = 0xaa3c000 > [pid 14119] brk(0xaa5f000) = 0xaa5f000 > [pid 14119] brk(0xaa82000) = 0xaa82000 > [pid 14119] brk(0xaaa3000) = 0xaaa3000 > [pid 14119] uname({sys="Linux", node="aine.fiee.net", ...}) = 0 This output looks quite messy because it is composed of the mixed output for five processes. I don't suspect the fault in luatex executable itself, however, mtxrun calls uname -m in a subshell and my guess is that it hangs there. Have you tried calling uname alone from your script? Next you could do strace -ff -o ctxtrace mtxrun --script context --batchmode --once hello This gives you hopefully four ctxtrace. files. Then do (for each pid) strace -p before killing luatex and see where it hangs. Interrupt and kill your luatex and have a look at the ctxtrace.* files. Florian ___ If your question is of interest to others as well, please add an entry to the Wiki! maillist : ntg-context@ntg.nl / http://www.ntg.nl/mailman/listinfo/ntg-context webpage : http://www.pragma-ade.nl / http://tex.aanhet.net archive : http://foundry.supelec.fr/projects/contextrev/ wiki : http://contextgarden.net ___
Re: [NTG-context] calling ConTeXt as a background process
Am 2010-10-08 um 10:12 schrieb Patrick Gundlach: I don't think it's in the LuaTeX area, for LuaTeX is just a regular process. Hm, I suspect it might be some system call inside LuaTeX that is blocked or the like. I have to admit that I have run into similar problems with different software I tried to run from a "server process" (whatever this means) and it always took me some time to get it right. At the end it was aways clear to me why it wouldn't run (but I can't remember anymore, since it is many years ago). So I suggest to replace the LuaTeX call with different other software and see where it breaks. Use software that reads STDIN, that writes to STDOUT/STDERR and alike and see if they are ok. A simple case like python -c "print 'Hello'" works ok. I tried shell commands and simple tools (echo, set, date, touch) before, and strace works, too. Even "texlua --help" works, so it's not texlua generally. texlua --script context --batchmode --once hello gives: "Script file context not found" (at least a usable error message, nothing hangs) The same with /var/opt/context/tex/setuptex && texlua --script context --batchmode -- once hello Greetlings from Lake Constance! Hraban --- http://www.fiee.net/texnique/ http://wiki.contextgarden.net https://www.cacert.org (I'm an assurer) ___ If your question is of interest to others as well, please add an entry to the Wiki! maillist : ntg-context@ntg.nl / http://www.ntg.nl/mailman/listinfo/ntg-context webpage : http://www.pragma-ade.nl / http://tex.aanhet.net archive : http://foundry.supelec.fr/projects/contextrev/ wiki : http://contextgarden.net ___
Re: [NTG-context] calling ConTeXt as a background process
On Fri, Oct 8, 2010 at 10:47 AM, Henning Hraban Ramm wrote: > Am 2010-10-08 um 10:37 schrieb luigi scarso: >> >> An user permissions/environment issue ? > > > I can't exclude that, but since the web server and its subprocesses run as > the same user that I (can) use for login and got no problems running TeX as > that user, I wouldn't know where to look for the wrong permissions. try to cd to a 777 dir -- luigi ___ If your question is of interest to others as well, please add an entry to the Wiki! maillist : ntg-context@ntg.nl / http://www.ntg.nl/mailman/listinfo/ntg-context webpage : http://www.pragma-ade.nl / http://tex.aanhet.net archive : http://foundry.supelec.fr/projects/contextrev/ wiki : http://contextgarden.net ___
Re: [NTG-context] calling ConTeXt as a background process
Am 2010-10-08 um 10:37 schrieb luigi scarso: An user permissions/environment issue ? I can't exclude that, but since the web server and its subprocesses run as the same user that I (can) use for login and got no problems running TeX as that user, I wouldn't know where to look for the wrong permissions. Greetlings from Lake Constance! Hraban --- http://www.fiee.net/texnique/ http://wiki.contextgarden.net https://www.cacert.org (I'm an assurer) ___ If your question is of interest to others as well, please add an entry to the Wiki! maillist : ntg-context@ntg.nl / http://www.ntg.nl/mailman/listinfo/ntg-context webpage : http://www.pragma-ade.nl / http://tex.aanhet.net archive : http://foundry.supelec.fr/projects/contextrev/ wiki : http://contextgarden.net ___
Re: [NTG-context] calling ConTeXt as a background process
Am 2010-10-08 um 09:59 schrieb Florian Wobbe: Thank you, but the problem is most probably something in my server setup that doesn't fit something in LuaTeX. As I mentioned: I got no problems calling TeX directly, just from the server process. All right. I have one more suggestion. If you don't want to dive into debugging via gdb you should try another attempt at strace. Previously you posted a tracedump of context. But it only shows context (the shell script) waiting and nothing related to luatex. You may want to make strace follow forks (-f). Sorry, I've no experience using strace. Here we are: This is how my process tree looked: 14025 ?Sl 0:01 | \_ /var/www/xxx/bin/python /var/ www/xxx/.../manage.py runfcgi ... 14116 ?S 0:00 | \_ /bin/sh -c set >> _context.log && strace -f context --batchmode --once hello >> _context.log 14117 ?S 0:00 | \_ strace -f context -- batchmode --once hello 14118 ?T 0:00 | \_ /bin/sh /var/opt/ context/tex/texmf-linux-64/bin/context --batchmode --once hello 14119 ?R 0:11 | \_ texlua /var/opt/ context/tex/texmf-linux-64/bin/mtxrun --script context --batchmode -- once hello The logfile was empty, texlua seemed to hang. I killed 14119 and got: execve("/var/opt/context/tex/texmf-linux-64/bin/context", ["context", "--batchmode", "--once", "hello"], [/* 10 vars */]) = 0 brk(0) = 0x8e2e000 mmap(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0x2b08959ac000 uname({sys="Linux", node="aine.fiee.net", ...}) = 0 access("/etc/ld.so.nohwcap", F_OK) = -1 ENOENT (No such file or directory) mmap(NULL, 8192, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0x2b08959ad000 access("/etc/ld.so.preload", R_OK) = -1 ENOENT (No such file or directory) open("/etc/ld.so.cache", O_RDONLY) = 3 fstat(3, {st_mode=S_IFREG|0644, st_size=20573, ...}) = 0 mmap(NULL, 20573, PROT_READ, MAP_PRIVATE, 3, 0) = 0x2b08959af000 close(3)= 0 access("/etc/ld.so.nohwcap", F_OK) = -1 ENOENT (No such file or directory) open("/lib/libncurses.so.5", O_RDONLY) = 3 read(3, "\177ELF\2\1\1\0\0\0\0\0\0\0\0\0\3\0>\0\1\0\0\0\320% \1\0\0\0\0\0@"..., 832) = 832 fstat(3, {st_mode=S_IFREG|0644, st_size=256288, ...}) = 0 mmap(NULL, 2353152, PROT_READ|PROT_EXEC, MAP_PRIVATE|MAP_DENYWRITE, 3, 0) = 0x2b0895bad000 mprotect(0x2b0895be8000, 2093056, PROT_NONE) = 0 mmap(0x2b0895de7000, 20480, PROT_READ|PROT_WRITE, MAP_PRIVATE| MAP_FIXED|MAP_DENYWRITE, 3, 0x3a000) = 0x2b0895de7000 close(3)= 0 access("/etc/ld.so.nohwcap", F_OK) = -1 ENOENT (No such file or directory) open("/lib/libdl.so.2", O_RDONLY) = 3 read(3, "\177ELF\2\1\1\0\0\0\0\0\0\0\0\0\3\0> \0\1\0\0\0\20\16\0\0\0\0\0\0@"..., 832) = 832 fstat(3, {st_mode=S_IFREG|0644, st_size=14616, ...}) = 0 mmap(NULL, 2109728, PROT_READ|PROT_EXEC, MAP_PRIVATE|MAP_DENYWRITE, 3, 0) = 0x2b0895dec000 mprotect(0x2b0895dee000, 2097152, PROT_NONE) = 0 mmap(0x2b0895fee000, 8192, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_FIXED| MAP_DENYWRITE, 3, 0x2000) = 0x2b0895fee000 close(3)= 0 access("/etc/ld.so.nohwcap", F_OK) = -1 ENOENT (No such file or directory) open("/lib/libc.so.6", O_RDONLY)= 3 read(3, "\177ELF\2\1\1\0\0\0\0\0\0\0\0\0\3\0> \0\1\0\0\0\300\342\1\0\0\0\0\0@"..., 832) = 832 fstat(3, {st_mode=S_IFREG|0755, st_size=1375536, ...}) = 0 mmap(NULL, 3482232, PROT_READ|PROT_EXEC, MAP_PRIVATE|MAP_DENYWRITE, 3, 0) = 0x2b0895ff mprotect(0x2b089613a000, 2093056, PROT_NONE) = 0 mmap(0x2b0896339000, 20480, PROT_READ|PROT_WRITE, MAP_PRIVATE| MAP_FIXED|MAP_DENYWRITE, 3, 0x149000) = 0x2b0896339000 mmap(0x2b089633e000, 17016, PROT_READ|PROT_WRITE, MAP_PRIVATE| MAP_FIXED|MAP_ANONYMOUS, -1, 0) = 0x2b089633e000 close(3)= 0 mmap(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0x2b0896343000 mmap(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0x2b0896344000 arch_prctl(ARCH_SET_FS, 0x2b0896343af0) = 0 mprotect(0x2b0896339000, 12288, PROT_READ) = 0 munmap(0x2b08959af000, 20573) = 0 rt_sigprocmask(SIG_BLOCK, NULL, [], 8) = 0 open("/dev/tty", O_RDWR|O_NONBLOCK) = -1 ENXIO (No such device or address) ioctl(0, SNDCTL_TMR_TIMEBASE or TCGETS, 0x7fffd5440e70) = -1 EINVAL (Invalid argument) brk(0) = 0x8e2e000 brk(0x8e2f000) = 0x8e2f000 brk(0x8e3) = 0x8e3 getuid()= 1005 getgid()= 1006 geteuid() = 1005 getegid() = 1006 rt_sigprocmask(SIG_BLOCK, NULL, [], 8) = 0 brk(0x8e31000) = 0x8e31000 open
Re: [NTG-context] calling ConTeXt as a background process
On Fri, Oct 8, 2010 at 10:00 AM, Henning Hraban Ramm wrote: > Summary so far: > > * Machine: virtual server on Debian 5.0 (x86_64) (I thought it was Ubuntu, > because I run that on my other Linux machines, sorry). > * Webserver: Django 1.2/fcgi (Python 2.5) behind Nginx > > I can run ConTeXt without any problems from the shell, also as the same user > that runs the server process. > From the shell it doesn't matter if I call it as "context" (that's a shell > script calling mtxrun), "mtxrun", other shell script or simple Python > script. > > I want to run ConTeXt from the server process (at the end I want it to run > via Celery/RabbitMQ, but that doesn't matter at the moment), but it hangs at > the texlua process and doesn't do anything. An user permissions/environment issue ? After this I have no other suggestions. -- luigi ___ If your question is of interest to others as well, please add an entry to the Wiki! maillist : ntg-context@ntg.nl / http://www.ntg.nl/mailman/listinfo/ntg-context webpage : http://www.pragma-ade.nl / http://tex.aanhet.net archive : http://foundry.supelec.fr/projects/contextrev/ wiki : http://contextgarden.net ___
Re: [NTG-context] calling ConTeXt as a background process
> Sorry for the OT thread, I suspected the problem being more in the LuaTeX > area. I don't think it's in the LuaTeX area, for LuaTeX is just a regular process. I have to admit that I have run into similar problems with different software I tried to run from a "server process" (whatever this means) and it always took me some time to get it right. At the end it was aways clear to me why it wouldn't run (but I can't remember anymore, since it is many years ago). So I suggest to replace the LuaTeX call with different other software and see where it breaks. Use software that reads STDIN, that writes to STDOUT/STDERR and alike and see if they are ok. Patrick ___ If your question is of interest to others as well, please add an entry to the Wiki! maillist : ntg-context@ntg.nl / http://www.ntg.nl/mailman/listinfo/ntg-context webpage : http://www.pragma-ade.nl / http://tex.aanhet.net archive : http://foundry.supelec.fr/projects/contextrev/ wiki : http://contextgarden.net ___
Re: [NTG-context] calling ConTeXt as a background process
Summary so far: * Machine: virtual server on Debian 5.0 (x86_64) (I thought it was Ubuntu, because I run that on my other Linux machines, sorry). * Webserver: Django 1.2/fcgi (Python 2.5) behind Nginx I can run ConTeXt without any problems from the shell, also as the same user that runs the server process. From the shell it doesn't matter if I call it as "context" (that's a shell script calling mtxrun), "mtxrun", other shell script or simple Python script. I want to run ConTeXt from the server process (at the end I want it to run via Celery/RabbitMQ, but that doesn't matter at the moment), but it hangs at the texlua process and doesn't do anything. - With an unmodified mtxrun that hangs at its external "uname" call (gets "defunct"); I patched it to just take "x86_64" instead of calling uname. - After that it doesn't matter how I call ConTeXt, it always hangs at texlua. - I tried os.system (that directly uses C's system call) and several variants of subprocess.Popen - I tried calling setuptex as well as setting PATH, TEXROOT and TEXMFOS for the environment of the process and checked that they exist in the call. - Calls of several linux tools via os.system or subprocess.Popen worked. I suspected SElinux settings, Taco hinted at AppArmor, but both aren't active, as far as I can tell. So, the problem is with the web server and texlua. Django/fcgi unfortunately swallows error messages (known problem of the fcgi mode). Next I'll try a different server setup. Unfortunately Apache with mod_wsgi or mod_python is too heavy for my virtual server (with Nginx I can run more that 5 Django sites, with Apache I had massive ressource problems with 3-4). Green Unicorn (http://gunicorn.org/) looks good for my purpose. Sorry for the OT thread, I suspected the problem being more in the LuaTeX area. Greetlings from Lake Constance! Hraban --- http://www.fiee.net/texnique/ http://wiki.contextgarden.net https://www.cacert.org (I'm an assurer) ___ If your question is of interest to others as well, please add an entry to the Wiki! maillist : ntg-context@ntg.nl / http://www.ntg.nl/mailman/listinfo/ntg-context webpage : http://www.pragma-ade.nl / http://tex.aanhet.net archive : http://foundry.supelec.fr/projects/contextrev/ wiki : http://contextgarden.net ___
Re: [NTG-context] calling ConTeXt as a background process
On Oct 8, 2010, at 09:36 , Henning Hraban Ramm wrote: >> I've got Ubunto and OSX here and could try to reproduce the error for you, >> if you post your script. > > Thank you, but the problem is most probably something in my server setup that > doesn't fit something in LuaTeX. > As I mentioned: I got no problems calling TeX directly, just from the server > process. All right. I have one more suggestion. If you don't want to dive into debugging via gdb you should try another attempt at strace. Previously you posted a tracedump of context. But it only shows context (the shell script) waiting and nothing related to luatex. You may want to make strace follow forks (-f). Florian ___ If your question is of interest to others as well, please add an entry to the Wiki! maillist : ntg-context@ntg.nl / http://www.ntg.nl/mailman/listinfo/ntg-context webpage : http://www.pragma-ade.nl / http://tex.aanhet.net archive : http://foundry.supelec.fr/projects/contextrev/ wiki : http://contextgarden.net ___
Re: [NTG-context] calling ConTeXt as a background process
Am 2010-10-07 um 18:32 schrieb Florian Wobbe: Well, remote sensing with computers is cumbersome and as I haven't got a clue what happens on your machine I'd rather see your the shell script. Don't keep fixed on that shell script - it was only one of many failed attempts. "context" itself is only a shell script that calls mtxrun. It doesn't matter how I call ConTeXt from the server process - it hangs at texlua. How is this script intended to be called: as a subshell from the prompt, as a cron job in the background, ...? At the end I want to run it via Celery/RabbitMQ (an AMQP message queue), but since that's another complication, I tried first to run it directly from my Django view (in other framework's talk that's the controller). I've got Ubunto and OSX here and could try to reproduce the error for you, if you post your script. Thank you, but the problem is most probably something in my server setup that doesn't fit something in LuaTeX. As I mentioned: I got no problems calling TeX directly, just from the server process. A shell script with just these 4 lines #!/bin/bash cd $HOME/context/tests export PATH=$PATH:$HOME/applic/context/tex/texmf-linux-64/bin context --batchmode test.tex works as expected on Ubunto 2.6.32-24-server x86_64 with an almost empty environment (none of the TEX* variables set). You just have to set the path to context - nothing else is necessary. As for the environment variables set: you can test for that in the script via 'env' and redirect the output to a file. Guess what I tested before complaining. I redirect the whole process output to a log file, unfortunately I catch no message from TeX whatsoever, e.g. with "set && date && context hello" I get the env settings and the date, but no message from context. I'll write a summary. Greetlings from Lake Constance! Hraban --- http://www.fiee.net/texnique/ http://wiki.contextgarden.net https://www.cacert.org (I'm an assurer) ___ If your question is of interest to others as well, please add an entry to the Wiki! maillist : ntg-context@ntg.nl / http://www.ntg.nl/mailman/listinfo/ntg-context webpage : http://www.pragma-ade.nl / http://tex.aanhet.net archive : http://foundry.supelec.fr/projects/contextrev/ wiki : http://contextgarden.net ___
Re: [NTG-context] calling ConTeXt as a background process
Henning Hraban Ramm wrote: I can imagine that it has something to do with SElinux settings, didn't dive in that yet. That does sound familiar, in fact. I had a hanging database server a while back because I moved it to a different location. apparmor was the culprit in that case, so perhaps it is here as well ? (Basically an application will get a 'resource unavailable' when it tries to access anything outside of the apparmor allowed directories for that service/binary). In any case, check /etc/apparmor and /etc/apparmor.d Best wishes, Taco ___ If your question is of interest to others as well, please add an entry to the Wiki! maillist : ntg-context@ntg.nl / http://www.ntg.nl/mailman/listinfo/ntg-context webpage : http://www.pragma-ade.nl / http://tex.aanhet.net archive : http://foundry.supelec.fr/projects/contextrev/ wiki : http://contextgarden.net ___
Re: [NTG-context] calling ConTeXt as a background process
> Am 2010-10-07 um 12:54 schrieb Florian Wobbe: > >> On Oct 7, 2010, at 12:19 , Henning Hraban Ramm wrote: >> >>> Am 2010-10-07 um 11:06 schrieb Taco Hoekwater: > Of course, prd_paket.tex is my input file. But that's there. Your shell quotes were off. It is looking for "./prd_paket --batchmode --once.tex" >>> >>> Ok, but it doesn't work with any other call of luatex - I also tried to >>> start another shell script that calls luatex, but it hangs in the same >>> manner. If I kill the luatex process, further commands work. >> >> could you send me the calling shell script along with all necessary files to >> run the context batch? or supply a minimal example instead of prd_paket.tex. > > Hi Florian, > > it doesn't depend on the tex file, it's the same with > > \starttext > Hello World! > \stoptext > > It also doesn't depend on context options (--batchmode etc.) > > And it doesn't seem to depend on the way of calling it from Python/Django: > - directly with subprocess.Popen (several options), subprocess.call, > subprocess.check_call > - indirectly via shell script (that contains just the context call) > - with context or mtxrun (didn't try texlua directly yet) > - with setuptex or setting TEXROOT and TEXMFOS directly (in the environment > for the subprocess call) > > I can imagine that it has something to do with SElinux settings, didn't dive > in that yet. > > I can call context manually as the same user that's running the server > process with all the options. > > It also works on my test system (osx-intel) with Django's dev server, just > not on my Ubuntu server (linux-64). Well, remote sensing with computers is cumbersome and as I haven't got a clue what happens on your machine I'd rather see your the shell script. How is this script intended to be called: as a subshell from the prompt, as a cron job in the background, ...? I've got Ubunto and OSX here and could try to reproduce the error for you, if you post your script. A shell script with just these 4 lines #!/bin/bash cd $HOME/context/tests export PATH=$PATH:$HOME/applic/context/tex/texmf-linux-64/bin context --batchmode test.tex works as expected on Ubunto 2.6.32-24-server x86_64 with an almost empty environment (none of the TEX* variables set). You just have to set the path to context - nothing else is necessary. As for the environment variables set: you can test for that in the script via 'env' and redirect the output to a file. Cheers, Florian ___ If your question is of interest to others as well, please add an entry to the Wiki! maillist : ntg-context@ntg.nl / http://www.ntg.nl/mailman/listinfo/ntg-context webpage : http://www.pragma-ade.nl / http://tex.aanhet.net archive : http://foundry.supelec.fr/projects/contextrev/ wiki : http://contextgarden.net ___
Re: [NTG-context] calling ConTeXt as a background process
Am 2010-10-07 um 17:48 schrieb luigi scarso: How's ConTeXt called in ConTeXtgarden Live? Any special means? http://github.com/pgundlach/contextlive/blob/master/runtexexec.c I use fork()/exec() combination in C. Ok, I won't use C... I saw you set TEXMFHOME - I don't need that for MkIV, do I? 1) Are you sure to have done source setuptex before run luatex tried that and tried to set TEXROOT and TEXMFOS "manually" (and the output of "set" looks right). 2) does os.system(...) in python only work ? Thank you, I didn't think of the easiest option! With strace I get a return code of 0, so the command seems to run through, nothing hangs, but also nothing happens. Without strace texlua hangs, and after killing it I get a 36608 (don't know what that is, it's not the PID). Greetlings from Lake Constance! Hraban --- http://www.fiee.net/texnique/ http://wiki.contextgarden.net https://www.cacert.org (I'm an assurer) ___ If your question is of interest to others as well, please add an entry to the Wiki! maillist : ntg-context@ntg.nl / http://www.ntg.nl/mailman/listinfo/ntg-context webpage : http://www.pragma-ade.nl / http://tex.aanhet.net archive : http://foundry.supelec.fr/projects/contextrev/ wiki : http://contextgarden.net ___
Re: [NTG-context] calling ConTeXt as a background process
Am 2010-10-07 um 12:54 schrieb Florian Wobbe: On Oct 7, 2010, at 12:19 , Henning Hraban Ramm wrote: Am 2010-10-07 um 11:06 schrieb Taco Hoekwater: Of course, prd_paket.tex is my input file. But that's there. Your shell quotes were off. It is looking for "./prd_paket --batchmode --once.tex" Ok, but it doesn't work with any other call of luatex - I also tried to start another shell script that calls luatex, but it hangs in the same manner. If I kill the luatex process, further commands work. could you send me the calling shell script along with all necessary files to run the context batch? or supply a minimal example instead of prd_paket.tex. Hi Florian, it doesn't depend on the tex file, it's the same with \starttext Hello World! \stoptext It also doesn't depend on context options (--batchmode etc.) And it doesn't seem to depend on the way of calling it from Python/ Django: - directly with subprocess.Popen (several options), subprocess.call, subprocess.check_call - indirectly via shell script (that contains just the context call) - with context or mtxrun (didn't try texlua directly yet) - with setuptex or setting TEXROOT and TEXMFOS directly (in the environment for the subprocess call) I can imagine that it has something to do with SElinux settings, didn't dive in that yet. I can call context manually as the same user that's running the server process with all the options. It also works on my test system (osx-intel) with Django's dev server, just not on my Ubuntu server (linux-64). Despairing greetlings, Hraban ___ If your question is of interest to others as well, please add an entry to the Wiki! maillist : ntg-context@ntg.nl / http://www.ntg.nl/mailman/listinfo/ntg-context webpage : http://www.pragma-ade.nl / http://tex.aanhet.net archive : http://foundry.supelec.fr/projects/contextrev/ wiki : http://contextgarden.net ___
Re: [NTG-context] calling ConTeXt as a background process
On Thu, Oct 7, 2010 at 5:44 PM, Henning Hraban Ramm wrote: > Am 2010-10-07 um 12:27 schrieb Patrick Gundlach: > >>> How's ConTeXt called in ConTeXtgarden Live? Any special means? >> >> http://github.com/pgundlach/contextlive/blob/master/runtexexec.c >> >> I use fork()/exec() combination in C. > > Ok, I won't use C... > > I guess it's a problem of my specific combination of Ubuntu/SElinux, > Django/fcgi and LuaTeX... :-( > I found others running latex from Django in exactly the same matter as > ConTeXt fails for me. > But I can call other programs that way, so part of the problem is in LuaTeX. 1) Are you sure to have done source setuptex before run luatex 2) does os.system(...) in python only work ? -- luigi ___ If your question is of interest to others as well, please add an entry to the Wiki! maillist : ntg-context@ntg.nl / http://www.ntg.nl/mailman/listinfo/ntg-context webpage : http://www.pragma-ade.nl / http://tex.aanhet.net archive : http://foundry.supelec.fr/projects/contextrev/ wiki : http://contextgarden.net ___
Re: [NTG-context] calling ConTeXt as a background process
Am 2010-10-07 um 12:27 schrieb Patrick Gundlach: How's ConTeXt called in ConTeXtgarden Live? Any special means? http://github.com/pgundlach/contextlive/blob/master/runtexexec.c I use fork()/exec() combination in C. Ok, I won't use C... I guess it's a problem of my specific combination of Ubuntu/SElinux, Django/fcgi and LuaTeX... :-( I found others running latex from Django in exactly the same matter as ConTeXt fails for me. But I can call other programs that way, so part of the problem is in LuaTeX. Greetlings, Hraban ___ If your question is of interest to others as well, please add an entry to the Wiki! maillist : ntg-context@ntg.nl / http://www.ntg.nl/mailman/listinfo/ntg-context webpage : http://www.pragma-ade.nl / http://tex.aanhet.net archive : http://foundry.supelec.fr/projects/contextrev/ wiki : http://contextgarden.net ___
Re: [NTG-context] calling ConTeXt as a background process
Am 2010-10-07 um 12:16 schrieb : Have you tried attaching to the stalled processes using strace (Linux) or dtruss (Mac)? This will tell you what system call is blocking, might give a clue. The defunct uname process in your original post suggests to me the parent did not calling wait() on the child (uname) and it exited. Here's the output of strace after I killed the stalled process: execve("/var/opt/context/tex/texmf-linux-64/bin/context", ["/var/opt/ context/tex/texmf-linux"..., "--batchmode", "--once", "prd_streifband- d"], [/* 10 vars */]) = 0 brk(0) = 0x1e4bd000 mmap(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0x2ababa192000 uname({sys="Linux", node="aine.fiee.net", ...}) = 0 access("/etc/ld.so.nohwcap", F_OK) = -1 ENOENT (No such file or directory) mmap(NULL, 8192, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0x2ababa193000 access("/etc/ld.so.preload", R_OK) = -1 ENOENT (No such file or directory) open("/etc/ld.so.cache", O_RDONLY) = 3 fstat(3, {st_mode=S_IFREG|0644, st_size=20573, ...}) = 0 mmap(NULL, 20573, PROT_READ, MAP_PRIVATE, 3, 0) = 0x2ababa195000 close(3)= 0 access("/etc/ld.so.nohwcap", F_OK) = -1 ENOENT (No such file or directory) open("/lib/libncurses.so.5", O_RDONLY) = 3 read(3, "\177ELF\2\1\1\0\0\0\0\0\0\0\0\0\3\0>\0\1\0\0\0\320% \1\0\0\0\0\0@"..., 832) = 832 fstat(3, {st_mode=S_IFREG|0644, st_size=256288, ...}) = 0 mmap(NULL, 2353152, PROT_READ|PROT_EXEC, MAP_PRIVATE|MAP_DENYWRITE, 3, 0) = 0x2ababa393000 mprotect(0x2ababa3ce000, 2093056, PROT_NONE) = 0 mmap(0x2ababa5cd000, 20480, PROT_READ|PROT_WRITE, MAP_PRIVATE| MAP_FIXED|MAP_DENYWRITE, 3, 0x3a000) = 0x2ababa5cd000 close(3)= 0 access("/etc/ld.so.nohwcap", F_OK) = -1 ENOENT (No such file or directory) open("/lib/libdl.so.2", O_RDONLY) = 3 read(3, "\177ELF\2\1\1\0\0\0\0\0\0\0\0\0\3\0> \0\1\0\0\0\20\16\0\0\0\0\0\0@"..., 832) = 832 fstat(3, {st_mode=S_IFREG|0644, st_size=14616, ...}) = 0 mmap(NULL, 2109728, PROT_READ|PROT_EXEC, MAP_PRIVATE|MAP_DENYWRITE, 3, 0) = 0x2ababa5d2000 mprotect(0x2ababa5d4000, 2097152, PROT_NONE) = 0 mmap(0x2ababa7d4000, 8192, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_FIXED| MAP_DENYWRITE, 3, 0x2000) = 0x2ababa7d4000 close(3)= 0 access("/etc/ld.so.nohwcap", F_OK) = -1 ENOENT (No such file or directory) open("/lib/libc.so.6", O_RDONLY)= 3 read(3, "\177ELF\2\1\1\0\0\0\0\0\0\0\0\0\3\0> \0\1\0\0\0\300\342\1\0\0\0\0\0@"..., 832) = 832 fstat(3, {st_mode=S_IFREG|0755, st_size=1375536, ...}) = 0 mmap(NULL, 3482232, PROT_READ|PROT_EXEC, MAP_PRIVATE|MAP_DENYWRITE, 3, 0) = 0x2ababa7d6000 mprotect(0x2ababa92, 2093056, PROT_NONE) = 0 mmap(0x2ababab1f000, 20480, PROT_READ|PROT_WRITE, MAP_PRIVATE| MAP_FIXED|MAP_DENYWRITE, 3, 0x149000) = 0x2ababab1f000 mmap(0x2ababab24000, 17016, PROT_READ|PROT_WRITE, MAP_PRIVATE| MAP_FIXED|MAP_ANONYMOUS, -1, 0) = 0x2ababab24000 close(3)= 0 mmap(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0x2ababab29000 mmap(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0x2ababab2a000 arch_prctl(ARCH_SET_FS, 0x2ababab29af0) = 0 mprotect(0x2ababab1f000, 12288, PROT_READ) = 0 munmap(0x2ababa195000, 20573) = 0 rt_sigprocmask(SIG_BLOCK, NULL, [], 8) = 0 open("/dev/tty", O_RDWR|O_NONBLOCK) = -1 ENXIO (No such device or address) ioctl(0, SNDCTL_TMR_TIMEBASE or TCGETS, 0x7fff561b3e10) = -1 EINVAL (Invalid argument) brk(0) = 0x1e4bd000 brk(0x1e4be000) = 0x1e4be000 brk(0x1e4bf000) = 0x1e4bf000 getuid()= 1005 getgid()= 1006 geteuid() = 1005 getegid() = 1006 rt_sigprocmask(SIG_BLOCK, NULL, [], 8) = 0 brk(0x1e4c) = 0x1e4c open("/proc/meminfo", O_RDONLY) = 3 fstat(3, {st_mode=S_IFREG|0444, st_size=0, ...}) = 0 mmap(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0x2ababa195000 read(3, "MemTotal: 2097152 kB\nMemFree"..., 4096) = 771 close(3)= 0 munmap(0x2ababa195000, 4096)= 0 brk(0x1e4c1000) = 0x1e4c1000 rt_sigaction(SIGCHLD, {SIG_DFL}, {SIG_DFL}, 8) = 0 rt_sigaction(SIGCHLD, {SIG_DFL}, {SIG_DFL}, 8) = 0 rt_sigaction(SIGINT, {SIG_DFL}, {SIG_DFL}, 8) = 0 rt_sigaction(SIGINT, {SIG_DFL}, {SIG_DFL}, 8) = 0 rt_sigaction(SIGQUIT, {SIG_DFL}, {SIG_DFL}, 8) = 0 rt_sigaction(SIGQUIT, {SIG_DFL}, {SIG_DFL}, 8) = 0 rt_sigprocmask(SIG_BLOCK, NULL, [], 8) = 0 rt_sigaction(SIGQUIT, {SIG_IGN}, {SIG_DFL}, 8) = 0 uname({sys="Linux", node="aine.fiee.net", ...}) = 0 stat("/var/www/xxx/_textemp", {st_mode=S_IFDIR|0755, st_size=4096, ...}) =
Re: [NTG-context] calling ConTeXt as a background process
Hi Hraban, > > How's ConTeXt called in ConTeXtgarden Live? Any special means? Please CC me for these kind of questions, for I don't follow the list every day. I've put context live into a repo at github: http://github.com/pgundlach/contextlive and take a look at http://github.com/pgundlach/contextlive/blob/master/runtexexec.c I use fork()/exec() combination in C. Patrick ___ If your question is of interest to others as well, please add an entry to the Wiki! maillist : ntg-context@ntg.nl / http://www.ntg.nl/mailman/listinfo/ntg-context webpage : http://www.pragma-ade.nl / http://tex.aanhet.net archive : http://foundry.supelec.fr/projects/contextrev/ wiki : http://contextgarden.net ___
Re: [NTG-context] calling ConTeXt as a background process
Am 2010-10-07 um 11:06 schrieb Taco Hoekwater: Of course, prd_paket.tex is my input file. But that's there. Your shell quotes were off. It is looking for "./prd_paket --batchmode --once.tex" Ok, but it doesn't work with any other call of luatex - I also tried to start another shell script that calls luatex, but it hangs in the same manner. If I kill the luatex process, further commands work. I found a lot of examples where people call other programs from Django in the same way as I tried with LuaTeX (i.e. without additional shell script), so my call isn't probably too wrong. Greetlings from Lake Constance! Hraban --- http://www.fiee.net/texnique/ http://wiki.contextgarden.net https://www.cacert.org (I'm an assurer) ___ If your question is of interest to others as well, please add an entry to the Wiki! maillist : ntg-context@ntg.nl / http://www.ntg.nl/mailman/listinfo/ntg-context webpage : http://www.pragma-ade.nl / http://tex.aanhet.net archive : http://foundry.supelec.fr/projects/contextrev/ wiki : http://contextgarden.net ___
Re: [NTG-context] calling ConTeXt as a background process
Posted by Henning Hraban Ramm [hra...@fiee.net] > Am 2010-10-07 um 10:42 schrieb Taco Hoekwater: > > > On 10/07/10 09:34, Henning Hraban Ramm wrote: > >> mtxrun --script context "prd_paket --batchmode --once" > > > > When I run that last thing on the command line, I get: > > > > ! I can't find file `./prd_paket --batchmode --once'. > > <*> "./prd_paket --batchmode --once" > > > > Please type another input file name: > > Of course, prd_paket.tex is my input file. But that's there. > > > But it is hard to say whether that is the problem. If it was > > me, I would attach a gdb to the process to see where it is at > > from the backtrace (but you need a non-stripped build of > > the binary for that). > > I'd like to avoid compiling a debugging version of luatex... Hraban, Have you tried attaching to the stalled processes using strace (Linux) or dtruss (Mac)? This will tell you what system call is blocking, might give a clue. The defunct uname process in your original post suggests to me the parent did not calling wait() on the child (uname) and it exited. Best, Robin ___ If your question is of interest to others as well, please add an entry to the Wiki! maillist : ntg-context@ntg.nl / http://www.ntg.nl/mailman/listinfo/ntg-context webpage : http://www.pragma-ade.nl / http://tex.aanhet.net archive : http://foundry.supelec.fr/projects/contextrev/ wiki : http://contextgarden.net ___
Re: [NTG-context] calling ConTeXt as a background process
On 10/07/10 11:00, Henning Hraban Ramm wrote: > Am 2010-10-07 um 10:42 schrieb Taco Hoekwater: > >> On 10/07/10 09:34, Henning Hraban Ramm wrote: >>> mtxrun --script context "prd_paket --batchmode --once" >> >> When I run that last thing on the command line, I get: >> >> ! I can't find file `./prd_paket --batchmode --once'. >> <*> "./prd_paket --batchmode --once" >> >> Please type another input file name: > > Of course, prd_paket.tex is my input file. But that's there. Your shell quotes were off. It is looking for "./prd_paket --batchmode --once.tex" ___ If your question is of interest to others as well, please add an entry to the Wiki! maillist : ntg-context@ntg.nl / http://www.ntg.nl/mailman/listinfo/ntg-context webpage : http://www.pragma-ade.nl / http://tex.aanhet.net archive : http://foundry.supelec.fr/projects/contextrev/ wiki : http://contextgarden.net ___
Re: [NTG-context] calling ConTeXt as a background process
Am 2010-10-07 um 10:42 schrieb Taco Hoekwater: On 10/07/10 09:34, Henning Hraban Ramm wrote: mtxrun --script context "prd_paket --batchmode --once" When I run that last thing on the command line, I get: ! I can't find file `./prd_paket --batchmode --once'. <*> "./prd_paket --batchmode --once" Please type another input file name: Of course, prd_paket.tex is my input file. But that's there. But it is hard to say whether that is the problem. If it was me, I would attach a gdb to the process to see where it is at from the backtrace (but you need a non-stripped build of the binary for that). I'd like to avoid compiling a debugging version of luatex... I guess the problem isn't luatex itself, but the probably somehow restricted environment. How's ConTeXt called in ConTeXtgarden Live? Any special means? Greetlings from Lake Constance! Hraban --- http://www.fiee.net/texnique/ http://wiki.contextgarden.net https://www.cacert.org (I'm an assurer) ___ If your question is of interest to others as well, please add an entry to the Wiki! maillist : ntg-context@ntg.nl / http://www.ntg.nl/mailman/listinfo/ntg-context webpage : http://www.pragma-ade.nl / http://tex.aanhet.net archive : http://foundry.supelec.fr/projects/contextrev/ wiki : http://contextgarden.net ___
Re: [NTG-context] calling ConTeXt as a background process
On 10/07/10 09:34, Henning Hraban Ramm wrote: > mtxrun --script context "prd_paket --batchmode --once" When I run that last thing on the command line, I get: ! I can't find file `./prd_paket --batchmode --once'. <*> "./prd_paket --batchmode --once" Please type another input file name: But it is hard to say whether that is the problem. If it was me, I would attach a gdb to the process to see where it is at from the backtrace (but you need a non-stripped build of the binary for that). Best wishes, Taco ___ If your question is of interest to others as well, please add an entry to the Wiki! maillist : ntg-context@ntg.nl / http://www.ntg.nl/mailman/listinfo/ntg-context webpage : http://www.pragma-ade.nl / http://tex.aanhet.net archive : http://foundry.supelec.fr/projects/contextrev/ wiki : http://contextgarden.net ___
Re: [NTG-context] calling ConTeXt as a background process
Am 2010-10-07 um 09:06 schrieb Henning Hraban Ramm: Hi together! I'm trying since about a week to get ConTeXt running as a background process of a web application. It's always hanging, and I get no log. "ps axef" shows the call stack: \_ /var/www/xxx/bin/python /var/www/xxx/releases/current/cerebrale/ manage.py runfcgi method=threaded ... \_ /bin/sh -c cd /var/www/xxx/_textemp && /var/opt/context/tex/ texmf-linux-64/bin/context ... \_ /bin/sh /var/opt/context/tex/texmf-linux-64/bin/context -- purgeall --batchmode --result=/var/www/xxx/_textemp/xxx.pdf \_ texlua /var/opt/context/tex/texmf-linux-64/bin/mtxrun --script context --purgeall --batchmode --result=/ ... \_ [uname] I.e. it looks like texlua calls uname in a way that lets the process hang forever. If I call the same from an active shell everything runs fine. Also uname alone works in the shell command. Any ideas what I could/should check? I don't think it could be the PATH, since ConTeXt's bin as well as all the system tools are in: PATH=/var/opt/context/tex/texmf-linux-64/bin:/var/www/cerebrale/bin:/ command:/usr/local/bin:/usr/local/sbin:/bin:/sbin:/usr/bin:/usr/ sbin:/usr/X11R6/bin TEXMFOS=/var/opt/context/tex/texmf-linux-64 TEXROOT=/var/opt/context/tex The user that runs the server process (incl. sub-shell and ConTeXt) owns all the TeX files and the directory. Maybe it's a Python problem, here's my call: subprocess.Popen(cmd.encode('utf-8'), shell=True, stderr=logfile, stdout=logfile) But I can call several other commands in this way without problems. The whole thing (incl. ConTeXt) works on my Mac (OSX 10.5.8) at home. Everything I call before ConTeXt shows up in the logfile. I tried to call mtxrun directly and replaced the calls to uname -m with their return value (x86_64). Now the texlua process itself hangs, even if not "defunct" it doesn't do anything: \_ /var/www/xxx/bin/python /var/www/xxx/releases/current/cerebrale/ manage.py runfcgi method=threaded ... \_ /bin/sh -c cd /var/www/xxx/_textemp && mtxrun --script context "prd_paket --batchmode --once" PATH=/var/opt/context/tex/texmf- linux-64/bin:/var/www/xxx/bin:... \_ texlua /var/opt/context/tex/texmf-linux-64/bin/mtxrun -- script context prd_paket --batchmode --once TEXMFOS=/var/opt/context/ tex/texmf-linux-64 VIRTUAL_ENV=/var/www/xxx PATH=/var/opt/context/tex/ texmf-linux-64/bin:... Help? Greetlings from Lake Constance! Hraban --- http://www.fiee.net/texnique/ http://wiki.contextgarden.net https://www.cacert.org (I'm an assurer) ___ If your question is of interest to others as well, please add an entry to the Wiki! maillist : ntg-context@ntg.nl / http://www.ntg.nl/mailman/listinfo/ntg-context webpage : http://www.pragma-ade.nl / http://tex.aanhet.net archive : http://foundry.supelec.fr/projects/contextrev/ wiki : http://contextgarden.net ___
[NTG-context] calling ConTeXt as a background process
Hi together! I'm trying since about a week to get ConTeXt running as a background process of a web application. It's always hanging, and I get no log. "ps axef" shows the call stack: \_ /var/www/xxx/bin/python /var/www/xxx/releases/current/cerebrale/ manage.py runfcgi method=threaded ... \_ /bin/sh -c cd /var/www/xxx/_textemp && /var/opt/context/tex/ texmf-linux-64/bin/context ... \_ /bin/sh /var/opt/context/tex/texmf-linux-64/bin/context -- purgeall --batchmode --result=/var/www/xxx/_textemp/xxx.pdf \_ texlua /var/opt/context/tex/texmf-linux-64/bin/mtxrun --script context --purgeall --batchmode --result=/ ... \_ [uname] I.e. it looks like texlua calls uname in a way that lets the process hang forever. If I call the same from an active shell everything runs fine. Also uname alone works in the shell command. Any ideas what I could/should check? I don't think it could be the PATH, since ConTeXt's bin as well as all the system tools are in: PATH=/var/opt/context/tex/texmf-linux-64/bin:/var/www/cerebrale/bin:/ command:/usr/local/bin:/usr/local/sbin:/bin:/sbin:/usr/bin:/usr/sbin:/ usr/X11R6/bin TEXMFOS=/var/opt/context/tex/texmf-linux-64 TEXROOT=/var/opt/context/tex The user that runs the server process (incl. sub-shell and ConTeXt) owns all the TeX files and the directory. Maybe it's a Python problem, here's my call: subprocess.Popen(cmd.encode('utf-8'), shell=True, stderr=logfile, stdout=logfile) But I can call several other commands in this way without problems. The whole thing (incl. ConTeXt) works on my Mac (OSX 10.5.8) at home. Everything I call before ConTeXt shows up in the logfile. Greetlings from Lake Constance! Hraban --- http://www.fiee.net/texnique/ http://wiki.contextgarden.net https://www.cacert.org (I'm an assurer) ___ If your question is of interest to others as well, please add an entry to the Wiki! maillist : ntg-context@ntg.nl / http://www.ntg.nl/mailman/listinfo/ntg-context webpage : http://www.pragma-ade.nl / http://tex.aanhet.net archive : http://foundry.supelec.fr/projects/contextrev/ wiki : http://contextgarden.net ___