Re: [fossil-users] 'fossil clone' from localhost is extremely slow
On Thu, 10 Jan 2013 13:57:54 +0100 Stephan Beal sgb...@googlemail.com wrote: On Thu, Jan 10, 2013 at 1:53 PM, Eduardo Morras emorr...@yahoo.es wrote: If he tries in shell 1 fossil open repository fossil server and in shell 2 fossil clone http://localhost:8080/ copy.fossil should workaround it, the repository is opened only once In theory, yes, but the web server forks a child process for each request and (as i learned via a recent thread in the sqlite list) the db must be opened by the child processes in order for locking to work properly. If BSD is not killing the child processes immediately after each request is finished then the db could be held open during the multiple requests sent by a clone operation. In theory/speculation. I thought that only in cgi mode it's forked for each request, I have to look that part of fossil code more deeply. I think in server mode fossil could/should/may keep running without exit after each request. It could make easier the xfer code for clone pull and push, so a cookie is not needed to continue the download/upload. -- - stephan beal http://wanderinghorse.net/home/stephan/ http://gplus.to/sgbeal --- --- Eduardo Morras emorr...@yahoo.es ___ fossil-users mailing list fossil-users@lists.fossil-scm.org http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users
Re: [fossil-users] 'fossil clone' from localhost is extremely slow
On Sat, Jan 12, 2013 at 9:58 AM, Eduardo Morras emorr...@yahoo.es wrote: I thought that only in cgi mode it's forked for each request, I have to look that part of fossil code more deeply. I think in server mode fossil could/should/may keep running without exit after each request. It's the other way around - CGI mode does not fork (because the web server launches one process per CGI request) but server mode is a thin shim around CGI mode which simply answers connections and forks the children in CGI mode. Fossil cannot really handle multiple requests from a single process because fossil makes heavy use of global state, so concurrent requests would all step on each other (e.g. the login info is held in global state). The server process itself keeps running, but sits in an endless loop doing select()/accept()/fork(), and does not touch the global repo state (that's done by the children), but it does touch _some_ global state (e.g. the http/sql tracing flags). i don't have the code handy here, but IIRC the server's listener loop is implemented in cgi.c and the server command code is in main.c. -- - stephan beal http://wanderinghorse.net/home/stephan/ http://gplus.to/sgbeal ___ fossil-users mailing list fossil-users@lists.fossil-scm.org http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users
Re: [fossil-users] 'fossil clone' from localhost is extremely slow
Hello, On Thu, Jan 10, 2013 at 1:03 AM, Joerg Sonnenberger jo...@britannica.bec.de wrote: On Wed, Jan 09, 2013 at 05:02:14PM +, Michai Ramakers wrote: apart from the question whether cloning from localhost makes sense or not (I use this from a script to make work from localhost or remote transparent), I experienced very slow network traffic - but no hang - while cloning like thus: Can you run echo 'analyze;' | fossil sqlite ori.fossil first and see if that helps? I didn't know the 'analyze' sqlite command, thank you. When trying (I am assuming you meant with '-R' option to point to repo), I got slightly more than I bargained for: --- michai@lime:/tmp$ uname -a NetBSD lime 6.0.1 NetBSD 6.0.1 (GENERIC) amd64 michai@lime:/tmp$ /tmp/fossil ver This is fossil version 1.25 [baa1ebb7d9] 2013-01-07 18:58:30 UTC michai@lime:/tmp$ echo 'analyze;' | /tmp/fossil sqlite -R 100MB.fossil michai@lime:/tmp$ echo 'analyze;' | /tmp/fossil sqlite -R 100MB.fossil Segmentation fault (core dumped) michai@lime:/tmp$ mkdir try michai@lime:/tmp$ cd try/ michai@lime:/tmp/try$ /tmp/fossil open ../100MB.fossil Segmentation fault (core dumped) michai@lime:/tmp/try$ gdb /tmp/fossil GNU gdb (GDB) 7.3.1 Copyright (C) 2011 Free Software Foundation, Inc. License GPLv3+: GNU GPL version 3 or later http://gnu.org/licenses/gpl.html This is free software: you are free to change and redistribute it. There is NO WARRANTY, to the extent permitted by law. Type show copying and show warranty for details. This GDB was configured as x86_64--netbsd. For bug reporting instructions, please see: http://www.gnu.org/software/gdb/bugs/... Reading symbols from /tmp/fossil...done. (gdb) run open ../100MB.fossil Starting program: /tmp/fossil open ../100MB.fossil Program received signal SIGSEGV, Segmentation fault. 0x in ?? () (gdb) bt #0 0x in ?? () #1 0x004b8fa2 in sqlite3VdbeExec (p=optimized out) at ./src/sqlite3.c:65390 #2 0x004a91e7 in sqlite3Step (p=0x7f7ff7e3b048) at ./src/sqlite3.c:62231 #3 sqlite3_step (pStmt=0x7f7ff7e3b048) at ./src/sqlite3.c:62305 #4 0x004aa17f in loadStat3 (zDb=0x6974df main, db=0x7f7ff7e03408) at ./src/sqlite3.c:79808 #5 sqlite3AnalysisLoad (db=0x7f7ff7e03408, iDb=optimized out) at ./src/sqlite3.c:14435 #6 0x004aa7c7 in sqlite3InitOne (db=0x7f7ff7e03408, iDb=0, pzErrMsg=0x7f7ff7e03c10) at ./src/sqlite3.c:93962 #7 0x004aa977 in sqlite3Init (db=0x7f7ff7e03408, pzErrMsg=0x7f7ff7e03c10) at ./src/sqlite3.c:94019 #8 0x004aa9ee in sqlite3ReadSchema (pParse=0x7f7ff7e03c08) at ./src/sqlite3.c:94056 #9 0x004ab23d in sqlite3LocateTable (pParse=0x7f7ff7e03c08, isView=0, zName=0x7f7ff7e2ed88 config, zDbase=0x0) at ./src/sqlite3.c:81103 #10 0x004ab3d5 in selectExpander (pWalker=0x7f7fd350, p=0x7f7ff7e2eb88) at ./src/sqlite3.c:97832 #11 0x004681f9 in sqlite3WalkSelect (pWalker=0x7f7fd350, p=0x7f7ff7e2eb88) at ./src/sqlite3.c:72494 #12 0x004684ce in sqlite3SelectExpand (pSelect=0x7f7ff7e2eb88, pParse=0x7f7ff7e03c08) at ./src/sqlite3.c:98063 #13 sqlite3SelectPrep (pParse=0x7f7ff7e03c08, p=0x7f7ff7e2eb88, pOuterNC=optimized out) at ./src/sqlite3.c:32611 #14 0x0048d8ac in sqlite3Select (pParse=0x7f7ff7e03c08, p=0x7f7ff7e2eb88, pDest=0x7f7fd6c0) at ./src/sqlite3.c:98412 #15 0x004a455a in yy_reduce (yyruleno=112, yypParser=0x7f7ff7e30008) at ./src/sqlite3.c:110655 #16 sqlite3Parser (yyp=optimized out, yymajor=1, yyminor=..., pParse=optimized out) at ./src/sqlite3.c:46121#17 0x004a6d6c in sqlite3RunParser (pParse=0x7f7ff7e03c08, zSql=optimized out, pzErrMsg=0x7f7fd7b0) at ./src/sqlite3.c:112494 #18 0x004a8a58 in sqlite3Prepare (db=0x7f7ff7e03408, zSql=0x7f7ff7e0b300 SELECT value FROM config WHERE name='allow-symlinks', nBytes=-1, saveSqlFlag=1, pReprepare=0x0, ppStmt=0x7f7fd8d0, pzTail=0x0) at ./src/sqlite3.c:94233 #19 0x004a8d2c in sqlite3LockAndPrepare (db=0x7f7ff7e03408, zSql=0x7f7ff7e0b300 SELECT value FROM config WHERE name='allow-symlinks', nBytes=-1, saveSqlFlag=1, pOld=0x0, ppStmt=0x7f7fd8d0, pzTail=0x0) at ./src/sqlite3.c:94325 #20 0x004a8ec8 in sqlite3_prepare_v2 (db=optimized out, zSql=optimized out, nBytes=optimized out, ppStmt=optimized out, pzTail=optimized out) at ./src/sqlite3.c:94401 #21 0x004113a5 in db_vprepare (pStmt=0x7f7fd8b0, errOk=0, zFormat=optimized out, ap=0x7f7fd8f0) at ./src/db.c:256 #22 0x00411492 in db_text (zDefault=0x0, zSql=optimized out) at ./src/db.c:620 #23 0x004123ac in db_get (zName=0x60553d allow-symlinks, zDefault=0x6bb4bf off) at ./src/db.c:1769 #24 0x00412ab9 in db_get_boolean (zName=optimized out, dflt=0) at ./src/db.c:1860 #25 0x00412b95 in db_open_repository (zDbName=0x7f7ff7e01060 /tmp/100MB.fossil) at ./src/db.c:1006 #26 0x004132d5 in cmd_open () at ./src/db.c:1963 #27 0x0042e90b in main (argc=optimized out,
Re: [fossil-users] 'fossil clone' from localhost is extremely slow
I didn't know the 'analyze' sqlite command, thank you. When trying (I am assuming you meant with '-R' option to point to repo), I got slightly more than I bargained for: [...] Segmentation fault (core dumped) Retried with newer fossil-version [c7133bd79d] from yesterday, and segfault is gone :-) (Perhaps either checkin [6b3e97a328] or [1a52914b38] (both sqlite-related) could have fixed it, I don't know.) The original 'slow clone from localhost' issue remains. Assuming this is not a fossil bug, sorry for the noise and thank you for thinking along. Michai ___ fossil-users mailing list fossil-users@lists.fossil-scm.org http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users
Re: [fossil-users] 'fossil clone' from localhost is extremely slow
On Thu, Jan 10, 2013 at 10:19:31AM +, Michai Ramakers wrote: Hello, On Thu, Jan 10, 2013 at 1:03 AM, Joerg Sonnenberger jo...@britannica.bec.de wrote: On Wed, Jan 09, 2013 at 05:02:14PM +, Michai Ramakers wrote: apart from the question whether cloning from localhost makes sense or not (I use this from a script to make work from localhost or remote transparent), I experienced very slow network traffic - but no hang - while cloning like thus: Can you run echo 'analyze;' | fossil sqlite ori.fossil first and see if that helps? I didn't know the 'analyze' sqlite command, thank you. When trying (I am assuming you meant with '-R' option to point to repo), I got slightly more than I bargained for: [snip] Heh, that's for drh. You said sqlite analyze works? Does it help or not? Joerg ___ fossil-users mailing list fossil-users@lists.fossil-scm.org http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users
Re: [fossil-users] 'fossil clone' from localhost is extremely slow
On Thu, Jan 10, 2013 at 11:22 AM, Joerg Sonnenberger I didn't know the 'analyze' sqlite command, thank you. When trying (I am assuming you meant with '-R' option to point to repo), I got slightly more than I bargained for: [snip] Heh, that's for drh. You said sqlite analyze works? Does it help or not? ah yes... forgot that :) It does not help. Note that taking 1 cpu core offline ('cpuctl offline 1' here) makes it fly. Michai ___ fossil-users mailing list fossil-users@lists.fossil-scm.org http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users
Re: [fossil-users] 'fossil clone' from localhost is extremely slow
On Thu, Jan 10, 2013 at 11:22 AM, Joerg Sonnenberger jo...@britannica.bec.de wrote: Heh, that's for drh. You said sqlite analyze works? Does it help or not? Also... see http://www.fossil-scm.org/xfer/info/85233c40c9bb05a87574cdc25402ec0d34b0b7ee : queries seem to be optimised to run without 'analyze' db info, but of course worth a try. Michai ___ fossil-users mailing list fossil-users@lists.fossil-scm.org http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users
Re: [fossil-users] 'fossil clone' from localhost is extremely slow
On Thu, Jan 10, 2013 at 12:32 PM, Michai Ramakers m.ramak...@gmail.comwrote: ah yes... forgot that :) It does not help. Note that taking 1 cpu core offline ('cpuctl offline 1' here) makes it fly This is just pure speculation, but ... perhaps there is a problem in BSD multi-core cases, such that the various requests made by fossil for cloning do not get closed/killed immediately, and each request is locking the db for longer than it should because its process it not being killed in a timely manner? It would be interesting to see if, when you see the hang-ups, if the child process forked for handling a request is still alive (on the other CPU) for a while after the subsequent request has already started. i.e... clone request 1 comes in and completes, but its process (or flock) is not closed/released in a timely manner. Then comes clone request #2, which is waiting on an flock which is still held by request 1. Ad nauseum for requests #3 to #N. The output of lsof the_repo_file might be useful here. (Again, purely speculation...) -- - stephan beal http://wanderinghorse.net/home/stephan/ http://gplus.to/sgbeal ___ fossil-users mailing list fossil-users@lists.fossil-scm.org http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users
Re: [fossil-users] 'fossil clone' from localhost is extremely slow
On Thu, 10 Jan 2013 13:21:50 +0100 Stephan Beal sgb...@googlemail.com wrote: On Thu, Jan 10, 2013 at 12:32 PM, Michai Ramakers m.ramak...@gmail.comwrote: ah yes... forgot that :) It does not help. Note that taking 1 cpu core offline ('cpuctl offline 1' here) makes it fly This is just pure speculation, but ... perhaps there is a problem in BSD multi-core cases, such that the various requests made by fossil for cloning do not get closed/killed immediately, and each request is locking the db for longer than it should because its process it not being killed in a timely manner? It would be interesting to see if, when you see the hang-ups, if the child process forked for handling a request is still alive (on the other CPU) for a while after the subsequent request has already started. i.e... clone request 1 comes in and completes, but its process (or flock) is not closed/released in a timely manner. Then comes clone request #2, which is waiting on an flock which is still held by request 1. Ad nauseum for requests #3 to #N. The output of lsof the_repo_file might be useful here. (Again, purely speculation...) If he tries in shell 1 fossil open repository fossil server and in shell 2 fossil clone http://localhost:8080/ copy.fossil should workaround it, the repository is opened only once -- - stephan beal http://wanderinghorse.net/home/stephan/ http://gplus.to/sgbeal --- --- Eduardo Morras emorr...@yahoo.es ___ fossil-users mailing list fossil-users@lists.fossil-scm.org http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users
Re: [fossil-users] 'fossil clone' from localhost is extremely slow
On Thu, Jan 10, 2013 at 1:53 PM, Eduardo Morras emorr...@yahoo.es wrote: If he tries in shell 1 fossil open repository fossil server and in shell 2 fossil clone http://localhost:8080/ copy.fossil should workaround it, the repository is opened only once In theory, yes, but the web server forks a child process for each request and (as i learned via a recent thread in the sqlite list) the db must be opened by the child processes in order for locking to work properly. If BSD is not killing the child processes immediately after each request is finished then the db could be held open during the multiple requests sent by a clone operation. In theory/speculation. -- - stephan beal http://wanderinghorse.net/home/stephan/ http://gplus.to/sgbeal ___ fossil-users mailing list fossil-users@lists.fossil-scm.org http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users
Re: [fossil-users] 'fossil clone' from localhost is extremely slow
On Wed, Jan 9, 2013 at 12:02 PM, Michai Ramakers m.ramak...@gmail.comwrote: Hello, apart from the question whether cloning from localhost makes sense or not (I use this from a script to make work from localhost or remote transparent), I experienced very slow network traffic - but no hang - while cloning like thus: in shell 1: # fossil server ori.fossil in shell 2: $ fossil clone http://michai@localhost:8080/ copy.fossil Might be faster and easier to do: fossil clone file::///path/to/ori.fossil copy.fossil Dunno why your localhost clone is slow. It is really fast for me here on Linux, Window8, MacOS, and Sparc-Solaris. -- D. Richard Hipp d...@sqlite.org ___ fossil-users mailing list fossil-users@lists.fossil-scm.org http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users
Re: [fossil-users] 'fossil clone' from localhost is extremely slow
On Wed, Jan 9, 2013 at 12:16 PM, Michai Ramakers m.ramak...@gmail.comwrote: Hello, in shell 1: # fossil server ori.fossil in shell 2: $ fossil clone http://michai@localhost:8080/ copy.fossil Might be faster and easier to do: fossil clone file::///path/to/ori.fossil copy.fossil Dunno why your localhost clone is slow. It is really fast for me here on Linux, Window8, MacOS, and Sparc-Solaris. Alright, thank you; I'll use that as a workaround. I will try this version of fossil on a 32-bit host and see what happens. Would you be interested in a repo that reproduces this (here), assuming I can make one? Yes. with kind regards, Michai Ramakers ___ fossil-users mailing list fossil-users@lists.fossil-scm.org http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users -- D. Richard Hipp d...@sqlite.org ___ fossil-users mailing list fossil-users@lists.fossil-scm.org http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users
Re: [fossil-users] 'fossil clone' from localhost is extremely slow
Michai Ramakers wrote: [...] ...resulting in about 70 kB/s traffic with very rare (a minute or more in between - didn't time exactly) single bursts. According to top(1), neither server- nor client-fossils seem to be doing anything CPU-intensive at that time. [...] Both 'local to local' clones were slow (70 kB/s), while both 'remote to local' clones were fast (around 7 MB/s). My instant gut reaction here is to say 'IPv6'. If your DNS setup is wrong on the 'local' machine, then whenever it tries to look up localhost it could well be spending ages trying to query it up the DNS hierarchy. This would explain the burst traffic: every HTTP connection would involve a very long pause for the lookup, and then normal traffic thereafter. I don't know if this is actually the case or not, but don't ask me how to fix it --- I don't use IPv6 on any machines I admin. But these links might be relevant: http://www.linuxquestions.org/questions/linux-networking-3/ipv6-stopping-dns-lookup-on-localhost-207066/ https://bugzilla.mozilla.org/show_bug.cgi?id=86449 -- ┌─── dg@cowlark.com ─ http://www.cowlark.com ─ │ There is nothing in the world so dangerous --- and I mean *nothing* │ --- as a children's story that happens to be true. --- Master Li Kao, │ _The Bridge of Birds_ signature.asc Description: OpenPGP digital signature ___ fossil-users mailing list fossil-users@lists.fossil-scm.org http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users
Re: [fossil-users] 'fossil clone' from localhost is extremely slow
Hello, On Wed, Jan 9, 2013 at 5:30 PM, David Given d...@cowlark.com wrote: Michai Ramakers wrote: [...] ...resulting in about 70 kB/s traffic with very rare (a minute or more in between - didn't time exactly) single bursts. According to top(1), neither server- nor client-fossils seem to be doing anything CPU-intensive at that time. [...] Both 'local to local' clones were slow (70 kB/s), while both 'remote to local' clones were fast (around 7 MB/s). My instant gut reaction here is to say 'IPv6'. If your DNS setup is wrong on the 'local' machine, then whenever it tries to look up localhost it could well be spending ages trying to query it up the DNS hierarchy. This would explain the burst traffic: every HTTP connection would involve a very long pause for the lookup, and then normal traffic thereafter. I don't know if this is actually the case or not, but don't ask me how to fix it --- I don't use IPv6 on any machines I admin. But these links might be relevant: http://www.linuxquestions.org/questions/linux-networking-3/ipv6-stopping-dns-lookup-on-localhost-207066/ https://bugzilla.mozilla.org/show_bug.cgi?id=86449 Thank you for the suggestion, however: $ /tmp/fossil clone http://michai@10.1.1.11:8080/ copy.fossil (same result - slow) Michai ___ fossil-users mailing list fossil-users@lists.fossil-scm.org http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users
Re: [fossil-users] 'fossil clone' from localhost is extremely slow
On Wed, Jan 9, 2013 at 5:20 PM, Richard Hipp d...@sqlite.org wrote: [...] Would you be interested in a repo that reproduces this (here), assuming I can make one? Yes. I can reproduce this with a 100 MB repo (first try - I guess it's not the repo then :-) I will now try to make it a bit smaller for mailing. Where can I put/send it? Michai ___ fossil-users mailing list fossil-users@lists.fossil-scm.org http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users
Re: [fossil-users] 'fossil clone' from localhost is extremely slow
Michai Ramakers wrote: [...] Thank you for the suggestion, however: $ /tmp/fossil clone http://michai@10.1.1.11:8080/ copy.fossil (same result - slow) N! My precious credibility! Well, it's good to be able to rule that out. I have no more ideas, unfortunately. -- ┌─── dg@cowlark.com ─ http://www.cowlark.com ─ │ There is nothing in the world so dangerous --- and I mean *nothing* │ --- as a children's story that happens to be true. --- Master Li Kao, │ _The Bridge of Birds_ signature.asc Description: OpenPGP digital signature ___ fossil-users mailing list fossil-users@lists.fossil-scm.org http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users
Re: [fossil-users] 'fossil clone' from localhost is extremely slow
On Wed, Jan 9, 2013 at 6:45 PM, Michai Ramakers m.ramak...@gmail.comwrote: I will now try to make it a bit smaller for mailing. Where can I put/send it? This doesn't answer your question, but your question got me thinking about using SQL to back up or vacuum repos... stephan@tiny:~/cvs/fossil$ gzip -c fossil.fsl fossil.fsl.gz stephan@tiny:~/cvs/fossil$ l fossil.fsl* -rw-rw-rw- 1 stephan stephan 35969024 Jan 8 17:55 fossil.fsl -rw-rw-r-- 1 stephan stephan 30286793 Jan 9 18:53 fossil.fsl.gz stephan@tiny:~/cvs/fossil$ echo '.dump' | sqlite3 fossil.fsl fossil.sql; l fossil.sql; gzip fossil.sql; l fossil.sql.gz -rw-rw-r-- 1 stephan stephan 46902857 Jan 9 18:54 fossil.sql -rw-rw-r-- 1 stephan stephan 24109773 Jan 9 18:54 fossil.sql.gz that SQL file, which is slightly smaller than fossil.fsl.gz, can then be fed back into sqlite3 to recreate the repo: stephan@tiny:~/cvs/fossil$ zcat fossil.sql.gz | sqlite3 foo.fsl; mkdir foo; cd foo stephan@tiny:~/cvs/fossil/foo$ l ../foo.fsl -rw-r--r-- 1 stephan stephan 29767680 Jan 9 18:56 ../foo.fsl stephan@tiny:~/cvs/fossil/foo$ f open ../foo.fsl | tail www/wikitheory.wiki project-name: Fossil repository: /home/stephan/cvs/fossil/foo/../foo.fsl local-root: /home/stephan/cvs/fossil/foo/ project-code: CE59BB9F186226D80E49D1FA2DB29F935CCA0333 checkout: 0fb6c829f2dbf6580315c05eb6c854065d1b9267 2013-01-08 16:55:39 UTC parent: be93ecb3a5758b84592f6fb0249420744f10c5ca 2013-01-08 15:58:42 UTC tags: trunk comment: Added missing commit example to 5-minute quickstart. (user: stephan) -- - stephan beal http://wanderinghorse.net/home/stephan/ http://gplus.to/sgbeal ___ fossil-users mailing list fossil-users@lists.fossil-scm.org http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users
Re: [fossil-users] 'fossil clone' from localhost is extremely slow
On Wed, Jan 9, 2013 at 5:20 PM, Richard Hipp d...@sqlite.org wrote: On Wed, Jan 9, 2013 at 12:16 PM, Michai Ramakers m.ramak...@gmail.com wrote: Hello, in shell 1: # fossil server ori.fossil in shell 2: $ fossil clone http://michai@localhost:8080/ copy.fossil Might be faster and easier to do: fossil clone file::///path/to/ori.fossil copy.fossil Dunno why your localhost clone is slow. It is really fast for me here on Linux, Window8, MacOS, and Sparc-Solaris. Alright, thank you; I'll use that as a workaround. I will try this version of fossil on a 32-bit host and see what happens. Would you be interested in a repo that reproduces this (here), assuming I can make one? Yes. ok, I made new repos containing... - 1 file of 1 MB of random: reproduces 3 out of 10 times here - 1 file of 10 MB of random: reproduces 4/10 times here - repo with around 2100 files, around 500 MB: reproduces 10/10 times here all are available, of course. I restarted the server between each test. Unless any suggestions come up, I'll stop with statistics for a while now... :-) What I did notice what that when a test failed (i.e. traffic became slow), it would typically show a few (let's say 5 or less) 'round trips' as shown in the progress feedback - no idea if this is relevant. Michai ___ fossil-users mailing list fossil-users@lists.fossil-scm.org http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users
Re: [fossil-users] 'fossil clone' from localhost is extremely slow
- repo with around 2100 files, around 500 MB: reproduces 10/10 times make that 100 MB. ___ fossil-users mailing list fossil-users@lists.fossil-scm.org http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users
Re: [fossil-users] 'fossil clone' from localhost is extremely slow
- repo with around 2100 files, around 500 MB: reproduces 10/10 times make that 100 MB. and 461 files... slight mixup. ___ fossil-users mailing list fossil-users@lists.fossil-scm.org http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users
Re: [fossil-users] 'fossil clone' from localhost is extremely slow
On Wed, Jan 09, 2013 at 05:02:14PM +, Michai Ramakers wrote: apart from the question whether cloning from localhost makes sense or not (I use this from a script to make work from localhost or remote transparent), I experienced very slow network traffic - but no hang - while cloning like thus: Can you run echo 'analyze;' | fossil sqlite ori.fossil first and see if that helps? Joerg ___ fossil-users mailing list fossil-users@lists.fossil-scm.org http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users