Changes between Version 28 and Version 29 of RemoteJob


Ignore:
Timestamp:
Apr 5, 2011, 7:37:51 AM (13 years ago)
Author:
tonig
Comment:

added FAQs

Legend:

Unmodified
Added
Removed
Modified
  • RemoteJob

    v28 v29  
    197197
    198198
    199 === Job submission ===
    200 
    201 Given that RBoinc takes care of uploading the files, will it be
    202 important to ensure that input files for different WUs have
    203 different names, so they don't conflict?
    204 
    205     The scientist can and reuse whatever names and contents he
    206     likes. The only thing that matters is the file content at the
    207     moment of submission. The RBoinc server will generate internal
    208     unique names (which are hidden from the user).
    209 
    210 
    211 Will files with identical names conflict
    212 when they end up in the ''download'' directory?
    213 
    214     No, it will be handled correctly: file names are made
    215     unique (in fact, equal to the file's MD5), even though the
    216     submitter used the same name for different contents. (It's fairly
    217     common e.g. for 1000 WUs to have in common at least one of the
    218     inputs).
    219 
    220 
    221 Will there be a problem if a user send simultaneously 1000 jobs by a
    222 script?
    223 
    224     It's perfectly fine (common indeed).
    225 
    226 === Retrieve ===
    227 
    228 How does the manual retrieve operation affect workunits once a job has
    229 been performed?
    230 
    231     Retrieve will free up space taken on the server by the results of
    232     the completed WUs. Currently running, scheduled, and future WUs
    233     will be otherwise unaffected (e.g., ''step1'' may depend on ''step0''
    234     through the chaining mechanism).
    235 
    236     Note, however, that a {{{boinc_retrieve -stop ...}}}  command will
    237     stop the chaining machinery, ie. the generation of new WUs. This is
    238     a ''stop'' operation, which should not be confused with a ''retrieve''.
    239 
    240 
    241 And what about the results if you don't perform a "retrieve"? 
    242 
    243     The results are kept on the server until the issuing scientist
    244     retrieves them. (In fact, they may fill up the server if
    245     forgotten.)
    246 
     199=== Portability and server coexistence ===
     200
     201Is your application directly compatible with
     202Windows?
     203
     204    The '''client''' scripts are likely portable with minor changes
     205    (just install a good Perl interpreter and the required modules).
     206
     207
     208Is it possible to use your commands with different BOINC projects?
     209
     210    Normally a ''RBoinc server'' is tied to the corresponding
     211    ''Boinc server'' (but it can talk to any of its applications).  The
     212    ''RBoinc clients'' can talk to multiple servers using different URLs
     213    ('''-url''' option).
     214
     215
     216I was thinking of having several Boinc projects to assign easily each
     217of them to a specific group of users.  Should I be installing multiple
     218servers with their own URLs?
     219
     220    Yes - that will be the most flexible solution (e.g. you wil be
     221    able to perform maintenance on each of them separately, make
     222    firewall rules, etc). Alternatively, you can create several
     223    applications on the same server, if you prefer.
     224
     225
     226=== Server and installation ===
    247227
    248228What is the relationship between RBoinc and the assimilator?
     
    266246
    267247
     248
     249
     250What is ''ENCODE_EXECUTABLE'' ?
     251
     252    It's an optional script to pre-process files before submission, e.g. for a server-side sanity check, or reformatting, etc.
     253
     254
     255What is ETC_DIR => "$cgi/etc" ?  Do I have to copy files included in server/etc ? Which ones will be used ?
     256
     257    Most notably, the files which are supplied as defaults (if any), e.g. those marked as         {{{<rboinc immutable="true"/>}}} or  {{{<rboinc optional="true"/> }}}     in the template.
     258
     259
     260What is  "process.sh"? Is it a daemon? What if server is restarted ?
     261
     262    ''process'' created automatically by the boinc_submit_server in each GROUP directory. (It's at the end of the server submission script.) It's invoked in two cases:
     263
     264      1. as the last step of the submission, when everything has been set up. It will create_work and exit.
     265      1. it will be invoked by the assimilator. if chaining is required,  it will create_work for the next "chain step" (depending on the one which just ended successfully) and exit.
     266
     267    Therefore its execution is short and usually triggered by a daemon; stopping/restarting boinc will not affect it.
     268
     269
     270=== Job submission ===
     271
     272Given that RBoinc takes care of uploading the files, will it be
     273important to ensure that input files for different WUs have
     274different names, so they don't conflict?
     275
     276    The scientist can and reuse whatever names and contents he
     277    likes. The only thing that matters is the file content at the
     278    moment of submission. The RBoinc server will generate internal
     279    unique names (which are hidden from the user).
     280
     281
     282Will files with identical names conflict
     283when they end up in the ''download'' directory?
     284
     285    No, it will be handled correctly: file names are made
     286    unique (in fact, equal to the file's MD5), even though the
     287    submitter used the same name for different contents. (It's fairly
     288    common e.g. for 1000 WUs to have in common at least one of the
     289    inputs).
     290
     291
     292Will there be a problem if a user send simultaneously 1000 jobs by a
     293script?
     294
     295    It's perfectly fine (common indeed).
     296
     297=== Retrieve ===
     298
     299How does the manual retrieve operation affect workunits once a job has
     300been performed?
     301
     302    Retrieve will free up space taken on the server by the results of
     303    the completed WUs. Currently running, scheduled, and future WUs
     304    will be otherwise unaffected (e.g., ''step1'' may depend on ''step0''
     305    through the chaining mechanism).
     306
     307    Note, however, that a {{{boinc_retrieve -stop ...}}}  command will
     308    stop the chaining machinery, ie. the generation of new WUs. This is
     309    a ''stop'' operation, which should not be confused with a ''retrieve''.
     310
     311
     312And what about the results if you don't perform a "retrieve"? 
     313
     314    The results are kept on the server until the issuing scientist
     315    retrieves them. (In fact, they may fill up the server if
     316    forgotten.)
     317
     318
     319
    268320The scientist may request the output at any time.  Is it also possible
    269321to retrieve automatically the output in a predefined folder when a job is
     
    278330
    279331
    280 === Portability and server coexistence ===
    281 
    282 Is your application directly compatible with
    283 Windows?
    284 
    285     The '''client''' scripts are likely portable with minor changes
    286     (just install a good Perl interpreter and the required modules).
    287 
    288 
    289 Is it possible to use your commands with different BOINC projects?
    290 
    291     Normally a ''RBoinc server'' is tied to the corresponding
    292     ''Boinc server'' (but it can talk to any of its applications).  The
    293     ''RBoinc clients'' can talk to multiple servers using different URLs
    294     ('''-url''' option).
    295 
    296 
    297 I was thinking of having several Boinc projects to assign easily each
    298 of them to a specific group of users.  Should I be installing multiple
    299 servers with their own URLs?
    300 
    301     Yes - that will be the most flexible solution (e.g. you wil be
    302     able to perform maintenance on each of them separately, make
    303     firewall rules, etc). Alternatively, you can create several
    304     applications on the same server, if you prefer.
    305 
    306332
    307333(Many thanks to Boris Dayma).