Hi There, I''m a student at NYIT, New York and I have project on xen hypervisor. Currently i''m working on xentrace and xenalyze to monitor the runstate events. What i''m trying do now is: While i''m capturing by xentrace I want to display the result on the screen using xenalyze. I tried to use loop in the main function but I got errors and when I tried to solve these errors I got more errors. int main(int argc, char *argv[]) { /* Start with warn at stderr. */ warn = stderr; argp_parse(&parser_def, argc, argv, 0, NULL, NULL); if (G.trace_file == NULL) exit(1); if ( (G.fd = open(G.trace_file, O_RDONLY|O_LARGEFILE)) < 0) { perror("open"); error(ERR_SYSTEM, NULL); } else { struct stat64 s; fstat64(G.fd, &s); G.file_size = s.st_size; } if ( (G.mh = mread_init(G.fd)) == NULL ) perror("mread"); if (G.symbol_file != NULL) parse_symbol_file(G.symbol_file); if(opt.dump_all) warn = stdout; for (int i=0;i<=1000;i++){ init_pcpus(); if(opt.progress) progress_init(); process_records(); if(opt.interval_mode) interval_tail(); if(opt.summary) summary(); if(opt.report_pcpu) report_pcpu(); sleep(2); } if(opt.progress) progress_finish(); return 0; } *Error: vcpu_next_update: FATAL: p->current not NULL! (d32768v0, runstate running)* How can I can make xenalyze to continuously read from xentrace output file ? _______________________________________________ Xen-users mailing list Xen-users@lists.xen.org http://lists.xen.org/xen-users
George Dunlap
2012-Dec-10 14:19 UTC
Re: how to make xenalyze continuously reading from xentrace file ?
On 08/12/12 06:02, M A wrote:> Hi There, > > I''m a student at NYIT, New York and I have project on xen hypervisor. > Currently i''m working on xentrace and xenalyze to monitor the runstate > events. What i''m trying do now is: While i''m capturing by xentrace I > want to display the result on the screen using xenalyze. I tried to use > loop in the main function but I got errors and when I tried to solve > these errors I got more errors. > > int main(int argc, char *argv[]) { > /* Start with warn at stderr. */ > > warn = stderr; > > argp_parse(&parser_def, argc, argv, 0, NULL, NULL); > > if (G.trace_file == NULL) > exit(1); > > if ( (G.fd = open(G.trace_file, O_RDONLY|O_LARGEFILE)) < 0) { > perror("open"); > error(ERR_SYSTEM, NULL); > } else { > struct stat64 s; > fstat64(G.fd, &s); > G.file_size = s.st_size; > } > > if ( (G.mh = mread_init(G.fd)) == NULL ) > perror("mread"); > > if (G.symbol_file != NULL) > parse_symbol_file(G.symbol_file); > > if(opt.dump_all) > warn = stdout; > > for (int i=0;i<=1000;i++){ > init_pcpus(); > > > if(opt.progress) > progress_init(); > > process_records(); > > if(opt.interval_mode) > interval_tail(); > > if(opt.summary) > summary(); > > if(opt.report_pcpu) > report_pcpu(); > sleep(2); > > } > if(opt.progress) > progress_finish(); > > return 0; > } > > > *Error: vcpu_next_update: FATAL: p->current not NULL! (d32768v0, > runstate running)* > > > How can I can make xenalyze to continuously read from xentrace output file ?(Adding xen-devel, since this is definitely a coding question) So I take it what you''re doing is this: 1. Start xentrace: # xentrace -e all /tmp/foo.trace & 2. Running your "looping" xenalyze on it: # xenalyze -s /tmp/foo.trace Is that correct? I don''t know what your exact problem is here, but one problem you''ll run into eventually is that xenalyze expects a certain "finished" file format, but there''s nothing here to synchronize xenalyze reading the file with xentrace writing the file. The result is that you''re bound at some point to read a file of which the end is only half-written. In any case, what you''re doing here is functionally not really that different from just doing it in bash: # while xenalyze -s /tmp/foo.trace && sleep 2 ; true; done -George