[Bug 63300] mod_status lists BusyWorkers IdleWorkers keys twice
https://bz.apache.org/bugzilla/show_bug.cgi?id=63300 --- Comment #2 from Stefan Neufeind --- Would be great to avoid duplicate values here. In the meantime others like collectd have implemented workarounds. But imho duplicate values should be prevented at the source. https://github.com/collectd/collectd/pull/3131 -- You are receiving this mail because: You are the assignee for the bug. - To unsubscribe, e-mail: bugs-unsubscr...@httpd.apache.org For additional commands, e-mail: bugs-h...@httpd.apache.org
[Bug 65158] delivered files corrupted on Linux with architecture ARM64
https://bz.apache.org/bugzilla/show_bug.cgi?id=65158 --- Comment #41 from Yann Ylavic --- > DocumentRoot "/usr/local/httpd/htdocs" > > Require all granted > > > and then copy "test.txt" in the directory "/usr/local/httpd/htdocs". Obviously this needs to remain the cifs mounted directory, so replace "/usr/local/httpd/htdocs" above with whatever that is.. -- You are receiving this mail because: You are the assignee for the bug. - To unsubscribe, e-mail: bugs-unsubscr...@httpd.apache.org For additional commands, e-mail: bugs-h...@httpd.apache.org
[Bug 65158] delivered files corrupted on Linux with architecture ARM64
https://bz.apache.org/bugzilla/show_bug.cgi?id=65158 --- Comment #40 from Yann Ylavic --- The logs from comment 35 seem to be missing many lines, notably the dumpio ones. There must be multiple ErroLog and/or LogLevel directives configured in the whole /usr/local/httpd/conf tree. I see that "/usr/local/httpd/conf/extra/010-std-80.conf" is included somehow, and that a VirtualHost is defined there, and possibly mod_php and other modules are loaded from there too. Let's try to avoid external configuration files Include(d) from "myhttpd.conf" now, and have a simple VirtualHost directly in the main configuration file like: Listen 80 ServerName localhost LogLevel trace7 DumpIOInput on DumpIOoutput on DocumentRoot "/usr/local/httpd/htdocs" Require all granted and then copy "test.txt" in the directory "/usr/local/httpd/htdocs". The loaded modules could be limited to something like: LoadModule unixd_module modules/mod_unixd.so LoadModule mpm_event_module modules/mod_mpm_event.so LoadModule authn_core_module modules/mod_authn_core.so LoadModule authz_core_module modules/mod_authz_core.so LoadModule log_config_module modules/mod_log_config.so LoadModule dumpio_module modules/mod_dumpio.so That is, a minimal configuration.. -- You are receiving this mail because: You are the assignee for the bug. - To unsubscribe, e-mail: bugs-unsubscr...@httpd.apache.org For additional commands, e-mail: bugs-h...@httpd.apache.org
[Bug 65158] delivered files corrupted on Linux with architecture ARM64
https://bz.apache.org/bugzilla/show_bug.cgi?id=65158 --- Comment #39 from hp4everyth...@googlemail.com --- ok, I checked it: same result -- You are receiving this mail because: You are the assignee for the bug. - To unsubscribe, e-mail: bugs-unsubscr...@httpd.apache.org For additional commands, e-mail: bugs-h...@httpd.apache.org
[Bug 65158] delivered files corrupted on Linux with architecture ARM64
https://bz.apache.org/bugzilla/show_bug.cgi?id=65158 --- Comment #38 from hp4everyth...@googlemail.com --- Created attachment 37764 --> https://bz.apache.org/bugzilla/attachment.cgi?id=37764=edit with corrected ServerRoot -- You are receiving this mail because: You are the assignee for the bug. - To unsubscribe, e-mail: bugs-unsubscr...@httpd.apache.org For additional commands, e-mail: bugs-h...@httpd.apache.org
[Bug 65158] delivered files corrupted on Linux with architecture ARM64
https://bz.apache.org/bugzilla/show_bug.cgi?id=65158 --- Comment #37 from Joe Orton --- It looks like uou are mixing modules compiled from the distribution and self-compiled in these traces: execve("/usr/local/httpd/bin/httpd", ["/usr/local/httpd/bin/httpd", "-X", "-f", "/usr/local/httpd/conf/myhttpd.co"...], 0xe4f49e88 /* 23 vars */) = 0 openat(AT_FDCWD, "/etc/httpd/modules/mod_authn_file.so", O_RDONLY|O_CLOEXEC) = 4 which is certainly not a good idea, though I'm not sure it can cause the symptoms described. Please pass "-d /usr/local/httpd" to httpd as well to avoid this, and/or adjust the config file used. -- You are receiving this mail because: You are the assignee for the bug. - To unsubscribe, e-mail: bugs-unsubscr...@httpd.apache.org For additional commands, e-mail: bugs-h...@httpd.apache.org
[Bug 65158] delivered files corrupted on Linux with architecture ARM64
https://bz.apache.org/bugzilla/show_bug.cgi?id=65158 --- Comment #36 from hp4everyth...@googlemail.com --- I've added the dumpio on commands. When reading the apache-docs I've seen that the dumpio-module should be configured with "dumpio:trace7". Is this the same as "LogLevel trace7" ? Here the terminal logs with the new timestamps: --- strace -o strace.out /usr/local/httpd/bin/httpd -X -f /usr/local/httpd/conf/myhttpd.conf [Mon Mar 08 17:30:40.293244 2021] [core:trace3] [pid 446273] core.c(3361): Setting LogLevel for all modules to trace7 ^C # /usr/local/httpd/bin/httpd -X -f /usr/local/httpd/conf/myhttpd.conf [Mon Mar 08 17:31:28.929678 2021] [core:trace3] [pid 446297] core.c(3361): Setting LogLevel for all modules to trace7 --- wget http://192.168.85.17/haushalt/test.txt --2021-03-08 17:30:52-- http://192.168.85.17/haushalt/test.txt Verbindungsaufbau zu 192.168.85.17:80 … verbunden. HTTP-Anforderung gesendet, auf Antwort wird gewartet … 200 OK Länge: 17 [text/plain] Wird in »test.txt.15« gespeichert. test.txt.15 100%[>] 17 --.-KB/sin 0s 2021-03-08 17:30:52 (706 KB/s) - »test.txt.15« gespeichert [17/17] # wget http://192.168.85.17/haushalt/test.txt --2021-03-08 17:31:42-- http://192.168.85.17/haushalt/test.txt Verbindungsaufbau zu 192.168.85.17:80 … verbunden. HTTP-Anforderung gesendet, auf Antwort wird gewartet … 200 Keine Header, vermutlich ist es HTTP/0.9. Länge: nicht spezifiziert Wird in »test.txt.16« gespeichert. test.txt.16 [ <=>] 324 --.-KB/sin 4,8s 2021-03-08 17:31:47 (68,1 B/s) - »test.txt.16« gespeichert [324] -- You are receiving this mail because: You are the assignee for the bug. - To unsubscribe, e-mail: bugs-unsubscr...@httpd.apache.org For additional commands, e-mail: bugs-h...@httpd.apache.org
[Bug 65158] delivered files corrupted on Linux with architecture ARM64
https://bz.apache.org/bugzilla/show_bug.cgi?id=65158 --- Comment #35 from hp4everyth...@googlemail.com --- Created attachment 37763 --> https://bz.apache.org/bugzilla/attachment.cgi?id=37763=edit correction with dumpio on -- You are receiving this mail because: You are the assignee for the bug. - To unsubscribe, e-mail: bugs-unsubscr...@httpd.apache.org For additional commands, e-mail: bugs-h...@httpd.apache.org
[Bug 65158] delivered files corrupted on Linux with architecture ARM64
https://bz.apache.org/bugzilla/show_bug.cgi?id=65158 --- Comment #34 from hp4everyth...@googlemail.com --- sorry, please forget the last results, probably I only loaded the dumpio-module, but forgot to activate it. I'll come back with new test results. -- You are receiving this mail because: You are the assignee for the bug. - To unsubscribe, e-mail: bugs-unsubscr...@httpd.apache.org For additional commands, e-mail: bugs-h...@httpd.apache.org
[Bug 65158] delivered files corrupted on Linux with architecture ARM64
https://bz.apache.org/bugzilla/show_bug.cgi?id=65158 --- Comment #33 from hp4everyth...@googlemail.com --- Sorry for changing the test-file. In the meantime I found that probably every type of file seems to be corrupted, even small text files. I'll try to keep the "test-environment" fixed in the future. Here the results corresponding to your test-sequence: ### first terminal vim /usr/local/httpd/conf/myhttpd.conf # grep LogLev /usr/local/httpd/conf/myhttpd.conf # LogLevel: Control the number of messages logged to the error_log. #LogLevel warn LogLevel trace7 # grep dumpio /usr/local/httpd/conf/myhttpd.conf LoadModule dumpio_module modules/mod_dumpio.so # strace -o strace.out /usr/local/httpd/bin/httpd -X -f /usr/local/httpd/conf/myhttpd.conf [Mon Mar 08 17:01:40.402877 2021] [core:trace3] [pid 445607] core.c(3361): Setting LogLevel for all modules to trace7 ^C ### second terminal # wget http://192.168.85.17/haushalt/test.txt --2021-03-08 17:01:57-- http://192.168.85.17/haushalt/test.txt Verbindungsaufbau zu 192.168.85.17:80 … verbunden. HTTP-Anforderung gesendet, auf Antwort wird gewartet … 200 OK Länge: 17 [text/plain] Wird in »test.txt.13« gespeichert. test.txt.13 100%[>] 17 --.-KB/sin 0s 2021-03-08 17:01:57 (952 KB/s) - »test.txt.13« gespeichert [17/17] ### ### first terminal # /usr/local/httpd/bin/httpd -X -f /usr/local/httpd/conf/myhttpd.conf [Mon Mar 08 17:02:23.622230 2021] [core:trace3] [pid 445627] core.c(3361): Setting LogLevel for all modules to trace7 ### second terminal # wget http://192.168.85.17/haushalt/test.txt --2021-03-08 17:02:35-- http://192.168.85.17/haushalt/test.txt Verbindungsaufbau zu 192.168.85.17:80 … verbunden. HTTP-Anforderung gesendet, auf Antwort wird gewartet … 200 Keine Header, vermutlich ist es HTTP/0.9. Länge: nicht spezifiziert Wird in »test.txt.14« gespeichert. test.txt.14 [ <=>] 324 --.-KB/sin 4,8s 2021-03-08 17:02:40 (68,1 B/s) - »test.txt.14« gespeichert [324] -- You are receiving this mail because: You are the assignee for the bug. - To unsubscribe, e-mail: bugs-unsubscr...@httpd.apache.org For additional commands, e-mail: bugs-h...@httpd.apache.org
[Bug 65158] delivered files corrupted on Linux with architecture ARM64
https://bz.apache.org/bugzilla/show_bug.cgi?id=65158 --- Comment #32 from hp4everyth...@googlemail.com --- Created attachment 37762 --> https://bz.apache.org/bugzilla/attachment.cgi?id=37762=edit strace.out and error_log unpack with "tar -xvzf" -- You are receiving this mail because: You are the assignee for the bug. - To unsubscribe, e-mail: bugs-unsubscr...@httpd.apache.org For additional commands, e-mail: bugs-h...@httpd.apache.org
[Bug 65169] Add escaped cert env variables
https://bz.apache.org/bugzilla/show_bug.cgi?id=65169 --- Comment #10 from Michael Osipov --- (In reply to Joe Orton from comment #9) > (In reply to Michael Osipov from comment #8) > > (In reply to Michael Osipov from comment #6) > > > (In reply to Joe Orton from comment #5) > > > > Adding SSL_{SERVER,CLIENT}_B64CERT variables (or whatever the name) is > > > > trivial, I would not want to add a config option for this definitely. > > > > > > One thing needs to be considered since +ExportCertData would provide both > > > it > > > would consume the double amount of memory even if the admin needs only one > > > format. What us your opinion on it? > > > > Having a Base64 version of the export option makes the system consistent. > > For the same reason I would export this chain too, this isn't much Tomcat > > specific, but for all upstream servers which want to inspect certs. > > Makes sense. > > > Why do you want to use Base64 URL instead of plain Base64? Other headers > > work with that too and Tomcat would need to add support first. > > I've a mild preference for base64url since it would allow safe use in more > contexts, but plain base64 would work too if that's difficult for Tomcat. I'd prefer plain Base64 for the sake of simplicity. -- You are receiving this mail because: You are the assignee for the bug. - To unsubscribe, e-mail: bugs-unsubscr...@httpd.apache.org For additional commands, e-mail: bugs-h...@httpd.apache.org
[Bug 65169] Add escaped cert env variables
https://bz.apache.org/bugzilla/show_bug.cgi?id=65169 --- Comment #9 from Joe Orton --- (In reply to Michael Osipov from comment #8) > (In reply to Michael Osipov from comment #6) > > (In reply to Joe Orton from comment #5) > > > Adding SSL_{SERVER,CLIENT}_B64CERT variables (or whatever the name) is > > > trivial, I would not want to add a config option for this definitely. > > > > One thing needs to be considered since +ExportCertData would provide both it > > would consume the double amount of memory even if the admin needs only one > > format. What us your opinion on it? > > Having a Base64 version of the export option makes the system consistent. > For the same reason I would export this chain too, this isn't much Tomcat > specific, but for all upstream servers which want to inspect certs. Makes sense. > Why do you want to use Base64 URL instead of plain Base64? Other headers > work with that too and Tomcat would need to add support first. I've a mild preference for base64url since it would allow safe use in more contexts, but plain base64 would work too if that's difficult for Tomcat. -- You are receiving this mail because: You are the assignee for the bug. - To unsubscribe, e-mail: bugs-unsubscr...@httpd.apache.org For additional commands, e-mail: bugs-h...@httpd.apache.org
[Bug 65169] Add escaped cert env variables
https://bz.apache.org/bugzilla/show_bug.cgi?id=65169 --- Comment #8 from Michael Osipov --- (In reply to Michael Osipov from comment #6) > (In reply to Joe Orton from comment #5) > > Adding SSL_{SERVER,CLIENT}_B64CERT variables (or whatever the name) is > > trivial, I would not want to add a config option for this definitely. > > One thing needs to be considered since +ExportCertData would provide both it > would consume the double amount of memory even if the admin needs only one > format. What us your opinion on it? Having a Base64 version of the export option makes the system consistent. For the same reason I would export this chain too, this isn't much Tomcat specific, but for all upstream servers which want to inspect certs. Why do you want to use Base64 URL instead of plain Base64? Other headers work with that too and Tomcat would need to add support first. -- You are receiving this mail because: You are the assignee for the bug. - To unsubscribe, e-mail: bugs-unsubscr...@httpd.apache.org For additional commands, e-mail: bugs-h...@httpd.apache.org
[Bug 65169] Add escaped cert env variables
https://bz.apache.org/bugzilla/show_bug.cgi?id=65169 --- Comment #7 from Joe Orton --- (In reply to Michael Osipov from comment #6) > (In reply to Joe Orton from comment #5) > > Adding SSL_{SERVER,CLIENT}_B64CERT variables (or whatever the name) is > > trivial, I would not want to add a config option for this definitely. > > One thing needs to be considered since +ExportCertData would provide both it > would consume the double amount of memory even if the admin needs only one > format. What us your opinion on it? I don't know what the usually recommended configuration is here but it shouldn't use ExportCertData exactly because it's so expensive. Better to use one of the configs using e.g. mod_headers which just extracts the variables which are required. If really required, we could add a +ExportBase64CertData or similar which just does the base64url(der), that is not as painful as adding a new config options at least. BTW, does Tomcat really want/need the CLIENT_CERT_CHAIN_n as well here? Proof of concept here - https://github.com/apache/httpd/pull/177 -- You are receiving this mail because: You are the assignee for the bug. - To unsubscribe, e-mail: bugs-unsubscr...@httpd.apache.org For additional commands, e-mail: bugs-h...@httpd.apache.org
[Bug 65169] Add escaped cert env variables
https://bz.apache.org/bugzilla/show_bug.cgi?id=65169 --- Comment #6 from Michael Osipov --- (In reply to Joe Orton from comment #5) > Adding SSL_{SERVER,CLIENT}_B64CERT variables (or whatever the name) is > trivial, I would not want to add a config option for this definitely. One thing needs to be considered since +ExportCertData would provide both it would consume the double amount of memory even if the admin needs only one format. What us your opinion on it? -- You are receiving this mail because: You are the assignee for the bug. - To unsubscribe, e-mail: bugs-unsubscr...@httpd.apache.org For additional commands, e-mail: bugs-h...@httpd.apache.org
[Bug 65169] Add escaped cert env variables
https://bz.apache.org/bugzilla/show_bug.cgi?id=65169 --- Comment #5 from Joe Orton --- Adding SSL_{SERVER,CLIENT}_B64CERT variables (or whatever the name) is trivial, I would not want to add a config option for this definitely. -- You are receiving this mail because: You are the assignee for the bug. - To unsubscribe, e-mail: bugs-unsubscr...@httpd.apache.org For additional commands, e-mail: bugs-h...@httpd.apache.org
[Bug 65169] Add escaped cert env variables
https://bz.apache.org/bugzilla/show_bug.cgi?id=65169 --- Comment #4 from Michael Osipov --- (In reply to Mark Thomas from comment #3) > Remove myself from CC. I intentionally added you to discuss how tomcat should handle it. -- You are receiving this mail because: You are the assignee for the bug. - To unsubscribe, e-mail: bugs-unsubscr...@httpd.apache.org For additional commands, e-mail: bugs-h...@httpd.apache.org
[Bug 65158] delivered files corrupted on Linux with architecture ARM64
https://bz.apache.org/bugzilla/show_bug.cgi?id=65158 --- Comment #31 from Yann Ylavic --- Yes the -X option is to keep httpd as a single process attached to the terminal, so it will start and "hang" until you stop it with ^C (and not with "-k stop" like when it's started with "-k start"). That's why I talked about starting httpd and wget in two separate terminals, leaving httpd alone ("hanging") in one terminal until you stop it. I'm a bit confused here too. This PR started with jpeg/pdf files (which I read as large files) corrupted with EnableMMap on. Now a tiny test.txt file of 17 bytes is also corrupted with EnableMMap on (only?) unless httpd is started by strace, right? Let's restart the test proceduire if you wish. 1. Configure httpd 1.a. Stop any running httpd: "httpd ... -k stop" and ^C on any terminal where you ran "strace ... httpd -X". There shouldn't be any httpd running as shown by "ps -ef" or alike. 1.b. In /usr/local/httpd/conf/httpd.conf: - modify LogLevel to trace7 - add/uncomment "LoadModule dumpio_module modules/mod_dumpio.so" - add "DumpIOInput on" and "DumpIOoutput on" (note, if mod_dumpio.so is not in the /usr/local/httpd/modules directory, run "./configure ... --enable-dumpio=shared" in the sources directory and "sudo make install" again). 1.c. rm /usr/local/httpd/logs/error_log 2. Test with strace 2.a. In one terminal (terminal 1), run: $ sudo strace -o strace.out /usr/local/httpd/bin/httpd -X -f /usr/local/httpd/conf/httpd.conf 2.b. In another terminal (terminal 2), run: $ wget http://192.168.85.17/haushalt/test.txt 2.c. In terminal 1, hit ^C stop stop httpd 2.d. Is the reponse correct/corrupted? 3. Test without strace 3.a. In one terminal (terminal 1), run: $ sudo /usr/local/httpd/bin/httpd -X -f /usr/local/httpd/conf/httpd.conf 3.b. In another terminal (terminal 2), run: $ wget http://192.168.85.17/haushalt/test.txt 3.c. In terminal 1, hit ^C stop stop httpd 3.d. Is the reponse correct/corrupted? 4. Attach strace.out and /usr/local/httpd/logs/error_log (.tar.gz) to this PR -- You are receiving this mail because: You are the assignee for the bug. - To unsubscribe, e-mail: bugs-unsubscr...@httpd.apache.org For additional commands, e-mail: bugs-h...@httpd.apache.org
[Bug 65169] Add escaped cert env variables
https://bz.apache.org/bugzilla/show_bug.cgi?id=65169 Mark Thomas changed: What|Removed |Added CC|ma...@apache.org| --- Comment #3 from Mark Thomas --- Remove myself from CC. -- You are receiving this mail because: You are the assignee for the bug. - To unsubscribe, e-mail: bugs-unsubscr...@httpd.apache.org For additional commands, e-mail: bugs-h...@httpd.apache.org
[Bug 65169] Add escaped cert env variables
https://bz.apache.org/bugzilla/show_bug.cgi?id=65169 Michael Osipov changed: What|Removed |Added CC||ma...@apache.org --- Comment #2 from Michael Osipov --- (In reply to Joe Orton from comment #1) > The obvious way to do this would be to use the base64url encoding of the > DER, rather than URL-encoding the PEM, which seems kind of daft. Can Tomcat > decode the former? Not without additional code. would you prefer to introduce new variables or rather a format co fig option? -- You are receiving this mail because: You are the assignee for the bug. - To unsubscribe, e-mail: bugs-unsubscr...@httpd.apache.org For additional commands, e-mail: bugs-h...@httpd.apache.org
[Bug 65158] delivered files corrupted on Linux with architecture ARM64
https://bz.apache.org/bugzilla/show_bug.cgi?id=65158 --- Comment #30 from hp4everyth...@googlemail.com --- to be precise: the start-command with -X hangs, but the httpd-server is running and produces the error as opposed to the strace version. -- You are receiving this mail because: You are the assignee for the bug. - To unsubscribe, e-mail: bugs-unsubscr...@httpd.apache.org For additional commands, e-mail: bugs-h...@httpd.apache.org
[Bug 65158] delivered files corrupted on Linux with architecture ARM64
https://bz.apache.org/bugzilla/show_bug.cgi?id=65158 --- Comment #29 from hp4everyth...@googlemail.com --- I've made an additional try with debug option -X but withut strace. /usr/local/httpd/bin/httpd -X -f /usr/local/httpd/conf/myhttpd.conf -k start That doesn't work, because the start command hangs till Ctrl-C Maybe -X is not supposed to work without strace? -- You are receiving this mail because: You are the assignee for the bug. - To unsubscribe, e-mail: bugs-unsubscr...@httpd.apache.org For additional commands, e-mail: bugs-h...@httpd.apache.org
[Bug 65158] delivered files corrupted on Linux with architecture ARM64
https://bz.apache.org/bugzilla/show_bug.cgi?id=65158 --- Comment #28 from hp4everyth...@googlemail.com --- does strace modify or influence the httpd-execution, e.g. by shifting the executable in memory to page boundaries? I hope, it's not a simple mistake on my side -- You are receiving this mail because: You are the assignee for the bug. - To unsubscribe, e-mail: bugs-unsubscr...@httpd.apache.org For additional commands, e-mail: bugs-h...@httpd.apache.org
[Bug 65158] delivered files corrupted on Linux with architecture ARM64
https://bz.apache.org/bugzilla/show_bug.cgi?id=65158 --- Comment #27 from hp4everyth...@googlemail.com --- that's my command history, I hope I didn't make a mistake... between 487 and 488 I started a wget from an other terminal, also after 491 486 strace -o strace.out /usr/local/httpd/bin/httpd -X -f /usr/local/httpd/conf/myhttpd.conf 487 /usr/local/httpd/bin/httpd -f /usr/local/httpd/conf/myhttpd.conf -k start 488 /usr/local/httpd/bin/httpd -f /usr/local/httpd/conf/myhttpd.conf -k stop 489 ls strace.out 490 rm strace.out 491 strace -o strace.out /usr/local/httpd/bin/httpd -X -f /usr/local/httpd/conf/myhttpd.conf 492 vi strace.out 493 ls -la strace.out 494 tar -cvzf strace.out.tar.gz 495 tar -cvzf strace.out.tar.gz strace.out -- You are receiving this mail because: You are the assignee for the bug. - To unsubscribe, e-mail: bugs-unsubscr...@httpd.apache.org For additional commands, e-mail: bugs-h...@httpd.apache.org
[Bug 65169] Add escaped cert env variables
https://bz.apache.org/bugzilla/show_bug.cgi?id=65169 --- Comment #1 from Joe Orton --- The obvious way to do this would be to use the base64url encoding of the DER, rather than URL-encoding the PEM, which seems kind of daft. Can Tomcat decode the former? -- You are receiving this mail because: You are the assignee for the bug. - To unsubscribe, e-mail: bugs-unsubscr...@httpd.apache.org For additional commands, e-mail: bugs-h...@httpd.apache.org
[Bug 65158] delivered files corrupted on Linux with architecture ARM64
https://bz.apache.org/bugzilla/show_bug.cgi?id=65158 --- Comment #26 from Yann Ylavic --- > I'm confused with the result: > > starting httpd without strace => corrupted file transfer and > HTTP-protocol-error > > starting httpd with strace => file transfer ok, HTTP-protocol ok Do you start httpd with the same command line, i.e. without strace: $ sudo /usr/local/httpd/bin/httpd -X -f /usr/local/httpd/conf/myhttpd.conf and with strace: $ sudo strace -o strace.out /usr/local/httpd/bin/httpd -X -f /usr/local/httpd/conf/myhttpd.conf ? -- You are receiving this mail because: You are the assignee for the bug. - To unsubscribe, e-mail: bugs-unsubscr...@httpd.apache.org For additional commands, e-mail: bugs-h...@httpd.apache.org
[Bug 65158] delivered files corrupted on Linux with architecture ARM64
https://bz.apache.org/bugzilla/show_bug.cgi?id=65158 --- Comment #25 from hp4everyth...@googlemail.com --- please note: the corrupted transfer takes several seconds, maybe the HTTP-response is pending or terminated by timeout? -- You are receiving this mail because: You are the assignee for the bug. - To unsubscribe, e-mail: bugs-unsubscr...@httpd.apache.org For additional commands, e-mail: bugs-h...@httpd.apache.org
[Bug 65158] delivered files corrupted on Linux with architecture ARM64
https://bz.apache.org/bugzilla/show_bug.cgi?id=65158 --- Comment #24 from hp4everyth...@googlemail.com --- I'm confused with the result: starting httpd without strace => corrupted file transfer and HTTP-protocol-error starting httpd with strace => file transfer ok, HTTP-protocol ok I did this 3 times with the same result: - # wget http://192.168.85.17/haushalt/test.txt --2021-03-08 11:32:33-- http://192.168.85.17/haushalt/test.txt Verbindungsaufbau zu 192.168.85.17:80 … verbunden. (translat: connected) HTTP-Anforderung gesendet, auf Antwort wird gewartet … 200 Keine Header, vermutlich ist es HTTP/0.9. (translat: 200 no header, probably HTTP/0.9) Länge: nicht spezifiziert (translat: length not specified ) Wird in »test.txt.1« gespeichert. test.txt.1 [ <=>] 324 --.-KB/sin 4,8s 2021-03-08 11:32:38 (68,1 B/s) - »test.txt.1« gespeichert [324] # wget http://192.168.85.17/haushalt/test.txt --2021-03-08 11:33:30-- http://192.168.85.17/haushalt/test.txt Verbindungsaufbau zu 192.168.85.17:80 … verbunden. HTTP-Anforderung gesendet, auf Antwort wird gewartet … 200 OK Länge: 17 [text/plain] (translat: length 17, which is ok ) Wird in »test.txt.2« gespeichert. test.txt.2 100%[>] 17 --.-KB/sin 0s -- You are receiving this mail because: You are the assignee for the bug. - To unsubscribe, e-mail: bugs-unsubscr...@httpd.apache.org For additional commands, e-mail: bugs-h...@httpd.apache.org
[Bug 65158] delivered files corrupted on Linux with architecture ARM64
https://bz.apache.org/bugzilla/show_bug.cgi?id=65158 --- Comment #23 from hp4everyth...@googlemail.com --- Created attachment 37761 --> https://bz.apache.org/bugzilla/attachment.cgi?id=37761=edit strace output please extract with "tar -xvzf strace.out.tar.gz" -- You are receiving this mail because: You are the assignee for the bug. - To unsubscribe, e-mail: bugs-unsubscr...@httpd.apache.org For additional commands, e-mail: bugs-h...@httpd.apache.org
[Bug 65158] delivered files corrupted on Linux with architecture ARM64
https://bz.apache.org/bugzilla/show_bug.cgi?id=65158 --- Comment #22 from Yann Ylavic --- Thanks. Please now start httpd with: $ sudo strace -o strace.out /usr/local/httpd/bin/httpd -X -f /usr/local/httpd/conf/myhttpd.conf (this will keep httpd attached to the terminal, with ^C to stop it) Then run wget (from another terminal/machine) and attach the "strace.out" file here. > Since this was not necessary with my previous build without your > buildconf-command and also is not necessary in the archlinux-distributed > apache, I'd like to know: > > - does buildconf lead to a configuration with preloaded or static > mpm-Module, which is not prefork? You can use: $ ./configure --with-mpm=prefork ... to select the default MPM at build time. -- You are receiving this mail because: You are the assignee for the bug. - To unsubscribe, e-mail: bugs-unsubscr...@httpd.apache.org For additional commands, e-mail: bugs-h...@httpd.apache.org
[Bug 64135] OCSP Stapling doesn't handle Keep-Alive responses properly, causing delays/timeouts
https://bz.apache.org/bugzilla/show_bug.cgi?id=64135 --- Comment #8 from Ruediger Pluem --- (In reply to Christophe JAILLET from comment #7) > Coming back to it, is this "Connection: close" needed at all? > We are using HTTP/1.0 here, and it should be closed in all cases. > > Did I miss something? Is there something else lurking around? It is probably caused by a misconfigured / buggy HTTP server on OCSP server side, but I see no harm in adding the "Connection: close" if it helps. I agree that the normal behavior of the server to a HTTP/1.0 request should be to close the connection after sending the response. -- You are receiving this mail because: You are the assignee for the bug. - To unsubscribe, e-mail: bugs-unsubscr...@httpd.apache.org For additional commands, e-mail: bugs-h...@httpd.apache.org