Re: Cygwin instabilities
2010/9/13 Al wanted stability stories: I run Cygwin on a double core machine with Vista. I estimate that 1 of 20 of my compilations break and need to be repeated. One weak point is the compression of man pages. I routinely run my perl compiler smokes on cygwin and linux machines. This involves automatic updates, failures which cause core dumps, 100%CPU or eating all available memory. cygwin (better Windows) is more stable than linux in this regard. I should really tune my ulimits on my linux but out of the box my linux box becomes unusable at certain tests. With cygwin not. Cygwin is just about 3x slower, and processes keep locking certain files. Why not. Others report that they don't use Cygwin because of instablilities, especially in the server context. Okay. While I'm running my nightly destructive smokes I also serve content for various projects. My linux laptop started with issues (heat or SW?) during night while people were up- and downloading via sftp, so I fell back to my cygwin box, and since then I kept running this cygwin server instead. Much more stable than my linux box so far. Of course that's most likely a HW issue, but cygwin was stable enough and better. For example I run vlc sessions (publicly broadcasting football games from brazil here in my town) with the cygwin perl testing sessions and cygwin stfp uploading in the background, and linux is not really as fast and stable in vlc client response as I with my windows box. What are the reasons? Will this be better with Windows 7? Can Cygwin become server stable? Windows 7 was not better for me. More system DLL's and footprint, much more rebase problems. Sometimes I can only stop MSIE and MS Outlook to continue to work in my mintty shells. Server stable in ISP terms of course not. It's still just Windows, with all its known weaknesses. But ISP's are still selling and using windows servers. -- Reini Urban http://phpwiki.org/ http://murbreak.at/ -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Re: Cygwin instabilities
What are the reasons? Will this be better with Windows 7? Can Cygwin become server stable? Windows 7 was not better for me. More system DLL's and footprint, much more rebase problems. Sometimes I can only stop MSIE and MS Outlook to continue to work in my mintty shells. :-( I always thought Vista was the ugly prototype and Windows 7 would become the lean new system, that brings back the fun. No, that brings the fun. There was never much fun in Windows, although I personally like the mere surface of Vista. Server stable in ISP terms of course not. It's still just Windows, with all its known weaknesses. But ISP's are still selling and using windows servers. Yes, in terms of ISP. It's not the Desktop users complaints. As a developer I can react immediatly, if something goes wrong. The admins can't. http://en.gentoo-wiki.com/wiki/Prefix/Cygwin I have done much progress with the evaluation of the Prefix bootstrapping process meanwhile. That was 4 weeks work to solve all issues. In a few days I will put that all into one big script. Then I will see, if the whole process will go through over night or where instabilities will occur. In that case I will be able to report more details of instabilities. Al -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Re: Cygwin instabilities
You'll need to be more specific about the issues you're encountering and your configuration. See http://cygwin.com/problems.html. The majority of instabilities I observerved while compressing manpages with bzip2 -9, a minority during execessiv link operations. Some others I can't relate to something. Since I have compiled bzip2 from the Gentoo sources, I haven't seen any instablilities from this side any more, but a also was more conservative and didn't do parallel builds. Al -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Cygwin instabilities
Hello, I run Cygwin on a double core machine with Vista. I estimate that 1 of 20 of my compilations break and need to be repeated. One weak point is the compression of man pages. Others report that they don't use Cygwin because of instablilities, especially in the server context. What are the reasons? Will this be better with Windows 7? Can Cygwin become server stable? Al -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Re: Cygwin instabilities
On 9/13/2010 9:19 AM, Al wrote: Hello, I run Cygwin on a double core machine with Vista. I estimate that 1 of 20 of my compilations break and need to be repeated. One weak point is the compression of man pages. Others report that they don't use Cygwin because of instablilities, especially in the server context. What are the reasons? Will this be better with Windows 7? Can Cygwin become server stable? You'll need to be more specific about the issues you're encountering and your configuration. See http://cygwin.com/problems.html. -- Larry Hall http://www.rfk.com RFK Partners, Inc. (508) 893-9779 - RFK Office 216 Dalton Rd. (508) 893-9889 - FAX Holliston, MA 01746 _ A: Yes. Q: Are you sure? A: Because it reverses the logical flow of conversation. Q: Why is top posting annoying in email? -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Re: Cygwin instabilities
Hello Larry, What are the reasons? Will this be better with Windows 7? Can Cygwin become server stable? You'll need to be more specific about the issues you're encountering and your configuration. See http://cygwin.com/problems.html. I am not asking this to debug my own setup. I am rather ask for an overall estimation of Cygwins current and future usability and stability. I am currently working on an OS product that depends on Cygwin as POSIX layer. I would like to estimate the chances it will have, to be usefull for many people. It not, how to tweak Cygwin to run on some machines. It's, how big is the percentage of windows machines, that will run a stable Cygwin with the standard setup.exe setup. Al -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Re: Cygwin instabilities
Al oss.el...@googlemail.com was heard to say: I am not asking this to debug my own setup. I am rather ask for an overall estimation of Cygwins current and future usability and stability. These two things are related. Remember that Cygwin is an open source project, and that it does not employ dozens of developers with the abstract task of increasing usability or stability. Both are increased by either debugging or at least properly reporting bugs. If you experience stability problems on your setup, then reporting this in all necessary detail is a sure step to increase future usability and stability. It not, how to tweak Cygwin to run on some machines. It's, how big is the percentage of windows machines, that will run a stable Cygwin with the standard setup.exe setup. Who would keep counters of stable or instable setups? These number are exceptionally hard to come by. Even if this list is now flooded with my setup works and mine too posts, these numbers would not be representative. Users may have given up on Cygwin due to instabilities without notifying the list. Others may run Cygwin so happily they never think about joining the list. All you could do is to scan the Cygwin archives for instabilities (this is your term and arguably far too unspecific) and compare it to the number of instabilities reported for any run-of-the mill Linux in the same timeframe. Just my 2cc Markus P.S. mine works -- Markus Hoenicka http://www.mhoenicka.de AQ score 38 -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Re: Cygwin instabilities
On 9/13/10, Markus Hoenicka markus.hoenicka xxx.duh wrote: Al oss.el...@xxx.xxx was heard to say: I am not asking this to debug my own setup. I am rather ask for an overall estimation of Cygwins current and future usability and stability. These two things are related. Remember that Cygwin is an open source project, and that it does not employ dozens of developers with the abstract task of increasing usability or stability. Both are increased by either debugging or at least properly reporting bugs. If you experience stability problems on your setup, then reporting this in all necessary detail is a sure step to increase future usability and stability. It not, how to tweak Cygwin to run on some machines. It's, how big is the percentage of windows machines, that will run a stable Cygwin with the standard setup.exe setup. Who would keep counters of stable or instable setups? These number are exceptionally hard to come by. Even if this list is now flooded with my setup works and mine too posts, these numbers would not be representative. Users may have given up on Cygwin due to instabilities without notifying the list. Others may run Cygwin so happily they never think about joining the list. All you could do is to scan the Cygwin archives for instabilities (this is your term and arguably far too unspecific) and compare it to the number of instabilities reported for any run-of-the mill Linux in the same timeframe. I would mention in response to earlier question that I'm quite happy with cygwin on 'doze 7. I routinely build and test mobile phone apps here as well as miscellaneous script based downloads and everything seems fine. The performance issues I reported are fine compared to any 'doze alternatives and so far most of what I have developed under cygwin runs under real linux. And it is a good way to learn linux too LOL. btw, is is possible to do something like the new iproute2 stuff in dohs? Just my 2cc Markus P.S. mine works -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Re: Cygwin instabilities
These two things are related. Remember that Cygwin is an open source Sure it is related, but that doesn't answer my question. Who would keep counters of stable or instable setups? These number are exceptionally hard to come by. Even if this list is now flooded with my setup works and mine too posts, these numbers would not be No that is not the way to go. I think there are people which run Cygwin on more than 1 machine, so they have a personal estimation and experience. representative. Users may have given up on Cygwin due to instabilities without notifying the list. Others may run Cygwin so happily they never Right. Similar I can't report the bugs for people telling me, they don't use Cygwin, because of stability issues they encountered in the past. Al -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Re: Cygwin instabilities
On 9/13/2010 11:57 AM, Al wrote: These two things are related. Remember that Cygwin is an open source Sure it is related, but that doesn't answer my question. Who would keep counters of stable or instable setups? These number are exceptionally hard to come by. Even if this list is now flooded with my setup works and mine too posts, these numbers would not be No that is not the way to go. I think there are people which run Cygwin on more than 1 machine, so they have a personal estimation and experience. representative. Users may have given up on Cygwin due to instabilities without notifying the list. Others may run Cygwin so happily they never Right. Similar I can't report the bugs for people telling me, they don't use Cygwin, because of stability issues they encountered in the past. OK. Well without any specifics about a particular issue and/or point in time, there's really no way for anyone on this list to respond to this type of query. If someone finds something, the best way to handle the problem is to report it here following the guidelines of http://cygwin.com/problems.html. -- Larry Hall http://www.rfk.com RFK Partners, Inc. (508) 893-9779 - RFK Office 216 Dalton Rd. (508) 893-9889 - FAX Holliston, MA 01746 _ A: Yes. Q: Are you sure? A: Because it reverses the logical flow of conversation. Q: Why is top posting annoying in email? -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Re: Cygwin instabilities
Al oss.el...@googlemail.com was heard to say: No that is not the way to go. I think there are people which run Cygwin on more than 1 machine, so they have a personal estimation and experience. Yes, and this is where the Cygwin archives come in handy. If there are users running Cygwin on lots of machines, they are likely to run into problems once in a while, and they're likely to use the list for reporting these problems. If Cygwin was as instable as you apparently suspect, you'd find plenty of reports in the archives. I haven't noticed any such large-scale reports about instabilities in several years of lurking on this list, but your thorough perusal of the archives may well prove me wrong. regards, Markus -- Markus Hoenicka http://www.mhoenicka.de AQ score 38 -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Re: Cygwin instabilities
On Mon, Sep 13, 2010 at 05:57:08PM +0200, Al wrote: These two things are related. Remember that Cygwin is an open source Sure it is related, but that doesn't answer my question. Who would keep counters of stable or instable setups? These number are exceptionally hard to come by. Even if this list is now flooded with my setup works and mine too posts, these numbers would not be No that is not the way to go. I think there are people which run Cygwin on more than 1 machine, so they have a personal estimation and experience. representative. Users may have given up on Cygwin due to instabilities without notifying the list. Others may run Cygwin so happily they never Right. Similar I can't report the bugs for people telling me, they don't use Cygwin, because of stability issues they encountered in the past. You seem to be approaching this problem as if people will say Ah! Stability issues! Right. Well, ok, here's what you need to know. If we knew of stability issues they would be fixed. Don't assume that since some nebulous person or persons couldn't get some version of Cygwin running it means that there are well-known problems that we will all clamor to proclaim. If you think that asking a group of strangers for help with a nebulously defined instability is really going to get you any useful responses then you must not be well-acquainted with open source. As a project lead, here's my advice: If you are concerned that Cygwin is unstable then buy support from Red Hat. Then you have a guaranteed recourse if something doesn't work. cgf -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Re: Cygwin instabilities
You seem to be approaching this problem as if people will say Ah! Stability issues! Right. Well, ok, here's what you need to know. It's not me thinking in the bug - solution pattern here. I don't ask for a solution. If you think that asking a group of strangers for help with a nebulously defined instability is really going to get you any useful responses then you must not be well-acquainted with open source. I ask for an estimation. I ask people that work intensively with Cygwin, to have a second opinion. Then I can compare it to the non-cygwin-advocates. So please don't take it the wrong way. I don't want to start a flame. Would be the wrong place anyway. :-) As a project lead, here's my advice: If you are concerned that Cygwin is unstable then buy support from Red Hat. Then you have a guaranteed recourse if something doesn't work. If I would sell closed source product, instead of devolping an OS one. ... but than I could advice the customer to buy the Interix layer instead. Al -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Re: Cygwin instabilities
On Mon, Sep 13, 2010 at 10:19 AM, Al oss.el...@googlemail.com wrote: Others report that they don't use Cygwin because of instablilities, especially in the server context. I've also been having instabilities in my server context (Windows Server 2008 R2), but I have a few more concrete details: I had 1.7.5 installed, and built gcc-3.4.6, gcc-4.2.4, and gcc-4.4.4 to use on a continuous integration test box. All went relatively fine. I updated to 1.7.7 and kept using the toolchains built with 1.7.5. I got that some of the build cycles would hang in make (in gcc.exe or tr.exe actually) with Bad address. The process would still appear in pstree with a * and I had to /usr/bin/kill -W it. Since it seemed to happen randomly, it wasn't easily reproducible, so I don't how or what exactly to debug. Is there any way to try and trap that error? I could also try to log everything with strace but I'm afraid that would produce tons of data (the builds could run for hours before such error occurred). -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Re: Cygwin instabilities
On 13/09/2010 17:35, Christopher Faylor wrote: You seem to be approaching this problem as if people will say Ah! Stability issues! Right. Well, ok, here's what you need to know. If we knew of stability issues they would be fixed. Well, I know of one, but haven't had time to fix it yet, so I keep this hack in my local builds. Can't run make check -jN without it, but even then it sometimes locks up. The problem I've run into is that on a heavily loaded system, a pinfo struct can get truncated into a redirector in between the time a syscall checks the process_state (using ISSTATE or NOTSTATE) and the time it subsequently attempts to access a pinfo member which it hoped to guard by that check. I have a bad feeling that the only way to totally resolve this is going to be adding lots of locking or mutexing around pinfo calls, which is almost bound to have performance implications :-( cheers, DaveK --- src.clean/winsup/cygwin/pinfo.h 2010-09-01 22:06:36.0 +0100 +++ src/winsup/cygwin/pinfo.h 2010-09-06 20:36:17.06250 +0100 @@ -51,8 +51,6 @@ public: DWORD exitcode; /* set when process exits */ -#define PINFO_REDIR_SIZE ((char *) myself.procinfo-exitcode - (char *) myself.procinfo) - /* 0 if started by a cygwin process */ DWORD cygstarted; @@ -64,9 +62,6 @@ public: signals. */ DWORD dwProcessId; - /* Used to spawn a child for fork(), among other things. */ - WCHAR progname[NT_MAX_PATH]; - /* User information. The information is derived from the GetUserName system call, with the name looked up in /etc/passwd and assigned a default value @@ -121,6 +116,12 @@ public: HANDLE wr_proc_pipe; DWORD wr_proc_pipe_owner; friend class pinfo; + + /* Used to spawn a child for fork(), among other things. */ + WCHAR progname[NT_MAX_PATH]; + /* Truncate it after execed process exits. */ +#define PINFO_REDIR_SIZE ((char *) myself.procinfo-progname[0] - (char *) myself.procinfo) + }; DWORD WINAPI commune_process (void *); -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Re: Cygwin instabilities
On Mon, Sep 13, 2010 at 03:52:55PM -0300, Ramiro Polla wrote: On Mon, Sep 13, 2010 at 10:19 AM, Al oss.el...@googlemail.com wrote: Others report that they don't use Cygwin because of instablilities, especially in the server context. I've also been having instabilities in my server context (Windows Server 2008 R2), but I have a few more concrete details: Ok. Two reports of instabilities. And what does that show? Here are the OP's original questions: What are the reasons? Dave theorized about one. Will this be better with Windows 7? No. Can Cygwin become server stable? It depends on what is meant by become. If it means will there be a concerted effort to harden Cygwin for a server then the answer is likely not unless someone pays for it. That points back to paying Red Hat for support. cgf -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Re: Cygwin instabilities
On Mon, Sep 13, 2010 at 08:34:42PM +0100, Dave Korn wrote: On 13/09/2010 17:35, Christopher Faylor wrote: You seem to be approaching this problem as if people will say Ah! Stability issues! Right. Well, ok, here's what you need to know. If we knew of stability issues they would be fixed. Well, I know of one, but haven't had time to fix it yet, so I keep this hack in my local builds. Can't run make check -jN without it, but even then it sometimes locks up. The problem I've run into is that on a heavily loaded system, a pinfo struct can get truncated into a redirector in between the time a syscall checks the process_state (using ISSTATE or NOTSTATE) and the time it subsequently attempts to access a pinfo member which it hoped to guard by that check. I have a bad feeling that the only way to totally resolve this is going to be adding lots of locking or mutexing around pinfo calls, which is almost bound to have performance implications :-( You've included a patch here but, FYI, I have no idea why. cgf -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Re: Cygwin instabilities
On 9/13/2010 12:34 PM, Dave Korn wrote: Even when I submit make check-c make check-fortran and make check-c++ in separate windows, occasionally an instance of expect will hang and fail to time out, so has to be killed to complete the check. More often than not, it's possible to complete the whole 3 day series without such a hang. Anyway, it's not specific to -jN. -- Tim Prince -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Re: Cygwin instabilities
On 13/09/2010 20:23, Christopher Faylor wrote: You've included a patch here but, FYI, I have no idea why. FTR only. It might be useful to others following the list. cheers, DaveK -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Re: Cygwin instabilities
It depends on what is meant by become. If it means will there be a concerted effort to harden Cygwin for a server then the answer is likely not unless someone pays for it. That points back to paying Red Hat for support. I think you use the term support in the wrong field. It is the enduser who buys support, not the developers. Also the enduser doesn't buy support to make a product more stable. It's works the other way. A stable product is choosen by more endusers, which in return buy more support. So it's Red Hat, who has to invest into stability of Cygwin to make it as succesfull as possible. If they don't see a challange in this, than Cygwin is in a kind of cul-de-sac with Red Hat I guess. Al -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Re: Cygwin instabilities
On Mon, Sep 13, 2010 at 09:48:49PM +0200, Al wrote: It depends on what is meant by become. ?If it means will there be a concerted effort to harden Cygwin for a server then the answer is likely not unless someone pays for it. That points back to paying Red Hat for support. I think you use the term support in the wrong field. It is the enduser who buys support, not the developers. Can't really parse that, especially given that I'm a developer and you're obviously not. Also the enduser doesn't buy support to make a product more stable. My supposition was that you would purchase the product from Red Hat, run it, report problems, and, gradually, have Red Hat improve stability. However, you could also purchase a contract with Red Hat with the goal of improving server stability. I guess the latter is not strictly a support contract but I don't see why there has to be minute hair splitting here. It's works the other way. A stable product is choosen by more endusers, which in return buy more support. So it's Red Hat, who has to invest into stability of Cygwin to make it as succesfull as possible. If they don't see a challange in this, than Cygwin is in a kind of cul-de-sac with Red Hat I guess. Red Hat puts as much money into the product as needed to make a profit. If they are satisfied with their customer base then they have no incentive to do anything to Cygwin. If *you* want to use Cygwin and you have a specific requirement that is not going to be met by the meandering ways of an open source project then you have two options: 1) work on improving the product yourself or 2) pay someone to do it for you. In the Cygwin scenario, the most likely place to find someone to do it for you is Red Hat. cgf -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Re: Cygwin instabilities
Can't really parse that, especially given that I'm a developer and you're obviously not. lol Al -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple