From: Digest To: "OS/2GenAu Digest" Date: Tue, 15 Jul 2003 00:02:00 EST-10EDT,10,-1,0,7200,3,-1,0,7200,3600 Subject: [os2genau_digest] No. 660 Reply-To: X-List-Unsubscribe: www.os2site.com/list/ ************************************************** Monday 14 July 2003 Number 660 ************************************************** Subjects for today 1 Warp Server : Graham Norton" 2 Re: Warp Server : John Angelico" 3 Re: Warp Server : Graham Norton" 4 SWAPPER.DAT : Alan Heiser 5 Re: SWAPPER.DAT : Ed Durrant 6 Re: Warp Server : Paul Smedley" 7 Re: SWAPPER.DAT : John Angelico" 8 Re: Warp Server : Ed Durrant 9 Re: Warp Server : Paul Smedley" 10 Re: Warp Server : bob 11 Re: Warp Server : Paul Smedley" 12 Re: Warp Server : John Angelico" 13 Re: Warp Server : Paul Smedley" 14 Re: SWAPPER.DAT : Mike O'Connor" 15 Re: Warp Server : bob **= Email 1 ==========================** Date: Mon, 14 Jul 2003 14:44:35 +0950 (CST) From: "Graham Norton" Subject: Warp Server Hi all It is with some pride that I discovered today that our Business warp Server file and print machine and a second PC - Internet gateway server have both been up and running without hiccup for 29days 22 hours and 23 minutes!!! and the reason for the server stopping almost a month ago was a violent electrical storm and power failure and the UPS monitor shut things down cleanly! Cheers Graham Norton Neurologist "when I need a hole in the head, I use Windows when I need a window on the brain, I use OS2!" ---------------------------------------------------------------------------------- **= Email 2 ==========================** Date: Mon, 14 Jul 2003 15:42:56 +1000 (AEST) From: "John Angelico" Subject: Re: Warp Server On Mon, 14 Jul 2003 14:44:35 +0950 (CST), Graham Norton wrote: >Hi all > >It is with some pride that I discovered today that our Business warp >Server file and print machine and a second PC - Internet gateway server >have both been up and running without hiccup for > >29days 22 hours and 23 minutes!!! > >and the reason for the server stopping almost a month ago was a violent >electrical storm and power failure and the UPS monitor shut things down >cleanly! > Good on you Graham. I trust that you have the latest available kernel which overcomes the 49 day problem. Yes, sadly folks, OS/2 suffers from a similar problem to the Win-dy systems. Although one wonders how MS ever found it or considered it a real issue... Best regards John Angelico OS/2 SIG os2 at melbpc dot org dot au or talldad at kepl dot com dot au _______________________________________ PMTagline v1.50 - Copyright, 1996-1997, Stephen Berg and John Angelico .... Would you trust a BMW controlled by Windows? ---------------------------------------------------------------------------------- **= Email 3 ==========================** Date: Mon, 14 Jul 2003 15:30:34 +0950 (CST) From: "Graham Norton" Subject: Re: Warp Server Oops! Now you have upset me! What is the 49 day issue!?? On Mon, 14 Jul 2003 15:42:56 +1000 (AEST), John Angelico wrote: >On Mon, 14 Jul 2003 14:44:35 +0950 (CST), Graham Norton wrote: > >>Hi all >> >>It is with some pride that I discovered today that our Business warp >>Server file and print machine and a second PC - Internet gateway server >>have both been up and running without hiccup for >> >>29days 22 hours and 23 minutes!!! >> >>and the reason for the server stopping almost a month ago was a violent >>electrical storm and power failure and the UPS monitor shut things down >>cleanly! >> > >Good on you Graham. > >I trust that you have the latest available kernel which overcomes the 49 >day problem. > >Yes, sadly folks, OS/2 suffers from a similar problem to the Win-dy >systems. Although one wonders how MS ever found it or considered it a real >issue... > > > >Best regards >John Angelico >OS/2 SIG >os2 at melbpc dot org dot au or talldad at kepl dot com dot au >_______________________________________ > >PMTagline v1.50 - Copyright, 1996-1997, Stephen Berg and John Angelico >... Would you trust a BMW controlled by Windows? > Graham Norton Neurologist "when I need a hole in the head, I use Windows when I need a window on the brain, I use OS2!" ---------------------------------------------------------------------------------- **= Email 4 ==========================** Date: Mon, 14 Jul 2003 18:46:34 +1000 From: Alan Heiser Subject: SWAPPER.DAT Stupid question - how do I reduce the size of swapper.dat It has taken over the whole system Thanks Alan ---------------------------------------------------------------------------------- **= Email 5 ==========================** Date: Mon, 14 Jul 2003 19:14:45 +1000 From: Ed Durrant Subject: Re: SWAPPER.DAT Add more real memory ! No really, if you want to limit it's size (and possibly change its location to another disk), find the statement in config.sys, which may look something like this: SWAPPATH=D:\ 4096 40960 The first numerical value (4096 in this example) is how much freespace (in kb) you decided you *MUST* keep free on the drive (in my example, drive D:). When this limit is reached OS/2 will start popping up messages to say that there's no more space. At which point, cancel some of your running applications. The second figure is the initial value to create the swapper.dat file with at boot up. OS/2 will keep increasing the size of the swapper.dat file as it needs to up to the limit of "must have free" that you set, however allocating a "chunk" together at the start is more efficient as the OS has less searching for free space to do. Best solution is to add more "Real" memory. How much do you have in the system at the moment ? Although Warp4 CAN run with as little as 4MB, I'd recommend 64 MB for current applications - more if you're running virtual PC or ODIN. Cheers/2 Ed. Alan Heiser wrote: > Stupid question - how do I reduce the size of swapper.dat > It has taken over the whole system > > Thanks > Alan > > ---------------------------------------------------------------------------------- **= Email 6 ==========================** Date: Mon, 14 Jul 2003 19:08:08 +0950 (CST) From: "Paul Smedley" Subject: Re: Warp Server Hey Graham, On Mon, 14 Jul 2003 15:30:34 +0950 (CST), Graham Norton wrote: >Oops! Now you have upset me! What is the 49 day issue!?? Well, strictly speaking, it's the 49.7 day issue. There's a variable in the kernel that records (from memory) the number of seconds since the system was rebooted. The problem is that the type of variable used only holds enough seconds for 49.7 days. There's been several attempts to correct - we definitely installed one of the Testcase kernels on the FIrewall machine in an attempt to prevent this hang at 49.7 days - I think the WSEB machine is still running the FP3 kernel. Most recent changes to do with the 49.7 day hang in the kernel were: JR15405: (probably) DosQueryAsyncTimer fails on timecritical thread after mscount rolls over after 49.7 days of uptime 20020401 revision 14.088g JR15405 part 2: more likely fix for 49.7 day hang (still in test) 20020802 revision 14.091a - Yet another attempt (the last, we hope) to fix the 49.7 day hang 20020822 revision 14.091b JR15405 (final) the 49.7 day change might not have been in 20020802. 20030213 14.093e JR17862 fix a problem with a thread hanging (but not the whole system) after the 49.7 day rollover doing a DosSleep (*Not* fixed in 1226) Can't remember where FP3 sits in the above.... Cheers, Paul. ---------------------------------------------------------------------------------- **= Email 7 ==========================** Date: Mon, 14 Jul 2003 19:43:51 +1000 (AEST) From: "John Angelico" Subject: Re: SWAPPER.DAT On Mon, 14 Jul 2003 19:14:45 +1000, Ed Durrant wrote: Hi Alan. >Add more real memory ! [chomp]ed Ed's excellent advice. But let's ask a question, as well... > >Alan Heiser wrote: > >> Stupid question - how do I reduce the size of swapper.dat >> It has taken over the whole system Is there some specific reason why it has taken over the whole system? Sounds like a runaway application gobbling up memory until the SWAPPER is occupying the entire drive and, as Ed said, it then pops up messages at you. Best regards John Angelico OS/2 SIG os2 at melbpc dot org dot au or talldad at kepl dot com dot au _______________________________________ PMTagline v1.50 - Copyright, 1996-1997, Stephen Berg and John Angelico .... Oh, no! Not *ANOTHER* learning experience! ---------------------------------------------------------------------------------- **= Email 8 ==========================** Date: Mon, 14 Jul 2003 19:45:04 +1000 From: Ed Durrant Subject: Re: Warp Server Is this a bug in newer releases ? I certainly have systems (some on pretty old versions of code) that have been running FAR longer than 49.7 days !! Cheers/2 Ed. Paul Smedley wrote: > Hey Graham, > > On Mon, 14 Jul 2003 15:30:34 +0950 (CST), Graham Norton wrote: > >Oops! Now you have upset me! What is the 49 day issue!?? > > Well, strictly speaking, it's the 49.7 day issue. There's a variable in the kernel that records > (from memory) the number of seconds since the system was rebooted. The problem is > that the type of variable used only holds enough seconds for 49.7 days. There's been > several attempts to correct - we definitely installed one of the Testcase kernels on the > FIrewall machine in an attempt to prevent this hang at 49.7 days - I think the WSEB > machine is still running the FP3 kernel. > > Most recent changes to do with the 49.7 day hang in the kernel were: > JR15405: (probably) DosQueryAsyncTimer fails on timecritical thread after > mscount rolls over after 49.7 days of uptime > > 20020401 revision 14.088g > JR15405 part 2: more likely fix for 49.7 day hang (still in test) > > 20020802 revision 14.091a > - Yet another attempt (the last, we hope) to fix the 49.7 day hang > > 20020822 revision 14.091b > JR15405 (final) the 49.7 day change might not have been in 20020802. > > 20030213 14.093e > JR17862 fix a problem with a thread hanging (but not the whole system) > after the 49.7 day rollover doing a DosSleep (*Not* fixed in 1226) > > Can't remember where FP3 sits in the above.... > > Cheers, > > Paul. > > ---------------------------------------------------------------------------------- **= Email 9 ==========================** Date: Mon, 14 Jul 2003 19:25:19 +0950 (CST) From: "Paul Smedley" Subject: Re: Warp Server On Mon, 14 Jul 2003 19:45:04 +1000, Ed Durrant wrote: >Is this a bug in newer releases ? I certainly have systems (some on pretty old versions of code) >that have been running FAR longer than 49.7 days !! To be honest - I don't know. I gather it must manifest in different hardware combinations in different ways. There have been MANY attempts to fix this - what I posted was just whats listed in the readme of the latest Testcase kernels. Obviously it was a non-trivial thing to fix!! Cheers, Paul. ---------------------------------------------------------------------------------- **= Email 10 ==========================** Date: Mon, 14 Jul 2003 17:57:13 +0800 From: bob Subject: Re: Warp Server On Mon, 14 Jul 2003 05:45 pm, you wrote: > Is this a bug in newer releases ? I certainly have systems (some on > pretty old versions of code) that have been running FAR longer than 49.7 > days !! The bug isn't a crash issue (like it is in win9x) all it does is wrap the "uptime" counter back to 0 after 49.7 days. The trivial workaround for this is to have a small .cmd write a log of boot times and refer to that. BTW, my record is 619 days. A 12 hr power outage got it in the end (a UPS can only do so much :) > Cheers/2 > Ed. -- By doing just a little every day, you can gradually let the task completely overwhelm you. ---------------------------------------------------------------------------------- **= Email 11 ==========================** Date: Mon, 14 Jul 2003 19:33:32 +0950 (CST) From: "Paul Smedley" Subject: Re: Warp Server Bob, On Mon, 14 Jul 2003 17:57:13 +0800, bob wrote: >On Mon, 14 Jul 2003 05:45 pm, you wrote: >> Is this a bug in newer releases ? I certainly have systems (some on >> pretty old versions of code) that have been running FAR longer than 49.7 >> days !! > >The bug isn't a crash issue (like it is in win9x) all it does is wrap the >"uptime" counter back to 0 after 49.7 days. The trivial workaround for this >is to have a small .cmd write a log of boot times and refer to that. Well according to Scott's changelog in the latest Testcase kernels - he fixed a 49.7 day hang - so there must have been a hang on some hardware :) A search on groups.google dot com shows posts about the hangs too :) Cheers, Paul. ---------------------------------------------------------------------------------- **= Email 12 ==========================** Date: Mon, 14 Jul 2003 20:07:48 +1000 (AEST) From: "John Angelico" Subject: Re: Warp Server On Mon, 14 Jul 2003 19:08:08 +0950 (CST), Paul Smedley wrote: Thanks Paul. I would still be searching for the details of the answer.... I do understand that all eCS 1.1 registered users can go to the eCS site and download the latest kernel 14.094 (not strictly WSEB but given the common kernel I would expect there is an equivalent download for you) which is reputed to totally fix the problem. Interesting note: it seems from Paul's notes to be called the "mscount" - somewhat ironic! [Dick Martin] I didn't know that!! Best regards John Angelico OS/2 SIG talldad at kepl dot com dot au __________________ >Hey Graham, > >On Mon, 14 Jul 2003 15:30:34 +0950 (CST), Graham Norton wrote: >>Oops! Now you have upset me! What is the 49 day issue!?? > >Well, strictly speaking, it's the 49.7 day issue. There's a variable in the kernel that records >(from memory) the number of seconds since the system was rebooted. The problem is >that the type of variable used only holds enough seconds for 49.7 days. There's been >several attempts to correct - we definitely installed one of the Testcase kernels on the >FIrewall machine in an attempt to prevent this hang at 49.7 days - I think the WSEB >machine is still running the FP3 kernel. > >Most recent changes to do with the 49.7 day hang in the kernel were: > JR15405: (probably) DosQueryAsyncTimer fails on timecritical thread after > mscount rolls over after 49.7 days of uptime > > 20020401 revision 14.088g > JR15405 part 2: more likely fix for 49.7 day hang (still in test) > > 20020802 revision 14.091a > - Yet another attempt (the last, we hope) to fix the 49.7 day hang > > 20020822 revision 14.091b > JR15405 (final) the 49.7 day change might not have been in 20020802. > > 20030213 14.093e > JR17862 fix a problem with a thread hanging (but not the whole system) > after the 49.7 day rollover doing a DosSleep (*Not* fixed in 1226) > >Can't remember where FP3 sits in the above.... > >Cheers, > >Paul. > > > > ---------------------------------------------------------------------------------- **= Email 13 ==========================** Date: Mon, 14 Jul 2003 19:45:58 +0950 (CST) From: "Paul Smedley" Subject: Re: Warp Server Hiya John, On Mon, 14 Jul 2003 20:07:48 +1000 (AEST), John Angelico wrote: >On Mon, 14 Jul 2003 19:08:08 +0950 (CST), Paul Smedley wrote: > >I would still be searching for the details of the answer.... Well I'm a testcase junky and always read the changelon - plus I hang out on Usenet a lot and saw the original posts! >I do understand that all eCS 1.1 registered users can go to the eCS site >and download the latest kernel 14.094 (not strictly WSEB but given the >common kernel I would expect there is an equivalent download for you) which >is reputed to totally fix the problem. Well the absolute latest kernels are currently sitting on IBM's Testcase site at ftp://testcase.boulder.ibm dot com/ps/fromibm/os2 for WSEB Uni-processor ftp://testcase.boulder.ibm dot com/ps/fromibm/os2/uni20030621.zip for WSEB SMP-Processor ftp://testcase.boulder.ibm dot com/ps/fromibm/os2/smp20030621.zip for Warp 4/eCS/MCP/MCP2 ftp://testcase.boulder.ibm dot com/ps/fromibm/os2/w420030621.zip >Interesting note: it seems from Paul's notes to be called the "mscount" - >somewhat ironic! Yeah :) or could be millisecond count :) but hey let's blame M$ for the heck of it :) Cheers, Paul ---------------------------------------------------------------------------------- **= Email 14 ==========================** Date: Mon, 14 Jul 2003 20:27:17 +1000 From: "Mike O'Connor" Subject: Re: SWAPPER.DAT Ed Durrant wrote: >Add more real memory ! > >No really, if you want to limit it's size (and possibly change its location to another >disk), find the statement in config.sys, which may look something like this: > >SWAPPATH=D:\ 4096 40960 > >The first numerical value (4096 in this example) is how much freespace (in kb) you >decided you *MUST* keep free on the drive (in my example, drive D:). When this limit >is reached OS/2 will start popping up messages to say that there's no more space. At >which point, cancel some of your running applications. The second figure is the >initial value to create the swapper.dat file with at boot up. OS/2 will keep >increasing the size of the swapper.dat file as it needs to up to the limit of "must >have free" that you set, however allocating a "chunk" together at the start is more >efficient as the OS has less searching for free space to do. > >Best solution is to add more "Real" memory. How much do you have in the system at the >moment ? Although Warp4 CAN run with as little as 4MB, I'd recommend 64 MB for current >applications - more if you're running virtual PC or ODIN. > >Cheers/2 > >Ed. > >Alan Heiser wrote: > > >>Stupid question - how do I reduce the size of swapper.dat >>It has taken over the whole system >> Hi Alan, Ed, Although with Warp3 Connect and 64MB RAM, after everything [NetBIOS etc.] loaded, you still have approx 20MB of RAM FREE, using any version of Warp4 that will show up as only 512KB of RAM FREE - just sufficient to run the virtual memory paging mechanism. If possible with Warp4 and better I would recommend not less than 96MB or 128MB - depending on the motherboard's SIMM/DIMM etc., capability - there are very cheap RAM modules around still. I saw 256MB SDRAM for AU$59 locally - should be even cheaper in one of the Capital Cities. -- Regards, Mike Failed the exam for -------------------- MCSE - Minesweeper Consultant and Solitaire Expert -------------------- ---------------------------------------------------------------------------------- **= Email 15 ==========================** Date: Mon, 14 Jul 2003 19:07:42 +0800 From: bob Subject: Re: Warp Server On Mon, 14 Jul 2003 05:43 pm, you wrote: > Well according to Scott's changelog in the latest Testcase kernels - he > fixed a 49.7 day hang - so there must have been a hang on some hardware > :) A search on groups.google dot com shows posts about the hangs too :) Thats not good. I'd be real unhappy if I couldn't get > 50 days uptime. Not one of the OS/2 servers I've had anything to do with has these sorts of problems... and I never heard anything other than rumours about these sorts of crashes, which I discounted as bad HW. > Cheers, > > Paul. As an aside... I wasn't aware of this before now but apparently a lot of Petrol servos run OS/2 on their control consoles. A friend of mine has just scored a job doing system support with a mob called Metric through his experience with OS/2 :) -- Feel disillusioned? I've got some great new illusions, right here! ----------------------------------------------------------------------------------