Re: problem with an old vm/esa environment after migration to a flex-estserver
Hi Gary, yes we are currently on the way to open an incident at the reseller. I will email you off-list the reseller's and the customer's name. Thank you all for your help and many ideas so far. Franz Josef - Original Message - From: Gary Eheman [EMAIL PROTECTED] To: IBMVM@LISTSERV.UARK.EDU Sent: Tuesday, December 19, 2006 2:11 AM Subject: Re: [IBMVM] problem with an old vm/esa environment after migration to a flex-estserver Franz Josef: It would be easier to handle some of the technical details off-list. Please open an incident through the normal support structure for FLEX-ES resellers and we will be happy to pursue. On Mon, 18 Dec 2006 21:01:51 +0100, Pohlen (Mailinglist) [EMAIL PROTECTED] wrote: Hi Gary and Ed, the answer why I have not answered the questions of Ed is simple. For whatever reason I haven't got them into my outlook. The other posts have all reached me. But following your hint I have searched the archives and fou nd the questions. Here are the answers What release is SQL/DS? I don't know currently. The system status is from approx. 1993 - 1994 Is the SQL/DS server using VM dataspaces? There is the dataspace MAP00 which I can see on IND SPACES. Ther e I also see the I/Os which Kris told me is no paging but the real DB i/o. I n fact there is currently no paging. I only didn't know that the database i/o is handled like paging and was misleaded from the high paging rate. Was their expanded storage on the old box? If so, how much? no, on the old box they had configured only 193 MB central storage on th e tserver they have 768 MB Is the Tserver configured with expanded storage? If so, how much? no Did the old system have cached DASD? it was a multiprise 2000 or 2003 with 32MB cache defined Are you using enough cache for the DASD on the Tserver? I don't know how much cache is on the raid controller of the tserver, bu t for sure it is much more than 32MB. The machine has physically 2 GB stor age. Are you paging on Unix/Linux at all on the Tserver? no Meanwhile after we have figured out that the high paging rate is DASD i/ o I don't think that memory is the problem. If I substract those i/os from paging there is zero paging. Gary, now one question to you. On a real mainframe there is the separate system assist processor to do the i/o. On an intel machine the i/o has to be done by the normal processor. We have sized the machine to a nearly equivalent mips rate of the old machine. Is it possible, that on a syste m with relative high i/o rate the processor eats too much from its capacit y for i/o handling which is then on the other hand missing for the non-i/o work? In other words, can the equavalent sizing of the mips rate be wron g on systems with high i/o AND high cpu, so that we need more mips on the tserver? regards Franz Josef - Original Message - From: Gary Eheman [EMAIL PROTECTED] To: IBMVM@LISTSERV.UARK.EDU Sent: Saturday, December 16, 2006 5:18 PM Subject: Re: [IBMVM] problem with an old vm/esa environment after migrat ion to a flex-estserver On Fri, 15 Dec 2006 15:30:38 +0100, Pohlen (Mailinglist) [EMAIL PROTECTED] wrote: Kris, I have now figured out from the IND SPACE every minute that the SQLDS u s er has only DASD paging between 300 and 700 pages per second in spaceid MAP00. The BASE spaceid has no paging. Because Xstore is not de f ined there is obviously no xstore paging. Would it make sense to define Xsto r e? I'm not sure if a CMS user uses it. Franz Josef The paradigm is different on a FLEX-ES system than on a conventional IBM mainframe with respect to xstore. Whereas paging to xstore may have been faster than I/O to disk could hav e been accomplished on the old mainframe, on a FLEX-ES system using intern a l dasd, with a cache hit we can satisfy an I/O from control unit cache in micro seconds under ideal conditions. The path length for xstore is long e r. The Oh, xstore is faster paradigm does not fit well at all in the FLEX-ES world. I do not generally recommend xstore in a VM environment, other than a to k en amount due to an idiosyncracy of the CP paging algorithms in VM that may function slightly better with a token amount available. And I never recommend MDC to xstore on a FLEX-ES system in light of the speed that I / O can be satisfied from controller cache. Ed Zell asked some excellent and pertinent questions in his post for whi c h I did not see an answer posted. -- Gary Eheman Fundamental Software, Inc http://www.funsoft.com =
Re: problem with an old vm/esa environment after migration to a flex-estserver
Kris, I have now figured out from the IND SPACE every minute that the SQLDS user has only DASD paging between 300 and 700 pages per second in spaceid MAP00. The BASE spaceid has no paging. Because Xstore is not defined there is obviously no xstore paging. Would it make sense to define Xstore? I'm not sure if a CMS user uses it. Franz Josef - Original Message - From: Kris Buelens To: Pohlen (Mailinglist) Sent: Tuesday, December 12, 2006 8:48 PM Subject: Re: problem with an old vm/esa environment after migration to a flex-estserver Here's my EXEC. Kris, IBM Belgium, VM customer support Pohlen (Mailinglist) [EMAIL PROTECTED] Sent by: The IBM z/VM Operating System IBMVM@LISTSERV.UARK.EDU 2006-12-12 17:45 Please respond to The IBM z/VM Operating System IBMVM@LISTSERV.UARK.EDU To IBMVM@LISTSERV.UARK.EDU cc Subject Re: problem with an old vm/esa environment after migration to a flex-es tserver Hi Kris, it seems to be dasd paging, because there is also heavy i/o on one paging dasd. I have found your data in memory techniques document, but there is no exec for checking ind spaces only qnssmap exec is listed there. But it was a good hint. I have created a small exec which does the ind spaces user xxx every minute and writes the result with a timestamp into a file. This I will let the customer run a few hours and check the results. I will not parse the command now because I have zVM 5.2 and I'm not sure if the command output is the same on VM/ESA 2.2. If I have the results available I will contact you again. regards Franz Josef - Original Message - From: Kris Buelens [EMAIL PROTECTED] To: IBMVM@LISTSERV.UARK.EDU Sent: Tuesday, December 12, 2006 3:44 PM Subject: Re: [IBMVM] problem with an old vm/esa environment after migration to a flex-es tserver Do you know if this paging is real paging or Dataspace I/O: when using VM dataspace support in DB2, all DB2 I/O is done by CP paging, hence high page rates. Quite some years ago, I created a document to explain the difference. It is called Data in memory techniques or alike and available on the VM web page. If you search for EXECLOAD, NUCXLOAD and, BUELENS you should quickly find it back. It contains an EXEC that you can run in a server every x minutes and that then will report how many real page in/out happend and how many dataspace read/writes (based on counts reported by CP IND SPACES) If you've got RTM/ESA, its DISPLAY SYSDASD command also reports the dataspace paging I/O Kris, IBM Belgium, VM customer support Pohlen (Mailinglist) [EMAIL PROTECTED] Sent by: The IBM z/VM Operating System IBMVM@LISTSERV.UARK.EDU 2006-12-12 14:54 Please respond to The IBM z/VM Operating System IBMVM@LISTSERV.UARK.EDU To IBMVM@LISTSERV.UARK.EDU cc Subject problem with an old vm/esa environment after migration to a flex-es tserver Hi listers, I have a customer with a really old VM/ESA 2.2, VSE/ESA 2.3 and SQLDS database. Now after migration to the tserver he complains about performance problems. What I have figured out that he has massive paging only on SQLDS service machine (300-500 pages/s). This I cannot understand, because first he has more memory available as before on multiprise (193 MB vs. 768 MB now) and second we haven't changed the system layout compared to the multiprise. We have migrated the system by dump/restoring the volumes with DDR. So what can cause only the sqlds machine to page so heavily? I have already doubled the virtual memory to 96 MB but this had no effect on the behaviour. Does anybody have an idea, where I can search for the problem? Mit freundlichen Grüßen / best regards Franz Josef Pohlen
problem with an old vm/esa environment after migration to a flex-es tserver
Hi listers, I have a customer with a really old VM/ESA 2.2, VSE/ESA 2.3 and SQLDS database. Now after migration to the tserver he complains about performance problems. What I have figured out that he has massive paging only on SQLDS service machine (300-500 pages/s). This I cannot understand, because first he has more memory available as before on multiprise (193 MB vs. 768 MB now) and second we haven't changed the system layout compared to the multiprise. We have migrated the system by dump/restoring the volumes with DDR. So what can cause only the sqlds machine to page so heavily? I have already doubled the virtual memory to 96 MB but this had no effect on the behaviour. Does anybody have an idea, where I can search for the problem? Mit freundlichen Grüßen / best regards Franz Josef Pohlen
Re: problem with an old vm/esa environment after migration to a flex-es tserver
Hi Kris, it seems to be dasd paging, because there is also heavy i/o on one paging dasd. I have found your data in memory techniques document, but there is no exec for checking ind spaces only qnssmap exec is listed there. But it was a good hint. I have created a small exec which does the ind spaces user xxx every minute and writes the result with a timestamp into a file. This I will let the customer run a few hours and check the results. I will not parse the command now because I have zVM 5.2 and I'm not sure if the command output is the same on VM/ESA 2.2. If I have the results available I will contact you again. regards Franz Josef - Original Message - From: Kris Buelens [EMAIL PROTECTED] To: IBMVM@LISTSERV.UARK.EDU Sent: Tuesday, December 12, 2006 3:44 PM Subject: Re: [IBMVM] problem with an old vm/esa environment after migration to a flex-es tserver Do you know if this paging is real paging or Dataspace I/O: when using VM dataspace support in DB2, all DB2 I/O is done by CP paging, hence high page rates. Quite some years ago, I created a document to explain the difference. It is called Data in memory techniques or alike and available on the VM web page. If you search for EXECLOAD, NUCXLOAD and, BUELENS you should quickly find it back. It contains an EXEC that you can run in a server every x minutes and that then will report how many real page in/out happend and how many dataspace read/writes (based on counts reported by CP IND SPACES) If you've got RTM/ESA, its DISPLAY SYSDASD command also reports the dataspace paging I/O Kris, IBM Belgium, VM customer support Pohlen (Mailinglist) [EMAIL PROTECTED] Sent by: The IBM z/VM Operating System IBMVM@LISTSERV.UARK.EDU 2006-12-12 14:54 Please respond to The IBM z/VM Operating System IBMVM@LISTSERV.UARK.EDU To IBMVM@LISTSERV.UARK.EDU cc Subject problem with an old vm/esa environment after migration to a flex-es tserver Hi listers, I have a customer with a really old VM/ESA 2.2, VSE/ESA 2.3 and SQLDS database. Now after migration to the tserver he complains about performance problems. What I have figured out that he has massive paging only on SQLDS service machine (300-500 pages/s). This I cannot understand, because first he has more memory available as before on multiprise (193 MB vs. 768 MB now) and second we haven't changed the system layout compared to the multiprise. We have migrated the system by dump/restoring the volumes with DDR. So what can cause only the sqlds machine to page so heavily? I have already doubled the virtual memory to 96 MB but this had no effect on the behaviour. Does anybody have an idea, where I can search for the problem? Mit freundlichen Grüßen / best regards Franz Josef Pohlen
Re: real ctc vs vctc on tcp/ip
Are the channel devices dedicated or shared. Can you show us the iocds definitions. regards Franz Josef - Original Message - From: Ron Greve [EMAIL PROTECTED] To: IBMVM@LISTSERV.UARK.EDU Sent: Monday, September 11, 2006 5:41 AM Subject: Re: [IBMVM] real ctc vs vctc on tcp/ip try setting the adapter number on the LINK statements to be the same: = Both 0s or both 1s David David We set them both to 1 as show below. We now get a configuration error DTCCTC059 (TCP/IP recognizes link adapter address is invalid.) We set them back to 0 and 1 and we can get the links up again. Netstat Devlinks show both sides Ready with no error status. Netstat Gate show the route with flags UHS. As soon as we ping from one side or the other we get the Unit Check. Netstat Devlinks show bytes out going up but nothing on bytes in. Side being pinged now show error status on Netstat Devlinks Ron lpar one -- attach 448 tcpip 720 attach 449 tcpip 721 lpar two -- attach 508 tcpip 720 attach 509 tcpip 721 ; lpar one tcp/ip device/link DEVICE CTCVDEV4 CTC 720 LINK CTCVLNK4 CTC 1 CTCVDEV4 ; lpar two tcp/ip device/link DEVICE CTCVDEV4 CTC 0720 LINK CTCVLNK4 CTC 1 CTCVDEV4 MTU 1500
Re: Problem activating FEA cards on MP3000
Hi Tom, check if there that there is no tcpip protocol active in mpts for the network cards you want to use for VM or VSE. There may be only SNA protocol active on it. The TCPIP protocol is only used for network cards which emulate console devices. Another possibility is to check if the adapter numbers. The onboard adapter should have adapter number zero and then the PCI Ethernet cards should follow. In my old self-written hints I have a remark, that in config.sys there must be the driver AWSLCSDD.SYS active before using TCPIP. Hope this helps Franz Josef - Original Message - From: Tom Cluster [EMAIL PROTECTED] To: IBMVM@LISTSERV.UARK.EDU Sent: Thursday, August 17, 2006 5:33 PM Subject: [IBMVM] Problem activating FEA cards on MP3000 This is posted both in the vse and vm listservs. We have a used 7060 for DR purposes. It has 3 FEA cards for IP traffic, to be used by VM and VSE TCP/IP. The IOCDS entries are copied from those for the FEA cards on our production 7060 and our TCP/IP for VM configurations are copied from our production system (with appropriate changes for different IP addresses). My EMIO configuration shows them as CTC devices. The EMIO channel is working because we're able to use the TN3270 session on the HMC. I've gone into various support element functions that purport to show the status of components on the 7060 and they show the cards as being installed and idle. The problem is that I cannot vary them online in VM - no channel path available. The cards seem to be powered up because the red lights are blinking (actually, they're only blinking on two of the cards - on the third card there are no lights, which I find interesting - this is one of the things I intend to talk with our CE about). I am trying to contact my IBM CE, and he may be able to figure out why I can't vary them online, but I was wondering if anyone has any tips. The cards are plugged into our network, but I don't know that the network drops are active. I wouldn't think, though, that an inactive network drop would prevent my varying them online in VM. Is there a configuration function on the HMC that pertains solely to the FEA's? You know, something that would configure them for IP traffic or SNA traffic, etc.? I'm not talking about the EMIO configuration panels. I appreciate any ideas you may have. Thanks! - Tom. Tom Cluster County of Sonoma Santa Rosa, CA (707) 565-3384 (Tuesdays and Wednesdays only)
Re: again a problem GCS/VTAM but now related to TUBES/VTAM
VTAM itself works well and there is only one segment each for both GCS and VTAM. VTUBES does not complain when the old original sized GCS segment is in effect, therefore VTAM complains about insufficient CSA storage. It is always a mess if customers do nothing to keep their system rather actual and when they change the hardware, in this case replacing a Multiprise with 3745-Tokenring-SNA-Connection by a t-server with XCA-Ethernet. On the old Multiprise everything has worked because the Tokenring adapters were handled by the 3745 and the Ethernet adapters on the t-server are handled by VTAM itself. The customer has connected more than 600 XID-PUs and then VTAM got the storage shortage in CSA. The increase of the GCS segment helped with the CSA shortage. Now VTUBES complains. Perhaps the increase of VTUBES to 128MB like Jim suggested helps. Otherwise I have advised the customer to change the terminal-PUs to a Communication Server which saves a lot of the PUs and XCA lines. - Original Message - From: Tom Duerbusch [EMAIL PROTECTED] To: IBMVM@LISTSERV.UARK.EDU Sent: Wednesday, March 29, 2006 6:17 PM Subject: Re: [IBMVM] again a problem GCS/VTAM but now related to TUBES/VTAM Just be sure you have the right saved segment. Stupid story.. Testing on our z/890 with z/VM 5.1 went fine. Conversion weekend. SPXTAPE DUMP all (from z/VM 4.2) SPXTAPE LOAD all (to z/VM 5.1) Hours later we IPL'ed VM, all hell broke loose. Turned out I dump/loaded all the saved segments. At IPL, the oldest ones (z/VM 4.2) were picked up. Tom Duerbusch THD Consulting [EMAIL PROTECTED] 3/29/2006 9:04 AM The level of VM should not have much to do with VTAM as GCS is very stable. There might be a PTF or two in the last few years. Are you running VM/VTAM 4.2.0? I see no reason to change the size or location of the VTAM segment from 600-6FF: q nss name vtama map FILE FILENAME FILETYPE MINSIZE BEGPAG ENDPAG TYPE CL #USERS PARMREGS VMGROUP 1370 VTAMADCSS N/A00600 006FF SR A 00016 N/A N/A Ready; T=0.01/0.01 09:54:08 With your new gcs like: q nss name gcsc map FILE FILENAME FILETYPE MINSIZE BEGPAG ENDPAG TYPE CL #USERS PARMREGS VMGROUP 1417 GCSC NSS 256K 0 C EW R 0 OMITTED YES 00700 0074E SR 0074F 0074F SW 00750 009FF SN 01000 0101A SR 0101B 04FFF SN I am not familiar with VTUBES. If the vendor says that it will not use storage above 16M, I do not know what to say. I would suggest bring the machine up to at least 80M to ensure all of GCS is inside the machine (as I stated we run ALL machines in the GCS group at 106M). It does sound as if VTUBES is I/O bound under 16M. Have you added a bunch of new devices to it? Jim -Original Message- From: The IBM z/VM Operating System [mailto:[EMAIL PROTECTED] On Behalf Of Pohlen (Mailinglist) Sent: Wednesday, March 29, 2006 9:52 AM To: IBMVM@LISTSERV.UARK.EDU Subject: again a problem GCS/VTAM but now related to TUBES/VTAM Hi listers, last week I asked for help on a CSA storage problem with pool 231. I got help by Marcy Cortes and Jim Stracka. Their suggestion for increasing the GCS shared segment worked. Now the customer has a storage problem with TUBES-VTAM. The Macro4 support says it cannot GETMAIN more than 5MB of storage. The virtual machine size is 64MB and MACHINE ESA. The support personnel of Macro4 does not have a clue what might cause that VTUBES does apperantly not use storage above 16MB. They told the customer to call IBM for help. But firstly it is VM/ESA 2.2 and secondly the customer has no contract on support (even if he had one it would not be useful because VM/ESA 2.2 is out of service for more than 5 years). Therefore I try it again with the list hoping someone has - as often before - a good idea, what might be the reason. I append the GCS segment definition currently in use for documentation. The 0148 was the old segment and the 0151 is the new definition. 0148 GCS NSS 256K 0 C EW R 0 OMITTED YES 00400 0044E SR 0044F 0044F SW 00450 005FF SN 01000 0101A SR 0101B 011FF SN 0151 GCS NSS 256K 0 C EW S 0 OMITTED YES 00700 0074E SR 0074F 0074F SW 00750 009FF SN 01000 0101A SR 0101B 04FFF SN HCPNSS440I Named Saved System (NSS) GCS was successfully saved in fileid 0151. After installing VTAM-PTFs I had to increase the VTAM segment from 600-6FF to 600-7FF. There were only a few K which did not fit into the old 600