Re: LTO5 Tuning
Thanks for your interest Mike. The san guy tells me it's 380 microseconds which equates to 76 km distance. Regards Steve On 30 Oct 2014 23:50, "Ryder, Michael S" wrote: > Steve: > > What is the latency on your intersite connection? > > Best regards, > > Mike, x7942 > RMD IT Client Services > > On Thu, Oct 30, 2014 at 8:35 AM, Steven Harris > wrote: > > > Hi All > > > > I have a TSM 6.3.4 server on newish P7 hardware and AIX V7.1. HBAs are > > all 8Gb. The sans behind it are 8Gb or 4Gb depending which path they > > take as we are in the middle of a SAN upgrade and there is still an old > > switch in the mix. > > > > Disk is XIV behind SVC. Tape is TS3500 and LTO5. > > > > According to the LTO wikipedia entry I should be able to get 140MB/sec > > raw out of the drive. I have an internal company document that suggests > > sustained 210MB/sec (compressed) is attainable in the real world. > > > > So far my server backs up 500GB per night of DB2 and Oracle databases on > > to file pools, without deduplication. Housekeeping then does a > > single-streamed simultaneous migrate and copy to onsite and offsite > > tapes. Inter site bandwidth is 4Gb and I have most of that to myself. > > > > That process takes over 5 hours so I'm seeing less than 100MB/sec. > > > > Accordingly I started a tuning exercise. I copied 50GB of my filepool > > twice to give me a test dataset and started testing, of course when > > there was no other activity on the TSM box. > > > > The data comes off disk at 500MB/sec to /dev/null, so that is not a > > bottleneck. > > > > Copying using dd to tape runs at a peak of 120MB/sec with periods of > > much lower than that, as measured using nmon's fc stats on the HBAs. I > > presume some of that slowdown is where the tape reaches its end and has > > to reverse direction. > > > > Elapsed time for 100GB is 18 min, with little variation so average speed > > is 95MB/sec > > > > dd ibs and obs values were varied and ibs=256K obs=1024K seems to give > > the best result. > > > > Elapsed time is very consistent. > > > > Copying to a local drive on the same switch blade as the tape HBA or > > copying across blades made no difference. > > > > Copying to a drive at the remote site increased elapsed time by 2 > > minutes, as one would expect with more switches in the path and a longer > > turnaround time. > > > > Tape to tape copy was not noticeably different to disk to tape. > > > > Reading from tape to /dev/null was no different. > > > > In all cases CPU time was about half of the elapsed time. > > > > lsattr on the drives shows that compression is on (this is also the > > default) > > > > The tape FC adapters are set to use the large transfer size. > > > > The test was also run using 64KB pages and svmon was used to verify the > > setting was effective. Again no difference. > > > > I'm running out of ideas here. num_cmd_elements on the hbas is 500 (the > > default) I'm thinking of increasing that to 2000, but it will require > > an outage and hence change control. > > > > Does anyone have any ideas, references I could look at or practical > > advice as to how to get this to perform? > > > > Thanks > > > > Steve > > > > Steven Harris > > TSM Admin > > Canberra Australia > > >
Re: Monitoring Agent, 2 on a single TSM Server?
You can do that on AIX and Linux, by installing the code into different directories. You can't however do that on Windows. But what issue are you dealing with between the versions? Lee Miller Tivoli Storage Manager for System Backup and Recovery Development and IBM Tivoli Monitoring for Tivoli Storage Manager Phone: 817-874-7484 From: "Vandeventer, Harold [OITS]" To: ADSM-L@VM.MARIST.EDU Date: 10/30/2014 03:58 PM Subject:[ADSM-L] Monitoring Agent, 2 on a single TSM Server? Sent by:"ADSM: Dist Stor Manager" Is it possible to install two Monitoring Agents on a single TSM server? My hope: leave the current remote agent that pushes data from TSM 6.3.x to the production Monitoring/Reporting/Admin Center server where production reports are produced. But, also install a second Remote Monitoring Agent (related to TSM 7.1.1) on that same TSM 6.3 server and configure it to push data to a new Monitoring/Reporting server. This would let me workout the wrinkles in new reporting and continue to use Remote Agents, not several local agents. Alternative, I suppose: create a second local Agent on the new Monitoring server and in time delete it, when I upgrade the Remote Agent to the new release. Thanks. Harold Vandeventer Systems Programmer State of Kansas - Office of Information Technology Services STE 751-S 910 SW Jackson (785) 296-0631 [Confidentiality notice:] *** This e-mail message, including attachments, if any, is intended for the person or entity to which it is addressed and may contain confidential or privileged information. Any unauthorized review, use, or disclosure is prohibited. If you are not the intended recipient, please contact the sender and destroy the original message, including all copies, Thank you. ***
Re: TSM 7.1 usage of volumes for dedupe
> Op 30 okt. 2014, om 22:04 heeft Colwell, William F. het > volgende geschreven: > > Hi Martha, > > I am glad this was useful to you. > > I have not reported this as a bug; I expect they would say > working-as-designed, try > submitting an rfe. I never understood that working as designed is a reason to close a call. If the design is bodged, then it should be fixed. I’ve been told that even TSM is designed by people, and I’m sure sometimes those do make mistakes. > > - bill > > -Original Message- > From: ADSM: Dist Stor Manager [mailto:ADSM-L@VM.MARIST.EDU] On Behalf Of > Martha M McConaghy > Sent: Thursday, October 30, 2014 10:09 AM > To: ADSM-L@VM.MARIST.EDU > Subject: Re: TSM 7.1 usage of volumes for dedupe > > Bill, > > I just wanted to let you know how much this information helped. I was > able to clear out all the problem volumes and have removed the full LUNs > from the devclass until there is enough space on them to be used again. > > This situation really seems strange to me. Why has TSM not been updated > to handle the out of space condition better? If it has a command that > shows how much space is left on the LUN, why can't TSM understand it is > time to stop allocating volumes on it? Forcing admins to do manual > clean up like this just to keep things healthy seems inconsistent with > how the rest of TSM functions. > > Has anyone ever reported this as a bug? > > Martha > > On 10/22/2014 2:38 PM, Colwell, William F. wrote: >> Hi Martha, >> >> I see this situation occur when a filesystem gets almost completely full. >> >> Do 'q dirsp ' to check for nearly full filesystems. >> >> The server doesn't fence off a filesystem like this, instead it keeps >> hammering on it, allocating new volumes. When it tries to write to a volume >> and gets an immediate out-of-space error, it marks the volume full so it >> won't >> try to use it again. >> >> I run this sql to find such volumes and delete them - >> >> select 'del v '||cast(volume_name as char(40)), cast(stgpool_name as >> char(30)), last_write_date - >> from volumes where upper(status) = 'FULL' and pct_utilized = 0 and >> pct_reclaim = 0 order by 2, 3 >> >> You should remove such filesystems from the devclass directory list until >> reclaim has emptied them a little bit. >> >> Hope his helps, >> >> Bill Colwell >> Draper Lab >> >> >> >> -Original Message- >> From: ADSM: Dist Stor Manager [mailto:ADSM-L@VM.MARIST.EDU] On Behalf Of >> Martha M McConaghy >> Sent: Wednesday, October 22, 2014 2:23 PM >> To: ADSM-L@VM.MARIST.EDU >> Subject: Re: TSM 7.1 usage of volumes for dedupe >> >> Interesting. Seems very similar, except the status of these volumes is >> "FULL", not "EMPTY". However, the %reclaimable space is 0.0. >> >> I think this is a bug. I would expect the volume to leave the pool once >> it is "reclaimed". It would be OK with me if it did not. However, since >> the status is "FULL", it will never be reused. That seems wrong. If it >> is going to remain attached to the dedupepool, the status should convert >> to EMPTY so the file can be reused. Or, go away altogether so the space >> can be reclaimed and reused. >> >> In looking at the filesystem on the Linux side (sorry I didn't mention >> this is running on RHEL), the file exists on /data0, but with no size: >> >> [urmm@tsmserver data0]$ ls -l *d57* >> -rw--- 1 tsminst1 tsmsrvrs 0 Oct 10 20:22 0d57.bfs >> >> /data0 is 100% utilized, so this file can never grow. Seems like it >> should get cleaned up rather than continue to exist. >> >> Martha >> >> On 10/22/2014 1:58 PM, Erwann SIMON wrote: >>> hi Martha, >>> >>> See if this can apply : >>> www-01.ibm.com/support/docview.wss?uid=swg21685554 >>> >>> Note that I had a situation where Q CONT returned that the volume was empty >>> but it wasn't in reality since it was impossible to delete it (without >>> discrading data). A select statement against the contents showed some >>> files. Unforunately, I don't know how this story finished... >>> >> -- >> Martha McConaghy >> Marist: System Architect/Technical Lead >> SHARE: Director of Operations >> Marist College IT >> Poughkeepsie, NY 12601 >> >> Notice: This email and any attachments may contain proprietary (Draper >> non-public) and/or export-controlled information of Draper Laboratory. If >> you are not the intended recipient of this email, please immediately notify >> the sender by replying to this email and immediately destroy all copies of >> this email. >> > > -- > Martha McConaghy > Marist: System Architect/Technical Lead > SHARE: Director of Operations > Marist College IT > Poughkeepsie, NY 12601 > > Notice: This email and any attachments may contain proprietary (Draper > non-public) and/or export-controlled information of Draper Laboratory. If you > are not the intended recipient of this email, please imm
Re: TSM 7.1 usage of volumes for dedupe
Hi Martha, I am glad this was useful to you. I have not reported this as a bug; I expect they would say working-as-designed, try submitting an rfe. - bill -Original Message- From: ADSM: Dist Stor Manager [mailto:ADSM-L@VM.MARIST.EDU] On Behalf Of Martha M McConaghy Sent: Thursday, October 30, 2014 10:09 AM To: ADSM-L@VM.MARIST.EDU Subject: Re: TSM 7.1 usage of volumes for dedupe Bill, I just wanted to let you know how much this information helped. I was able to clear out all the problem volumes and have removed the full LUNs from the devclass until there is enough space on them to be used again. This situation really seems strange to me. Why has TSM not been updated to handle the out of space condition better? If it has a command that shows how much space is left on the LUN, why can't TSM understand it is time to stop allocating volumes on it? Forcing admins to do manual clean up like this just to keep things healthy seems inconsistent with how the rest of TSM functions. Has anyone ever reported this as a bug? Martha On 10/22/2014 2:38 PM, Colwell, William F. wrote: > Hi Martha, > > I see this situation occur when a filesystem gets almost completely full. > > Do 'q dirsp ' to check for nearly full filesystems. > > The server doesn't fence off a filesystem like this, instead it keeps > hammering on it, allocating new volumes. When it tries to write to a volume > and gets an immediate out-of-space error, it marks the volume full so it won't > try to use it again. > > I run this sql to find such volumes and delete them - > > select 'del v '||cast(volume_name as char(40)), cast(stgpool_name as > char(30)), last_write_date - > from volumes where upper(status) = 'FULL' and pct_utilized = 0 and > pct_reclaim = 0 order by 2, 3 > > You should remove such filesystems from the devclass directory list until > reclaim has emptied them a little bit. > > Hope his helps, > > Bill Colwell > Draper Lab > > > > -Original Message- > From: ADSM: Dist Stor Manager [mailto:ADSM-L@VM.MARIST.EDU] On Behalf Of > Martha M McConaghy > Sent: Wednesday, October 22, 2014 2:23 PM > To: ADSM-L@VM.MARIST.EDU > Subject: Re: TSM 7.1 usage of volumes for dedupe > > Interesting. Seems very similar, except the status of these volumes is > "FULL", not "EMPTY". However, the %reclaimable space is 0.0. > > I think this is a bug. I would expect the volume to leave the pool once > it is "reclaimed". It would be OK with me if it did not. However, since > the status is "FULL", it will never be reused. That seems wrong. If it > is going to remain attached to the dedupepool, the status should convert > to EMPTY so the file can be reused. Or, go away altogether so the space > can be reclaimed and reused. > > In looking at the filesystem on the Linux side (sorry I didn't mention > this is running on RHEL), the file exists on /data0, but with no size: > > [urmm@tsmserver data0]$ ls -l *d57* > -rw--- 1 tsminst1 tsmsrvrs 0 Oct 10 20:22 0d57.bfs > > /data0 is 100% utilized, so this file can never grow. Seems like it > should get cleaned up rather than continue to exist. > > Martha > > On 10/22/2014 1:58 PM, Erwann SIMON wrote: >> hi Martha, >> >> See if this can apply : >> www-01.ibm.com/support/docview.wss?uid=swg21685554 >> >> Note that I had a situation where Q CONT returned that the volume was empty >> but it wasn't in reality since it was impossible to delete it (without >> discrading data). A select statement against the contents showed some files. >> Unforunately, I don't know how this story finished... >> > -- > Martha McConaghy > Marist: System Architect/Technical Lead > SHARE: Director of Operations > Marist College IT > Poughkeepsie, NY 12601 > > Notice: This email and any attachments may contain proprietary (Draper > non-public) and/or export-controlled information of Draper Laboratory. If you > are not the intended recipient of this email, please immediately notify the > sender by replying to this email and immediately destroy all copies of this > email. > -- Martha McConaghy Marist: System Architect/Technical Lead SHARE: Director of Operations Marist College IT Poughkeepsie, NY 12601 Notice: This email and any attachments may contain proprietary (Draper non-public) and/or export-controlled information of Draper Laboratory. If you are not the intended recipient of this email, please immediately notify the sender by replying to this email and immediately destroy all copies of this email.
Monitoring Agent, 2 on a single TSM Server?
Is it possible to install two Monitoring Agents on a single TSM server? My hope: leave the current remote agent that pushes data from TSM 6.3.x to the production Monitoring/Reporting/Admin Center server where production reports are produced. But, also install a second Remote Monitoring Agent (related to TSM 7.1.1) on that same TSM 6.3 server and configure it to push data to a new Monitoring/Reporting server. This would let me workout the wrinkles in new reporting and continue to use Remote Agents, not several local agents. Alternative, I suppose: create a second local Agent on the new Monitoring server and in time delete it, when I upgrade the Remote Agent to the new release. Thanks. Harold Vandeventer Systems Programmer State of Kansas - Office of Information Technology Services STE 751-S 910 SW Jackson (785) 296-0631 [Confidentiality notice:] *** This e-mail message, including attachments, if any, is intended for the person or entity to which it is addressed and may contain confidential or privileged information. Any unauthorized review, use, or disclosure is prohibited. If you are not the intended recipient, please contact the sender and destroy the original message, including all copies, Thank you. ***
Re: exchange 2010
Hi Del, thanks, that's what I'd expected. What I'll do is something along the lines of: tdpexcc backup * incremental /excludedb="these_go_full_today" /preferdagpassive (and more stuff) tdpexcc backup "these_go_full_today" full /preferdagpassive (and more stuff) and have the script determine what "these_go_full_today" will be for each day of the week, based on a simple array. It'll probably be powershell because I like it and it does provide the functionality I'll need like actual access to a simple function to indicate the day of the week and arrays and stuff. I'll lookup the URL that you provided. O, and yes, what Erwann said. Op 30 okt. 2014, om 20:48 heeft Del Hoobler het volgende geschreven: > Hi Remco, > > The PREFERDAGPASSIVE tells DP/Exchange to only back up an active database > if there > are no current healthy passive copies of the database. If no healthy > passive copy > is available, the backup is made from the active copy. > > Most of the DAG backup best practices and methodology is in the Exchange > book > under "Database Availability Group (DAG) backups": > > > http://www-01.ibm.com/support/knowledgecenter/SSTG2D_7.1.1/com.ibm.itsm.mail.exc.doc/c_dpfcm_bup_replica_exc.html?lang=en > > > The key to "load balance" when setting up the scheduled backup script is > to have > a separate invocation of each database. > > For example: > > TDPEXCC BACKUP DB1 FULL /MINIMUMBACKUPINTERVAL=720 /PREFERDAGPASSIVE > TDPEXCC BACKUP DB2 FULL /MINIMUMBACKUPINTERVAL=720 /PREFERDAGPASSIVE > TDPEXCC BACKUP DB3 FULL /MINIMUMBACKUPINTERVAL=720 /PREFERDAGPASSIVE > TDPEXCC BACKUP DB4 FULL /MINIMUMBACKUPINTERVAL=720 /PREFERDAGPASSIVE > TDPEXCC BACKUP DB5 FULL /MINIMUMBACKUPINTERVAL=720 /PREFERDAGPASSIVE > > Then, run this command from each of the Exchange servers at or about the > same time. > > > > Del > > > > "ADSM: Dist Stor Manager" wrote on 10/30/2014 > 02:42:01 PM: > >> From: Remco Post >> To: ADSM-L@VM.MARIST.EDU >> Date: 10/30/2014 02:42 PM >> Subject: exchange 2010 >> Sent by: "ADSM: Dist Stor Manager" >> >> Hi All, >> >> we're sort of struggling to get the full backups of our exchange >> environment completed. The good news is that the mail admins have >> decided to split the exchange env. into many small databases, so I >> was thinking to create a script that backups a number of databases >> full and the rest incrementally and rotating the selection of >> databases to be backed up fully each day. >> >> I was just wondering about the preferdagpassive parameter. If I say >> 'backup db1,db2,db3,db4 full /preferdagpasive' and db3 and db4 are >> active databases on this DAG node, will they or will they not be >> backed up? In other words, is there a difference in handling for * >> vs. explicitly naming the databases to be backed up? >> >> (of course I'll run the exact same command on the other DAG node at >> more or less the same time to backup those databases that are >> passive over there). >> >> -- >> >> Met vriendelijke groeten/Kind Regards, >> >> Remco Post >> r.p...@plcs.nl >> +31 6 248 21 622 >> -- Met vriendelijke groeten/Kind Regards, Remco Post r.p...@plcs.nl +31 6 248 21 622
Re: exchange 2010
Hi Del, It would be nice to have a feature like vmmaxparallel for easy Exchange databases parallel backups. Le 30 octobre 2014 20:48:33 CET, Del Hoobler a écrit : >Hi Remco, > >The PREFERDAGPASSIVE tells DP/Exchange to only back up an active >database >if there >are no current healthy passive copies of the database. If no healthy >passive copy >is available, the backup is made from the active copy. > >Most of the DAG backup best practices and methodology is in the >Exchange >book >under "Database Availability Group (DAG) backups": > > >http://www-01.ibm.com/support/knowledgecenter/SSTG2D_7.1.1/com.ibm.itsm.mail.exc.doc/c_dpfcm_bup_replica_exc.html?lang=en > > >The key to "load balance" when setting up the scheduled backup script >is >to have >a separate invocation of each database. > >For example: > > TDPEXCC BACKUP DB1 FULL /MINIMUMBACKUPINTERVAL=720 /PREFERDAGPASSIVE > TDPEXCC BACKUP DB2 FULL /MINIMUMBACKUPINTERVAL=720 /PREFERDAGPASSIVE > TDPEXCC BACKUP DB3 FULL /MINIMUMBACKUPINTERVAL=720 /PREFERDAGPASSIVE > TDPEXCC BACKUP DB4 FULL /MINIMUMBACKUPINTERVAL=720 /PREFERDAGPASSIVE > TDPEXCC BACKUP DB5 FULL /MINIMUMBACKUPINTERVAL=720 /PREFERDAGPASSIVE > >Then, run this command from each of the Exchange servers at or about >the >same time. > > > >Del > > > >"ADSM: Dist Stor Manager" wrote on 10/30/2014 >02:42:01 PM: > >> From: Remco Post >> To: ADSM-L@VM.MARIST.EDU >> Date: 10/30/2014 02:42 PM >> Subject: exchange 2010 >> Sent by: "ADSM: Dist Stor Manager" >> >> Hi All, >> >> we're sort of struggling to get the full backups of our exchange >> environment completed. The good news is that the mail admins have >> decided to split the exchange env. into many small databases, so I >> was thinking to create a script that backups a number of databases >> full and the rest incrementally and rotating the selection of >> databases to be backed up fully each day. >> >> I was just wondering about the preferdagpassive parameter. If I say >> 'backup db1,db2,db3,db4 full /preferdagpasive' and db3 and db4 are >> active databases on this DAG node, will they or will they not be >> backed up? In other words, is there a difference in handling for * >> vs. explicitly naming the databases to be backed up? >> >> (of course I'll run the exact same command on the other DAG node at >> more or less the same time to backup those databases that are >> passive over there). >> >> -- >> >> Met vriendelijke groeten/Kind Regards, >> >> Remco Post >> r.p...@plcs.nl >> +31 6 248 21 622 >> -- Erwann SIMON Envoyé de mon téléphone Android avec K-9 Mail. Excusez la brièveté.
Re: exchange 2010
Hi Remco, The PREFERDAGPASSIVE tells DP/Exchange to only back up an active database if there are no current healthy passive copies of the database. If no healthy passive copy is available, the backup is made from the active copy. Most of the DAG backup best practices and methodology is in the Exchange book under "Database Availability Group (DAG) backups": http://www-01.ibm.com/support/knowledgecenter/SSTG2D_7.1.1/com.ibm.itsm.mail.exc.doc/c_dpfcm_bup_replica_exc.html?lang=en The key to "load balance" when setting up the scheduled backup script is to have a separate invocation of each database. For example: TDPEXCC BACKUP DB1 FULL /MINIMUMBACKUPINTERVAL=720 /PREFERDAGPASSIVE TDPEXCC BACKUP DB2 FULL /MINIMUMBACKUPINTERVAL=720 /PREFERDAGPASSIVE TDPEXCC BACKUP DB3 FULL /MINIMUMBACKUPINTERVAL=720 /PREFERDAGPASSIVE TDPEXCC BACKUP DB4 FULL /MINIMUMBACKUPINTERVAL=720 /PREFERDAGPASSIVE TDPEXCC BACKUP DB5 FULL /MINIMUMBACKUPINTERVAL=720 /PREFERDAGPASSIVE Then, run this command from each of the Exchange servers at or about the same time. Del "ADSM: Dist Stor Manager" wrote on 10/30/2014 02:42:01 PM: > From: Remco Post > To: ADSM-L@VM.MARIST.EDU > Date: 10/30/2014 02:42 PM > Subject: exchange 2010 > Sent by: "ADSM: Dist Stor Manager" > > Hi All, > > we're sort of struggling to get the full backups of our exchange > environment completed. The good news is that the mail admins have > decided to split the exchange env. into many small databases, so I > was thinking to create a script that backups a number of databases > full and the rest incrementally and rotating the selection of > databases to be backed up fully each day. > > I was just wondering about the preferdagpassive parameter. If I say > 'backup db1,db2,db3,db4 full /preferdagpasive' and db3 and db4 are > active databases on this DAG node, will they or will they not be > backed up? In other words, is there a difference in handling for * > vs. explicitly naming the databases to be backed up? > > (of course I'll run the exact same command on the other DAG node at > more or less the same time to backup those databases that are > passive over there). > > -- > > Met vriendelijke groeten/Kind Regards, > > Remco Post > r.p...@plcs.nl > +31 6 248 21 622 >
exchange 2010
Hi All, we're sort of struggling to get the full backups of our exchange environment completed. The good news is that the mail admins have decided to split the exchange env. into many small databases, so I was thinking to create a script that backups a number of databases full and the rest incrementally and rotating the selection of databases to be backed up fully each day. I was just wondering about the preferdagpassive parameter. If I say 'backup db1,db2,db3,db4 full /preferdagpasive' and db3 and db4 are active databases on this DAG node, will they or will they not be backed up? In other words, is there a difference in handling for * vs. explicitly naming the databases to be backed up? (of course I'll run the exact same command on the other DAG node at more or less the same time to backup those databases that are passive over there). -- Met vriendelijke groeten/Kind Regards, Remco Post r.p...@plcs.nl +31 6 248 21 622
Re: TSM 7.1 usage of volumes for dedupe
Bill, I just wanted to let you know how much this information helped. I was able to clear out all the problem volumes and have removed the full LUNs from the devclass until there is enough space on them to be used again. This situation really seems strange to me. Why has TSM not been updated to handle the out of space condition better? If it has a command that shows how much space is left on the LUN, why can't TSM understand it is time to stop allocating volumes on it? Forcing admins to do manual clean up like this just to keep things healthy seems inconsistent with how the rest of TSM functions. Has anyone ever reported this as a bug? Martha On 10/22/2014 2:38 PM, Colwell, William F. wrote: Hi Martha, I see this situation occur when a filesystem gets almost completely full. Do 'q dirsp ' to check for nearly full filesystems. The server doesn't fence off a filesystem like this, instead it keeps hammering on it, allocating new volumes. When it tries to write to a volume and gets an immediate out-of-space error, it marks the volume full so it won't try to use it again. I run this sql to find such volumes and delete them - select 'del v '||cast(volume_name as char(40)), cast(stgpool_name as char(30)), last_write_date - from volumes where upper(status) = 'FULL' and pct_utilized = 0 and pct_reclaim = 0 order by 2, 3 You should remove such filesystems from the devclass directory list until reclaim has emptied them a little bit. Hope his helps, Bill Colwell Draper Lab -Original Message- From: ADSM: Dist Stor Manager [mailto:ADSM-L@VM.MARIST.EDU] On Behalf Of Martha M McConaghy Sent: Wednesday, October 22, 2014 2:23 PM To: ADSM-L@VM.MARIST.EDU Subject: Re: TSM 7.1 usage of volumes for dedupe Interesting. Seems very similar, except the status of these volumes is "FULL", not "EMPTY". However, the %reclaimable space is 0.0. I think this is a bug. I would expect the volume to leave the pool once it is "reclaimed". It would be OK with me if it did not. However, since the status is "FULL", it will never be reused. That seems wrong. If it is going to remain attached to the dedupepool, the status should convert to EMPTY so the file can be reused. Or, go away altogether so the space can be reclaimed and reused. In looking at the filesystem on the Linux side (sorry I didn't mention this is running on RHEL), the file exists on /data0, but with no size: [urmm@tsmserver data0]$ ls -l *d57* -rw--- 1 tsminst1 tsmsrvrs 0 Oct 10 20:22 0d57.bfs /data0 is 100% utilized, so this file can never grow. Seems like it should get cleaned up rather than continue to exist. Martha On 10/22/2014 1:58 PM, Erwann SIMON wrote: hi Martha, See if this can apply : www-01.ibm.com/support/docview.wss?uid=swg21685554 Note that I had a situation where Q CONT returned that the volume was empty but it wasn't in reality since it was impossible to delete it (without discrading data). A select statement against the contents showed some files. Unforunately, I don't know how this story finished... -- Martha McConaghy Marist: System Architect/Technical Lead SHARE: Director of Operations Marist College IT Poughkeepsie, NY 12601 Notice: This email and any attachments may contain proprietary (Draper non-public) and/or export-controlled information of Draper Laboratory. If you are not the intended recipient of this email, please immediately notify the sender by replying to this email and immediately destroy all copies of this email. -- Martha McConaghy Marist: System Architect/Technical Lead SHARE: Director of Operations Marist College IT Poughkeepsie, NY 12601
Access to VE api
Hi again. last one from me and then I'll shut up. I've been toying for years with the idea of wrapping the TSM API so it could be accessed from modern scripting languages. One reason would be to re-implement something like the old adsmpipe program. Others would be to clean up leftovers of decommissioned nodes or trump Oracle DBAs who unilaterally exceed their agreed retention periods. Lately my thinking has been to do this in Java as then it would be available to anything using the JVM, such as Javascript with Rhino, Jruby, Jython, even Clojure if that suits. The Java FFI makes calling the C api possible. Unless you have been under a rock you will have heard about Docker. This technology is sweeping all before it and I can see it becoming ubiquitous very quickly. One of the features of Docker is the Docker volume which is like a private file system for it. So I was considering how to back up a docker volume, and one way would be to back it up as a single object from the host OS. That would be an excellent use case for the TSM VE API... that is, break the docker volume into 128MB chunks and run a subfile-style delta backup on each, with full copies being taken once the amount of change or number of deltas gets too large. However, looking at the API Docs, I see no exposure of the VE API. Anyone know of plans to make the VE API publicly available? Regards Steve Steven Harris Dreamer of programming way beyond his ability Canberra Australia
multi-threaded Solaris backups
Hi Again. A new customer has been signed, and while I don't think I'll get handed the job, the implications got me thinking. The customer is a strong Solaris user, and our solution is to coalesce the workload from multiple solaris boxes on to two latest generation machines one at each of two datacentres. There are many hundreds of servers and over 2000 existing cpus and this is all going to run on just two hi-spec machines, in Solaris containers. So question is how do we back this up? The simple way would be to install a client in every local zone, but that is unweildy and requires a lot of management. Its really not intelligent to work that way today. Backing up from the global zone would seem to be reasonable, but you'd need a stack of client instances running to get through the workload, possibly taking ZFS snapshots for consistency. If you wanted one node for each zone, then scheduling would be a problem becuase each schedule would need ASNODE specified and that means one schedule per node. Running multiple client processes would have us back in the bad place we were with fullVM backups prior to VE, where one stuck backup would derail a whole stream. Another issue would be automatically detecting new zones and added filesystems. Pretty much what is needed is a re-implementation of VE and the multi-threaded scheduler for the Solaris Cluster environment. Now down here in oz we are nowhere near the bleeding edge so someone else in the world must also be having these issues. Is anyone aware of any plans for such a beast? Has anyone wrestled with the problem and be willing to share some insights? Regards Steve Steven Harris TSM Admin Canberra Australia
Re: LTO5 Tuning
Steve: What is the latency on your intersite connection? Best regards, Mike, x7942 RMD IT Client Services On Thu, Oct 30, 2014 at 8:35 AM, Steven Harris wrote: > Hi All > > I have a TSM 6.3.4 server on newish P7 hardware and AIX V7.1. HBAs are > all 8Gb. The sans behind it are 8Gb or 4Gb depending which path they > take as we are in the middle of a SAN upgrade and there is still an old > switch in the mix. > > Disk is XIV behind SVC. Tape is TS3500 and LTO5. > > According to the LTO wikipedia entry I should be able to get 140MB/sec > raw out of the drive. I have an internal company document that suggests > sustained 210MB/sec (compressed) is attainable in the real world. > > So far my server backs up 500GB per night of DB2 and Oracle databases on > to file pools, without deduplication. Housekeeping then does a > single-streamed simultaneous migrate and copy to onsite and offsite > tapes. Inter site bandwidth is 4Gb and I have most of that to myself. > > That process takes over 5 hours so I'm seeing less than 100MB/sec. > > Accordingly I started a tuning exercise. I copied 50GB of my filepool > twice to give me a test dataset and started testing, of course when > there was no other activity on the TSM box. > > The data comes off disk at 500MB/sec to /dev/null, so that is not a > bottleneck. > > Copying using dd to tape runs at a peak of 120MB/sec with periods of > much lower than that, as measured using nmon's fc stats on the HBAs. I > presume some of that slowdown is where the tape reaches its end and has > to reverse direction. > > Elapsed time for 100GB is 18 min, with little variation so average speed > is 95MB/sec > > dd ibs and obs values were varied and ibs=256K obs=1024K seems to give > the best result. > > Elapsed time is very consistent. > > Copying to a local drive on the same switch blade as the tape HBA or > copying across blades made no difference. > > Copying to a drive at the remote site increased elapsed time by 2 > minutes, as one would expect with more switches in the path and a longer > turnaround time. > > Tape to tape copy was not noticeably different to disk to tape. > > Reading from tape to /dev/null was no different. > > In all cases CPU time was about half of the elapsed time. > > lsattr on the drives shows that compression is on (this is also the > default) > > The tape FC adapters are set to use the large transfer size. > > The test was also run using 64KB pages and svmon was used to verify the > setting was effective. Again no difference. > > I'm running out of ideas here. num_cmd_elements on the hbas is 500 (the > default) I'm thinking of increasing that to 2000, but it will require > an outage and hence change control. > > Does anyone have any ideas, references I could look at or practical > advice as to how to get this to perform? > > Thanks > > Steve > > Steven Harris > TSM Admin > Canberra Australia >
LTO5 Tuning
Hi All I have a TSM 6.3.4 server on newish P7 hardware and AIX V7.1. HBAs are all 8Gb. The sans behind it are 8Gb or 4Gb depending which path they take as we are in the middle of a SAN upgrade and there is still an old switch in the mix. Disk is XIV behind SVC. Tape is TS3500 and LTO5. According to the LTO wikipedia entry I should be able to get 140MB/sec raw out of the drive. I have an internal company document that suggests sustained 210MB/sec (compressed) is attainable in the real world. So far my server backs up 500GB per night of DB2 and Oracle databases on to file pools, without deduplication. Housekeeping then does a single-streamed simultaneous migrate and copy to onsite and offsite tapes. Inter site bandwidth is 4Gb and I have most of that to myself. That process takes over 5 hours so I'm seeing less than 100MB/sec. Accordingly I started a tuning exercise. I copied 50GB of my filepool twice to give me a test dataset and started testing, of course when there was no other activity on the TSM box. The data comes off disk at 500MB/sec to /dev/null, so that is not a bottleneck. Copying using dd to tape runs at a peak of 120MB/sec with periods of much lower than that, as measured using nmon's fc stats on the HBAs. I presume some of that slowdown is where the tape reaches its end and has to reverse direction. Elapsed time for 100GB is 18 min, with little variation so average speed is 95MB/sec dd ibs and obs values were varied and ibs=256K obs=1024K seems to give the best result. Elapsed time is very consistent. Copying to a local drive on the same switch blade as the tape HBA or copying across blades made no difference. Copying to a drive at the remote site increased elapsed time by 2 minutes, as one would expect with more switches in the path and a longer turnaround time. Tape to tape copy was not noticeably different to disk to tape. Reading from tape to /dev/null was no different. In all cases CPU time was about half of the elapsed time. lsattr on the drives shows that compression is on (this is also the default) The tape FC adapters are set to use the large transfer size. The test was also run using 64KB pages and svmon was used to verify the setting was effective. Again no difference. I'm running out of ideas here. num_cmd_elements on the hbas is 500 (the default) I'm thinking of increasing that to 2000, but it will require an outage and hence change control. Does anyone have any ideas, references I could look at or practical advice as to how to get this to perform? Thanks Steve Steven Harris TSM Admin Canberra Australia