nav[aria-label="Primary Navigation"] { padding: 0; & ul { list-style: none; width: 100%; display: flex; flex-direction: row; justify-content: start; align-items: start; gap: 30px; padding: 0; & li { margin: 0; } & ul li { list-style: none; } } }

deleting branches

msrich

When I delete a branch in TS 5.5.2 (W2K, SP2), I see a batch job running (seen in IWStat output). How can I tell how much longer the batch job will run? Can I get any stats about the batch job itself?

We've had mixed results deleting branches: some deletes linger in batch for days, and some finish within a half hour.

Find more posts tagged with

Comments

gwen1

Information I have received is that there is no way to tell how long the job will run. But branches with more content and larger amount of history will take longer. We are currently removing some branches and larger branches have taken between 18-48 hours to complete Cleanup.

Adam Stoller

You can run iwstat -c which among other things should show you the number of dirty items in the cache there are to be processed. It is not a linear algorithm to determine how long it will take though. As I recall the fewer dirty items to be processed the faster it goes (I realize that sounds obvious, but I mean in terms of number of items cleared per time unit).

I believe, though am not certain, that the better equipped your hardware is, the better perfomance you will have with respect to this.

A dirty item is essentially data that is currently being held in memory that needs to be flushed out to disk - so yes, the bigger the branch (files, versions, editions) the longer it will take to remove it.

Sorry I cannot provide more specific numbers for the above, but I hope it helps.

--fish
(Interwoven Senior Technical Consultant)

3nodestable55.ssd.zip

msrich

Thanks very much for the information. I'm a little concerned, then, because I have two branch deletes that have been running since 5 p.m. Monday 8/11. We have put batch jobs on hold during business hours, but they've been free to run all night for at least three nights. I see no 'dirty' pages at all in iwstat output. Are these batch jobs (and perhaps the workarea deletes that have queued up behind them) stuck, in some fashion?

C:\Profiles\svcbackup>iwstat -c
Store Status
default Running
workflow Running

ID Thread User Duration Store Operation
0x5d69bb1 0x538 vanguard\UMIR 0.000 - GetServerStatus
0x5d69bac 0xd78 VANGUARD\ATEAP 0.109 default CFILE_SETATTR_svc (ce,400262bf) 48c4c7/5e2847
0x5aa3d73 0x9b4 - 58807.594 default Batch Jobs

Store: default
Batch Jobs (running):
Queued User Job Object ID
[Mon Aug 11 17:04:29 2003] vanguard\UMIR 0x00002250000000000015e400
[Mon Aug 11 17:04:29 2003] vanguard\UMIR 0x00002250000000000030b600
[Tue Aug 12 11:25:12 2003] VANGUARD\URG4 Cleanup Destroyed Workarea u4jw 0x0000210000000000005e28e3
[Tue Aug 12 11:25:38 2003] VANGUARD\URG4 Cleanup Destroyed Workarea u4jw 0x0000210000000000005e27cb
[Tue Aug 12 11:26:02 2003] VANGUARD\URG4 Cleanup Destroyed Workarea u4jw 0x0000210000000000005e2cf6
[Tue Aug 12 11:26:20 2003] VANGUARD\URG4 Cleanup Destroyed Workarea u4jw 0x0000210000000000005e29e3
[Tue Aug 12 11:26:41 2003] VANGUARD\URG4 Cleanup Destroyed Workarea u4jw 0x0000210000000000005e2e02
[Tue Aug 12 11:27:19 2003] VANGUARD\URG4 Cleanup Destroyed Workarea u4jw 0x0000210000000000005e2ae3
[Tue Aug 12 11:27:37 2003] VANGUARD\URG4 Cleanup Destroyed Workarea u4jw 0x0000210000000000005e2be2
[Tue Aug 12 11:27:59 2003] VANGUARD\URG4 Cleanup Destroyed Workarea u4jw 0x0000210000000000002d0436
[Tue Aug 12 16:32:10 2003] VANGUARD\URDG Cleanup Destroyed Workarea 4Q03 0x0000210000000000005e4470

Cache Active Dirty To Purge Available
default 179742 0 0
workflow 9 0 0
Total 179751 0 0 20249

Minutes Thruput Avg op Load
1 48 0.0417 2.00
15 46 0.0644 2.96
60 29 0.0471 1.36
300 41 0.0401 1.64
1440 102 0.0118 1.20
2880 94 0.0113 1.06
5760 227 0.0002 0.04
9620 169 0.0028 0.47

Adam Stoller

I'd suggest you contact Interwoven Support - it sounds like something might be amiss here, and it would be far better for you to work with them to resolve it.

--fish
(Interwoven Senior Technical Consultant)

Migrateduser

Batch jobs are done sequentially. So everything is backed up behind that first "vanguard\UMIR" job. The others are not taking a long time, they are not running at all. Branch deletions can take a long time, depending on how big it is. I've had (pathological) test deletes take days of continuous running.

You might be able to suspend or abort the offending job, letting the others proceed.

Does iwtrace.log show any problems?

James E. Pace
Quality Engineer
Interwoven

Canceled.lxf.zip

output.pdf

Migrateduser

Hi,

How do you suspend a branch deleting job?

Regards,

james1

I think that "iwfreeze -j" will suspend all batch jobs. See "iwfreeze -h" for more information.

-- James

--
James H Koh
Interwoven Engineering